TRANSACTIONS OFTIIE SPWLA THIRTY-SIXTH ANNUAL LOGGING SYMPOSIUM Sponsored by THE SOCIETY OF PROFESSIONAL WELL LOG ANALYSTS, INC. 8866 Gulf Freeway, Suite 320 Houston, Texas 77017 Presented at THE PALAIS des CONGRI% Paris, France June 26-29, 1995 NOTICE TO EDITORS: Permission is hereby granted to publish elsewhere any of these transactions after June 29, 1995, provided that conspicuous acknowledgement is given to the original presentation of the paper and the authors of the paper have agreed to the republication. (The statements and opinions expressed in these transactions are those of the authors and should not be construed as an official action or opinion of the Society of Professional Well Log Analysts, Inc.) th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER A TOWARDS FAST QUANTITATIVE MODELLING OF PULSED NEUTRON LOGGING TOOLS R.L. Jennings and G.A. Weber Koninklijke/Shell Exploratie en Produktie Laboratorium, Rijswijk, The Netherlands ABSTRACT Current practice for the interpretation of pulsed neutron logging data involving empirical fitting formulae together with the application of ad hoc corrections is not ideal. This is because the fitting procedure is too simplistic, and furthermore the correction factors are both difficult and expensive to acquire. To remedy this we have developed a new physical model of pulsed neutron logging based upon a macroscopic theory of neutrons and γ-rays. Such a macroscopic description is much simpler than Monte Carlo simulations which track individual neutrons. In order that the model is suitable for practical logging purposes we have considered only those nuclear processes that we deemed essential. However, we shall show that these simplifications are justified. Principally, we model the cooling of highly energetic neutrons, the diffusion and capture of thermal neutrons together with the transport of γ-rays in a borehole-casing-formation environment. This yields a predicted γ-count rate at the detector which can be compared with the count rate actually recorded. Via this approach it is possible to extract the thermal neutron capture cross section of the formation. In order to validate our model we show a comparison of our model’s results with experimental data collected in a well defined environment. The agreement thus obtained is very satisfactory and provides evidence that our model captures the essence of pulsed neutron logging. Specifically, our results demonstrate the importance of treating the diffusion of thermal neutrons as well as the need to include the metal casing in the modelling. This latter point is noteworthy, since to date the casing’s role has not been sufficiently emphasised. In conclusion, we hope to replace existing interpretation schemes with a method that is underpinned by physical principles. Moreover, we wish to stress that the model is sufficiently flexible that it can easily be extended to more realistic logging, configurations. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER B THE ACCURACY AND PRECISION OF FEMWD DENSITY TOOLS Daniel C. Minette Baker Hughes INTEQ, Houston, TX Tracey Flynn Baker Hughes INTEQ, Aberdeen, Scotland ABSTRACT The accuracy of the density measurement is critical to formation evaluation. In order for FEMWD (Formation Evaluation Measurement While Drilling) technology to provide a true standalone formation evaluation service, a consistent and accurate density measurement must be provided, This consistency and accuracy must be demonstrated for more than a few ideal conditions. It must be shown for the full range of formations and muds encountered downhole. While the formations and mud systems encountered by FEMWD and wireline density tools are the same, there are significant operational, environmental and design differences between wireline and FEMWD density logging. FEMWD tools are usually run in boreholes that are in better condition than those available to wireline tools. The formation viewed by the FEMWD tools is routinely less invaded. FEMWD tools must be able to acquire accurate data whether rotating or sliding unlike the mode of acquisition of wireline density data, which is run only in sliding mode. With these facts in mind, a laboratory program was developed to fully characterize two density tool sizes that have recently undergone a significant upgrade in both precision and accuracy: the 6 ¾” and the 8 ¼” FEMWD density. For this characterization, formations with densities ranging from 1.7 g/cc to 3.0 g/cc and Pe spanning 2.2 to 6.0 barns/electron are used. Water standoffs of 0” to 1/2”, as well as mudcakes ranging from 1/8” to 3/8” thickness were used to determine the standoff response. The density and Fe in these mudcakes vary from 1.0 to 2.4 g/cc, and 0.3 to 120 barns/electron (b/c), respectively. Data were taken for individual standoffs and were also integrated over a range of standoffs in order to properly emulate downhole conditions. In addition to laboratory data, field data are required to verify the accuracy of any formation evaluation tool. In this paper, several field examples are used to demonstrate the tool accuracy. These data also span a wide range of formation conditions. Formation densities range from less than 2.00 g/cc to over 2.70 g/cc. Mud weights vary from about 1.2 g/cc to 1.9 g/cc. In addition, both sliding and rotating data are considered. This analysis includes consideration of both the accuracy and precision of the measurement. Accuracy is determined by comparison with other measurements such as wireline logs and cores. Precision will be determined by the consistency of the measurement. In this manner, the response of the tool is verified in fields under conditions that are representative of those encountered by the European formation evaluation community. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER C ADAPTING WIRELINE LOGGING TOOLS FOR ENVIRONMENTAL LOGGING APPLICATIONS D. V. Ellis, R. A. Perchonok, H. D. Scott, and C. Stoller Schlumberger ABSTRACT Recently developed environmental awareness has spawned governmental regulations which will require cleanup of numerous hazardous waste sites. Many of the largest sites currently targeted for cleanup in the US have been associated, in the past, with nuclear weapons production. Subsurface pollutants at these sites range from chemicals and heavy metals to man-made radioactive substances, including plutonium. A first step in the remediation process is site assessment: the mapping and monitoring of the subsurface contamination, and the quantification of the hydrogeologic parameters necessary to support models that predict contaminant transport, and health risks, and support the selection of cleanup alternatives. The potential of wireline logging to map subsurface properties is relatively unappreciated by the environmental community. Although existing wireline logging technology may not be able to respond to all subsurface contamination mapping needs, a number of logging techniques seem well suited for the quantification of hydrogeologic parameters, like the measurement of soil moisture and detection of radioactive wastes. Under a cooperative research and development agreement (CRADA) with two groups at the Hanford Site, we were invited to adapt a neutron porosity tool and a natural gamma ray spectrometer for monitoring soil moisture and detecting several manmade radioisotopes in the vadose zone. To accomplish the adaptation, the response of a neutron porosity tool was determined, through the use of Monte Carlo modeling and experimental measurements, for the unique operating conditions of the vadose zone: a nearly dry 30 to 40 p.u. formation with air-filled cased holes. Additional gamma ray spectral standards were determined experimentally for the detection and quantification of 137Cs and 60Co behind the casing. The modified algorithms were demonstrated on data acquired in test tanks at the Hanford Site and in several monitoring wells. Log examples illustrate detection of 137Cs and 60Co contamination in several wells and the simultaneous real-time measurement of formation moisture and variations in man-made radioisotopes at another contaminated location. . th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER D THE RESPONSE OF INDUCTION TOOLS TO DIPPING, ANISOTROPIC FORMATIONS Barbara I. Anderson, Thomas D. Barber, and Martin G. Lüling Schlumberger ABSTRACT Directional wells, especially horizontal wells, are commonly drilled today to enhance reservoir productivity and minimize unwanted production of water or gas. At the steep apparent dip angles encountered, logging tool response characteristics change. Induction tools become more sensitive to bed boundary location. They also detect resistivity anisotropy, which remained largely invisible in vertical wells. The interpretation of induction logs in directional wells poses several challenges. Like all logging tools, induction tools were developed for wellbores perpendicular to the bedding planes. The measurements provide several radial depths of investigation. Separation of the logs is generally caused by invasion, and this separation provides a radial resistivity profile. However, in directional wells, a cap shale or an aquifer can cause induction curves to separate because the multiple depths of investigation have different sensitivities to beds adjacent to the zone of interest. Thus curve separation no longer indicates invasion exclusively. In anisotropic formations, induction tools sense a weighted average of the horizontal and vertical resistivities. This observed resistivity may differ considerably from the resistivity of a nearby vertical reference well where induction tools sense only the horizontal resistivity. In these complex formation geometries, forward modeling can provide a reliable resistivity interpretation of tool response in layered, anisotropic media. The computer modeling program generates logs for either dual induction or array induction tools. The tool can be tilted at any user-provided angle against a layer-cake formation. Each layer may have an intrinsic resistivity anisotropy. A major advantage of the program is that computer run-time is independent of the number of beds modeled. The modeling program is used to study the sensitivity of both array induction and dual induction tools to anisotropy. In thick beds, anisotropy alone can cause the induction curves to separate because the mixing of the horizontal and vertical resistivities distorts the skin effect correction. Curve separation is most noticeable in conductive zones and at high dip angles. At steeply dipping bed boundaries, polarization horns appear, similar to those occurring on 2MHz resistivity logs. These horns are most prominent on the long arrays. A field log example is used to illustrate the use of tool response modeling in an iterative inversion for Rt. The induction interpretation is corroborated by 2-MHz resistivity measurements obtained while drilling. Logs are modeled in both vertical and near-horizontal wells in the same layered formation. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER E INTEGRATED 2-D INTERPRETATION OF RESISTIVITY LOGGING MEASUREMENTS BY INVERSION METHODS AG. Mezzatesta, M.H. Eckard, and K.-M. Straack Western Atlas Logging Services, Western Atlas International, Houston, TX ABSTRACT Accurate estimates of residual and movable hydrocarbon require precise estimation of the formation resistivities in the invaded and uncontaminated zones. Several resistivity measurements based on different physical principles and providing different vertical and radial resolutions are normally available to the log analyst. The integration of these measurements leads to a single distribution of resistivities around the borehole, capable of explaining the entire set, and allowing the interpreter to rationalize the differences among them. Conventional interpretation methods of electromagnetic measurements use approximate and independent techniques for laterolog and induction logging measurements. These techniques are typically designed for the interpretation of the measurements that correspond to a single suite of logs, causing results that may differ when applied to the different sets of log measurements available in a particular borehole. The final interpretation is then left to the skill of the log analyst. An integrated interpretation of laterolog and induction measurements based on a 2-dimensional inversion process, provides a single resistivity distribution (perpendicular and parallel to the borehole direction) consistent with all resistivity measurements. The inversion process allows the tool behavior to be simulated and simultaneously accounts for borehole, invasion, and shoulder effects. The inversion does not require the tedious process of normalizing the data in vertical resolution. In addition, the inversion process provides statistical results that help the log analyst assess the quality of the interpretation results. Combining measurements based on different physical principles also improves the resolution of the estimated formation parameters, such as formation resistivities, invasion depths, and bed boundaries. This work demonstrates how the interpretation is improved by using an algorithm that simultaneously considers all available resistivity measurements. The work also shows the importance of the inversion process in extracting the information content from the data in cases that would not be possible by using conventional interpretation methods, as well as the use of statistical results to assess the quality of the interpretation. Several field cases are presented that show the importance of unified interpretations in improving the accuracy of the formation resistivity estimation, and the enhancement of radially and vertically layered earth models that result. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER F COMPARISONS OF WIRELINE AND LWD RESISTIVITY HIGHLIGHT RESISTIVITY FREQUENCY DISPERSION IN SEDIMENTARY FORMATIONS Roland Chemali, Dale Heysse, G. A. Merchant, and Charles Jackson Halliburton Energy Services ABSTRACT Some differences in resistivity measurements between wireline induction and LWD propagation instruments may be explained by the phenomenon of resistivity frequency dispersion. The resistivity of sedimentary formations varies with measurement frequency. From a petrophysical point of view, the frequency dispersion of resistivity is related to shaliness, pore tortuosity and pore fluid salinity. In recent developments, LWD resistivity has been matched to wireline resistivity to a certain extent by postulating a realistic value for the dielectric constant of the formation, and entering this parameter into the wave propagation model of the LWD resistivity device. However, when applied to a newly developed dual spaced 2 MHz LWD device, this approach sometimes gave unsatisfactory results, in which the correction for dielectric constant could not adequately bridge the gap between all four measurements of the 2 MHz dual spaced LWD and the wireline induction. The concept of variability of formation resistivity with frequency is confirmed by published theoretical and experimental results. Several models for electrical rock properties implicitly predict that the resistivity of sedimentary formations decrease significantly as the frequency increases from near 0 Hz to the 1 GHz range. One such model is Wyllie’s equation, which was originally developed for acoustic waves in porous media, then successfully applied to electromagnetic microwaves. In that application, Wyllie’s equation is better known as Complex Refractive Index Method or CRIM. The formal expression of CRIM contains a function of frequency which can be readily expanded to lower frequencies. Surprisingly, as the frequency approaches those of induction and laterolog the CRIM formula coincides with Archie’s equation thus reconciling Archie and Wyllie in a unified theory. In the mid-frequency range of 2 MIHz the resistivity predicted by CRIM may be significantly lower than the resistivity measured at the wireline induction frequency of 20 kHz. Resistivity measurements performed above 100 MHz exhibit the same dispersion to an even larger extent. Other mixing laws, such as the HanaiBruggemann or the Maxwell-Wagner-Sillars equations, also predict frequency dispersion of resistivity in porous media as a function of pore or grain geometry. The dispersive nature of sedimentary formations is illustrated in log examples where resistivity logs were run using instruments operating at frequencies extending from near d-c to microwave frequency, including the 2 MHz range. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER G INDUCTION LOG FORWARD MODELING: A RIGOROUS AND SYSTEMATIC APPROACH TO MODEL CONSTRUCTION W. D. Kennedy Mobil Exploration and Producing Technical Center, Dallas, TX ABSTRACT In many instances the apparent resistivity response of the 6FF40 induction logging instrument is not a good estimate of the true formation resistivity. In some cases this is true in beds up to 200 feet thick. Traditional chart book corrections do not usually remedy this condition; however, when conditions warrant one-dimensional forward modeling can convert induction log apparent resistivity responses into more accurate estimates of formation resistivity. With the advent of commercially available, fast modeling codes packaged in easy-to-use interfaces, application of l-d forward modeling has become a feasible interpretation option with the potential of substantially increasing reserve estimates. Unfortunately, the modeling of several hundred to several thousand feet of induction log response may seem a daunting task regardless of the speed of computer codes and the convenience of user interfaces and the experience of the analyst. However, it has been discovered that, regardless of how complicated a log may appear, log responses can be catalogued into six easily recognized responses. The responses have been named for convenience: (1) the impulse response; (2) the step response; (3) the ramp response; (4) the whole space response; (5) the thin bed (blind frequency and anti-correlation) responses; (6) mixed responses. Responses not falling into these categories can be recognized and identified as two-dimensional responses or in some cases erroneous responses due to, e.g., incorrectly set sonde errors. The six responses are founded in the tool physics, but they can also be used as practical rules of thumb; using the six responses to construct initial models and refine the subsequent results significantly reduces modeling time. Both field and theoretical examples of each type of response are illustrated. Recognition of the cataloged responses permits efficient forward modeling. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER H TOTAL POROSITY ESTIMATION IN SHALY SANDS FROM SHEAR MODULUS Raghu Ramamoorthy and William F. Murphy Schlumberger-Doll Research, Ridgefield, CT, U.S.A. Carolina Coil Maraven S.A., Caracas, Venezuela ABSTRACT Measured shear velocities in clastic reservoir rocks have been shown to be independent of the type of fluid present in the pore space while being influenced by the porosity. There is promise that in hydrocarbon (particularly gas) reservoirs one may derive an independent estimate of porosity from the shear velocity of the rock. It has been shown that sonic velocities are best interpreted when decomposed into moduli that are intrinsic measures of rock frame and pore fluid mechanical properties. While the effect of porosity on the shear modulus has been documented for clean sandstone, these relationships break down in the presence of even limited amounts of clay mineral in the sandstone. Existing correlations between shear velocity, porosity and clay content appear to poorly represent the physics of acoustic transport in the rock. We propose a method to estimate total porosity in shaly sands from the shear modulus and a knowledge of the clay content of the rock. The formula accounts for the clay effect in a manner consistent with the known behavior of clay in rock. The method is successfully demonstrated in two well examples with porosity ranging from 10 p.u. to 38 p.u., dry clay volumes from 0% to 60%, and in zones with water, oil or gas in the pore space. Clay concentrations are determined with an accuracy of +/-2%. Comparisons are also presented with existing correlations showing the superior fit of the proposed technique over a wide range of porosity and clay content. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER I CASED HOLE ACOUSTIC LOGGING — A SOLUTION TO A PROBLEM Steve Chudy and Gordon Mcintyre Western Atlas Logging Services, Aberdeen, Scotland Paul Schuh Conoco (UK) Limited, Aberdeen, Scotland ABSTRACT Whether due to hole conditions or to reduce rig time, openhole logging is sometimes foregone over certain intervals of a well. However, the depth-derived formation data from these and subsequently logged intervals require the use of acoustic velocity measurements over as large a depth range as possible to confirm the depth-time relationship with surface seismic data. With the introduction of array-type acoustic logging tools, acquiring full-waveform data in the cased section of a wellbore is now possible whilst pulling out of the hole following a deeper logging suite. These waveform data can be processed to extract the acoustic velocities from the cased interval, thereby foregoing the requirement to log the interval in open hole, and thus saving the associated time and costs. This case study focuses on waveform acquisition from array-type acoustic instruments run in a number of wells in both the southern and northern sectors of the North Sea. The postprocessing techniques and acquisition constraints will be reviewed. Acoustic data acquired over the same intervals in both open and cased wellbores will be used to demonstrate the accuracy of the measurements. Finally, synthetic seismic techniques will be used to demonstrate the response validity in cases where poor cement would preclude the use of other acoustic devices. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER J FRACTURE AND STRESS EVALUATION USING DIPOLE-SHEAR ANISOTROPY LOGS C. Esmersoy, M. Kane Schlumberger-Doll Research A. Boyd and S. Denoo Schlumberger Wireline & Testing ABSTRACT Fractures and stresses, important factors in production and development of reservoirs, influence the propagation of shear waves. This effect has been documented as anisotropy, where the shear waves split into two; fast- and slowshear, with orthogonal particle-motion polarization directions. In fractured or horizontally stressed rocks the fastshear waves are polarized along the direction of the fracture strike and/or maximum horizontal stress. The amount of anisotropy, e.g., the percentage difference in slownesses, varies with fracture frequency or the amount of unbalanced subsurface stress. Seismic and VSP applications of shear-wave anisotropy are used to determine the large-scale character of the fractures and stresses. Dipole sonic logging, for the first time, makes high-resolution, sonic-scale shear anisotropy measurement possible. Wellbore imaging tools give valuable information regarding fracture orientation and frequency in open-hole environments, but current technology cannot be used in cased wells to detect fractures. Furthermore, although stress azimuth can be inferred from the borehole ovality, measured by wellbore imaging tools, this method is not always robust. Therefore, an acoustic anisotropy log could be a very valuable complement to the existing techniques in fracture and stress evaluation. We describe acquisition and processing techniques for dipole-shear anisotropy logging, and present two examples. One example deals with predicting the azimuth of a hydraulically induced fracture. The dipole-shear prediction was verified later by tiltmeter measurements. Another example demonstrate the use of this technique in a cased well for detecting fractured zones. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER K HYDROCARBON IDENTIFICATION IN FRESH-WATER BEARING RESERVOIRS USING DYNAMIC POISSON’S RATIO: A CASE STUDY Luis Guedes Condessa Petroleo Brasileiro S. A., Rio de Janeiro, Brazil ABSTRACT The characterization of pay zones in fresh-water bearing reservoirs has been troublesome for Petrobras log analysts in many oilfields of Potiguar Basin, northeastern Brazil. Routinely, the MSFL curve has been used to identify intervals with hydrocarbons in addition to acoustic, radioactivity and porosity data. Meanwhile, several unsuccessful exploration results produced a quest for new solutions to aid the identification of potential sequences in this kind of reservoir. One of them is the dynamic Poisson’s Ratio, an elastic property derived from the processing of sonic data recorded with a full waveform acoustic tool, which has proven to be an efficient qualitative evaluation technique in some areas of the Potiguar Basin. In this paper we present a detection technique for gas-bearing intervals in clean or shaly fresh-water saturated sandstones and provide information that can be used to help on decisions respecting drill stem tests in future wells, in that area. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER L SLOWNESS AND DELAY IN ACOUSTIC LOGGING J.L. Marl, F. Coppens Institut Francais du Pétrole, Rueil-Malmaison, France E. Blondin Sartrouville, France E. Wicquart Gaz de France, La Plaine Saint Denis, France ABSTRACT Formation slowness is commonly estimated by using the first arrival times of the compressional headwave recorded by the different receivers of an acoustic tool. In the presence of near borehole changes such as in altered or damaged zones (due to drilling), refracted borehole arrivals may pass from several inches to several feet from the borehole wall and be time delayed. The delays are computed using the intercept time method which has been widely used for the interpretation of refracted arrivals in land surface seismic surveys. They can be used to determine the altered zone extension. However delay computation requires an accurate picking of the true first break times of refracted arrivals. The picked times can then be used to obtain both slowness logs and delay logs. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER M NMR CORE ANALYSIS OF LOWER SAN ANDRES/GLORIETA/UPPER CLEAR FORK (PERMIAN) CARBONATES: CENTRAL BASIN PLATFORM, WEST TEXAS Shinichi Sakuraf and Robert G. Loucks ARCO Exploration and Production Technology John S. Gardner NUMAR Corporation ABSTRACT Lower San Andres/Glorieta/Upper Clear Fork carbonates on the Central Basin platform in West Texas are geologically and petrophysically complex. Within these carbonates, it is very difficult and time consuming to develop a well-defined porosity-permeability relationship. Nuclear Magnetic Resonance (NMR) technology was tested to see whether T2 distributions, derived from this technique, can be used to improve our understanding of the petrophysics of carbonates. Pore-body size distribution from thin-section point-count analysis and pore-throat size distribution from capillary pressure analysis are compared to derived pore-size distribution from NMR analysis. The overall correlation of core porosity to core permeability in the carbonates studied is poor because of the complex mineralogy, broad pore-size distribution, and range of pore-system types. NMR technology used to evaluate the carbonate rocks was based solely on core analyzer measurements made in a laboratory environment Sample selection was designed to cover a wide range of porosity and permeability values, as well as different pore networks from different fades. NMR porosity was obtained from curve fitting of the echo trains using an unconstrained bi-exponential model. Dimensions of pores were measured from thin-section analysis. Mercury injection capillary pressure tests were made to determine pore-throat size distribution. A comparison of “T2 distribution” measured by NMR and “pore-body-size distribution” from thin-sections made from the end of plug show a strikingly similar pattern. NMR T2 distribution also agrees well with “pore-throat-size distribution’ by mercury injection on the same plug used for the NMR tests. An excellent relationship between macroporosity, obtained from core NMR T2 measurements, and core permeability was obtained. The elimination of non-effective micro-porosity from the correlation is necessary to develop a useful porosity-permeability transform. Within these carbonates, NMR technology can be used to help estimate pore-size distribution and permeability. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER N NMR LOGGING OF NATURAL GAS RESERVOIRS R. Akkurt Shell Offshore Inc., New Orleans, LA H. J. Vinegar, P.N. Tutunjian Shell Development Co., Houston, TX A.J. Guillory Shell Offshore Inc., New Orleans, LA. ABSTRACT The phenomenon causing reduced NMR porosities in gas reservoirs, the so called “gas effect”, has lately become a subject of great interest in the petrophysical community. Contrary to the industry wide belief, NMR logging tools can detect gas provided that the pulse sequences are chosen properly and that the logging tool has adequate depthof-investigation. Furthermore, gradient-based logging tools such as the MRIL-C can be used to unambiguously identify the gas phase in the reservoir. Failure to recognize gas may result in gas being misinterpreted as bound fluid, which in turn may result in excessively high irreducible water saturations and incorrect permeability estimates. The NMR properties of gas are quite different from those of water and oil under typical reservoir conditions and this can be used to quantify the gas phase in a reservoir. A new NMR-only interpretation approach based on this principle, called the Differential Spectrum Method (DSM), has been developed and successfully tested in the Gulf of Mexico. This method utilizes properly selected NMR pulse sequences and does not require resistivity or other porosity logs. The DSM can be used in reservoirs containing gas and/or oil. Another technique exploiting the diffusion properties of gas, called the Shifted Spectrum Method (SSM), is also introduced. Hydrocarbon saturations computed using the Differential and Shifted Spectrum Methods show very good agreement with those obtained conventionally. The methods are mineralogy independent and insensitive to clay bound water, and ideal for shaly sand applications, th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER O CORE AND LOG NMR MEASUREMENTS OF AN IRON-RICH, GLAUCONITIC SANDSTONE RESERVOIR Wm. Scott Dodge Sr. ESSO Australia Ltd., Melbourne, Victoria, Australia John L. Shafer and Angel O. Guzman-Garcia Exxon Production Research Company, Houston, TX ABSTRACT NMR porosity and relaxation time measurements from an iron-rich. glauconitic sandstone reservoir show quantifiable effects of mineral iron content on NMR T2 relaxation times. This result has significant impact upon measuring irreducible water pore volume where the surface relaxation mechanism is nonconstant. Centrifuge air/brine drainage capillary pressure measurements show that the standard 30 msec T2 cutoff must be lowered to calibrate irreducible water saturation computed from NMR. Although the effects of iron are observable on T2 distributions, permeability estimation from NIMR, using either the Coates or Schlumberger relationships, show excellent agreement to permeability on core plugs. Quantitative mineral composition on core plugs using both XRD and XRF, show iron-rich glauconite to vary from 3 to 31 weight percent. The bulk rock total iron oxide content ranges from 1 to 17 weight percent. High iron content within this reservoir raised concern that NMR surface relaxation would be affected, leading to errors in irreducible water saturation and producible porosity derived from NMR measurements. NMR measurements were acquired using a pulsed field gradient logging tool operating at 530 kHz and on core plugs with a 1000 kHz laboratory spectrometer. Homogenous field NMR core plug measurements are used to show the accuracy of the logging tool to measure NMR porosity, and permeability. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER P APPLYING LOG MEASUREMENTS OF RESTRICTED DIFFUSION AND T2 TO FORMATION EVALUATION G. R. Coates, D. L. Miller, D. Mardon, and J. S. Gardner NUMAR Corporation ABSTRACT We describe a method for interpreting borehole measurements of pore fluid diffusion made with a fixed-gradient, pulsed NMR logging tool. The method is based on a simple mathematical model for relaxation and diffusion in oilwater mixtures contained in a single, water-wet pore. The model incorporates the effects of pore surface relaxation on the water as well as restricted diffusion in both the oil and the water. The log interpretation procedure uses cross plots of intrinsic transverse relaxation time (T2) vs. the effective diffusion coefficient D of the pore fluid mixture. Locating a point on the cross plot simultaneously yields the near-wellbore water saturation and the rock pore size. Limited laboratory NMR measurements on Bentheim sandstone cores containing oil-water mixtures provide a preliminary check on the validity of the model, A trial application of the cross plot technique to interpreting well log data yields saturation and pore size data that are consistent with other logs and the reservoir geology. The results point to a potential advantage of diffusion logging over standard T2 logging for characterizing formation pore size and related reservoir flow properties. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER Q RESIDUAL OIL SATURATION MEASUREMENTS IN CARBONATES WITH PULSED NMR LOGS John P. Horkowitz Pennzoil Exploration and Production Co., Houston, TX Harold J. Vinegar Shell Development Co., Bellaire Technology Center, Houston, TX Donald E. Hartman Pennzoil Exploration and Production Co., Houston, TX George R. Coates NUMAR Corporation, Houston, TX Edward A. Clerke Pennzoil Exploration and Production Co., Houston, TX ABSTRACT A new technique has been developed for determining residual oil saturation using pulsed nuclear magnetic resonance (NMR) logging. This technique retains the accuracy and reduces the costs and operational problems of the Mn-EDTA inject-log technique, which is considered to be the most accurate method of determining residual oil saturation in open holes. The NMR log is unique because, unlike other logging tools, the signal measured is derived only from the pore fluids. In the inject-log technique the brine is doped with a paramagnetic ion to “kill” the water signal, i.e. shorten its relaxation time below the dead time of the tool, so that only oil is detected. The new generation of pulsed NMR logging tools has greatly simplified this method, both by eliminating the need for doping the drilling mud with magnetite and by reducing the tool dead time from 30 ms to about 1 ms. Using a new NMR doping method, we have found that in carbonate reservoirs in West Texas the starch-based drilling mud can be successfully doped with MnCI2 , which is an order of magnitude less expensive than chelated Mn-EDTA. The Mn++ ion also has greater relaxivity for water protons than Mn-EDTA so that less dopant is required. The drilling mud is doped before drilling the target zone so that spurt-loss is primarily responsible for Mn++ invasion to the measurement annulus. There is, therefore, no need to pack-off and inject in the target interval as in the inject-log technique. The NMR measurement annulus is sufficiently deep that the oil saturation is not stripped below waterflood residual. The Mn concentration is chosen to separate widely the oil and brine T1 signals but not to shorten the brine signal below detectability as in the older technique. Thus, the pulsed NMR tool can measure porosity and oil saturation in the same pass without the need to log-inject-log. Moreover, the oil viscosity is determined by either relaxation or diffusion measurements. The Mn++ ion also serves as a tracer for flushing in core. The result is an accurate and virtually painless measurement of residual oil saturation. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER R ΣA LABORATORY MEASUREMENTS FOR GEOLOGICAL SAMPLES (BY PULSED METHOD) Jan A. Czubek, Krzysztof Drozdowicz, Barbara Gabañska, Andrzej Igielski, Ewa Krynicka and Urszula Woznicka Henryk Niewodniczaalski Institute of Nuclear Physics, Krakow, Poland ABSTRACT The advances in a method of measurement of the thermal neutron macroscopic absorption cross section Σa of small rock samples are presented. Theoretical principles of the method have been established in a one diffusion group approach, where thermal neutron parameters used have been averaged over the modified Maxwellian. In the consecutive measurements the investigated sample is enveloped in shells of a moderator of varying thickness. The neutron parameters of the moderator should be well known. The entire sample-moderator system is irradiated by a pulsed beam of fast neutrons. The neutrons are slowed down in the system and the dieaway rate of escaping thermal neutron flux is measured. The time decay constant versus the thickness of the moderator creates an experimental curve. The absorption cross section of the sample is found from the intersection of the experimental curve with a theoretical one which is calculated for the case, when the dynamic material buckling of the inner sample is set to zero. No reference absorption standard is required and the method is independent of the transport cross section of the measured sample. For geological formations the measured samples are composed from the crushed rock matrix saturated with a highly absorbing liquid. A calculation of the mass absorption cross section of the rock matrix requires the knowledge of the following quantitites: the decay constant of the neutron die-away curve, the neutron parameters of the external moderator (Plexigaiss), and the absorption cross section of the saturating fluid. The sample volume is about 170 ccm. The standard deviation of the measured mass absorption cross section of the rock matrix is in the range of 4 to 20 per cent of the measured value and for brines is of the order of 0.5 per cent. The highest value of the standard deviation arrives for the lowest values of the measured absorption cross section i.e. for pure silica. About one hundred rock samples of different lithology and stratigraphy have been measured until now. The Polish neutron calibration facility (at Zielona Gora) was also sampled by this method. Another faster and cheeper procedure was developed in the frame of this method for its routine applications. This method was established through a careful consideration of all experimental results obtained till now. It requires a single die-away measurement performed with only one size of the Plexiglass moderator, A detailed discussion of the procedure has allowed to apply a computer simulation of the behaviour of the statistical errors to estimate the accuracy for each fast assay of the sample. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER S COMPREHENSIVE ANALYSIS OF RUSSIAN PETROPHYSICAL MEASUREMENTS J. Craig Tingey Exxon Exploration Company, Houston, Texas Richard J. Nelson and Kent F. Newsham Amoco Production Company, Houston, Texas ABSTRACT Wireline log and core measurements form the basis for most of the current petrophvsical evaluation efforts throughout the Former Soviet Union (F.S.U.) by the Western petroleum industry. Understanding specific logging tool response characteristics and applying corrections for the effects of borehole environment can identify valid input data necessary for integrated interpretation methods. Many Russian log analysis textbooks and tool-specific chart books are available for electrical, nuclear and acoustic logs. Their application provides a wealth of insight to the methodologies and algorithms used by several generations of Russian petrophysicists Multiple techniques are needed to quantify Rt, Rxo, and Di. These techniques are based on unfocused and focused electrode and induction electrical measurements. Unfocused electrode (BKZ) methods often must be understood even when focused electrical logs are available, Forward modeling single and multiple inversion and combinations of these methods are used for lateral response interpretation. Published integrated radial geometrical and pseudogeometrical factor curves for focused electrical logs are used to generate Western-style tornado charts for 1-D invasion modeling. Common nuclear measurements include natural gamma ray, neutron-gamma (capture) neutron-neutron (thermal and epithermal), and gamma-gamma density. Newer neutron logging tools may be configured as single or dual-detector devices. Nuclear logs used for porosity determination require specific knowledge of the tool calibration methods. Environmental corrections are necessary to avoid systematic errors in the estimation of porosity. Some Russian core measurement terminology differs from Western definitions and therefore must be understood for comparison with log measurements. Depth control during core acquisition is a common problem; hence, normalization of log measurements by histogram techniques rather than regression analysis may be appropriate. The rationale for coring and core measurements varies by area and reservoir type. Examples will demonstrate that the application of appropriate Russian tool response functions and environmental corrections can yield the necessary information for informed reservoir assessment. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER T BOREHOLE IMAGE LOGS: APPLICATIONS IN FRACTURED AND THINLY BEDDED RESERVOIRS K. B. Sullivan and K. J. Schepel Exxon Exploration Company, Houston, TX, U.S.A. ABSTRACT In recent years, borehole images have enhanced our ability to assess fractured and thinly bedded reservoirs by providing a high-resolution, two-dimensional view of the borehole wall. However, image log-derived interpretations were considered largely qualitative because of the lack of calibration of the imaging tools. In recent case studies of a fractured and a thinly-bedded reservoir, comparisons of image log-based interpretations to core and production log data support more quantitative applications of image log data. In the fractured Monterey oil reservoir of the Santa Ynez Unit (SYU), offshore California, electrical borehole image data analysis has provided continuous fracture data that were previously unavailable from conventional logging tools, In at least one SYU production well, fracture density trends from borehole image analysis were well correlated to core fracture descriptions and fluid entry profiles from production log analysis. Furthermore, in several new development wells, fracture orientation histograms have revealed a preferred drainage pattern that is consistent with previous investigations of nearby fields, coastal outcrops, and the regional stress regime. Finally, interactive workstation analysis of borehole image data has yielded significantly improved structural interpretations over dipmeter-based methods. These and other applications of image-based interpretations have enhanced several aspects of SYU’s development including well planning, completions, and performance assessment efforts. In one example from the Gulf of Mexico, the integration of borehole and core image data facilitated a significantly improved assessment of a thinly-bedded, deep-water turbidite reservoir. In this assessment, a procedure was developed that incorporates electrical borehole image data, open-hole log data, and detailed core-to-log calibration techniques to obtain a net hydrocarbon sand thickness. First, individual pad conductivity values were aligned and calibrated to an open-hole, shallow-focused resistivity curve. Next, a static-processed image log was produced from the calibrated conductivity data, and the digitized core images were aligned with the log. After alignment, a threshold conductivity value was established that approximated the net sand thickness observed in core. Finally, the conductivity cutoff value was extrapolated to intervals where no core was available. Net sand thicknesses determined by this method resulted in a 20 to 30 percent increase in net reservoir volume as compared to more conventional log analysis methods. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER U THE USE OF FUZZY ARTMAP FOR LITHOFACIES CLASSIFICATION: A COMPARISON STUDY P.M. Wong Centre for Petroleum Engineering, The University of New South Wales, Sydney, Australia T.D. Gedeon School of Computer Science and Engineering, The University of New South Wales, Sydney, Australia I.J. Taggart Centre for Petroleum Engineering, The University of New South Wales, Sydney, Australia ABSTRACT It is well-known that lithofacies information is important for estimating porosity and permeability values from wireline logs at the un-cored well intervals, however, predicting lithofacies from logs is not an easy task. This paper uses a new approach called simplified Fuzzy ARTMAP (SFAM) for lithofacies classification using wireline log data. This technique is primarily based on the structure of a neural network. The objective of this study is to demonstrate via three actual field examples how the SFAM method performs in relation to the statistical discriminant analysis (SDA) and backpropagation neural network (BPNN) methods. The results of this study show that the neural network-based methods perform better than the SDA method. The SFAM gives competitive results compared to BPNN, but does not suffer the problems associated with excessive training time and prior specification of network topology. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER V PORO-ELASTICITY OF CLASTIC ROCKS: A UNIFIED MODEL Shiyu Xu and Roy White Research School of Geological & Geophysical Sciences, Birkbeck College & University College London, London ABSTRACT We describe a practical velocity model for clastic rocks developed from the Kuster and Toksoz, effective medium and Gassmann theories. In this model, the total pore space is divided into two parts: (1) pores associated with sand grains and (2) pores associated with clay minerals (including bounding water). The essential feature of the model is that the difference in pore geometry implies a difference in elastic compliance. Clay components are made much more compliant with increasing porosity than sands are. The model is able to simulate the combined effect of lithology, clay content, water-saturation and fluid type on Pwave and S-wave velocities. In contrast to common applications of Gassmann’s theory to simulate fluid relaxation, we use Kuster and Toksöz and effective medium theories to derive the elastic moduli of the dry rock frame. The predictions are therefore derived from tabulated grain matrix parameters. The model was used to study the relationship between porosity, clay content and P- and S-wave velocities. Numerical results indicate that clays affect the elastic behaviour of clastic rocks in two ways: (1) by reducing the elastic moduli of the grain matrix, and (2) by creating pores with small aspect ratios. The porosity-velocity relationship predicted for shales agrees with Wood’s suspension model rather than Wyllie’s time average equation or Raymer’s model. The model was also used to predict S-wave logs from other logs. An S-wave log can be predicted in three ways: from porosity (φ) and shale volume (Vsh). from the P-wave sonic log (DT) and Vsh , or from DT and φ. Comparisons of predictions with log and laboratory measurements demonstrated that prediction from DT and Vsh or φ was more robust than that from φ and Vsh . DT is normally less error-prone than φ or Vsh . Velocity dispersion was studied by considering the effect on the model of the relaxed and unrelaxed extremes of fluid flow. This allows the dependence of velocity dispersion in a clastic rock on porosity, clay content, and elastic properties of the grain matrix and pore fluid to be simulated. The comparisons between our predictions and laboratory and well logging data span a wide range of conditions and rock composition and demonstrate the flexibility, applicability and reliability of the model. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER W DEVELOPMENTS IN CORROSION LOGGING USING ULTRASONIC IMAGING A. J. Hayrnan, P. Parent, G. Rouault, S. Zurquiyah and P. Verges. Schlumberger Wireline and Testing K. Liang, F. E. Stanke and P. Herve Schiumberger-Doll Research ABSTRACT The ultrasonic imaging platform rotates a pulse-echo transducer immersed in the well fluid. The tool was initially applied to cement evaluation with simultaneous casing inspection, using a resonance method (200-700 kHz) to determine the cement impedance and casing thickness. The spatial resolution of these measurements is about 1.2 in. Subsequently, a focused 250/500 kHz transducer was provided for openhole imaging and detailed internal casing inspection. A focused 2-MHz transducer has now been developed for finer-resolution internal and external corrosion imaging, with a spot size of 0.11 in. The aim is to quantify small-scale defects such as pitting. The casing radius and thickness are measured from the arrival times of the internal and external casing echoes, using down-hole digital processing. The casing is examined every 2 degrees in azimuth and every 0.2 in. vertically. Laboratory results demonstrate quantification of 0.16-in. diameter flat-bottom holes. Field tests illustrate the detection of small-scale internal and external corrosion and damage and show the improved resolution of the new 2 MHz measurement compared to the resonance method. Thus three casing-inspection services are provided by the same platform: first a baseline inspection (which can be made at the same time as the cement evaluation), second a detailed internal inspection (which can be performed in the same pass as an openhole log), and third a higher-resolution log of internal and external corrosion. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER X IMPROVED POROSITY ESTIMATION IN INVADED GAS RESERVOIRS USING A PULSED NEUTRON POROSITY DEVICE C. R. Case, C. Flaum, P. D. Wraight Schlumberger-Doll Research, Ridgefield, Connecticut T. DasGupta Schiumberger Wireline & Testing, Sugar Land, Texas ABSTRACT Borehole fluid invasion into a gas reservoir significantly affects the responses of both the density and neutron devices. Specifically, invasion reduces the separation between the density and neutron porosity logs, the classic “crossover’ gas indicator. Since conventional neutron devices have a larger depth of investigation than density tools, quantification of this effect, to obtain an accurate formation porosity estimate, is difficult (Wiley and Patchett, 1994). It is possible, however, to use forward modeling (Patchett and Wiley, 1994) to correct for invasion effects, but this requires assumptions about the formation and invasion properties, e.g., fluid densities, radial extent of the invasion front, etc., which are not always accurate. With the introduction of the new APT Accelerator Porosity Tool (Scott, et al, 1994), the log analyst now has an epithermal neutron porosity device which samples a volume of the formation shallower than the traditional thermal neutron porosity tool, the CNL Compensated Neutron Log, more closely matching that of the density tool. This paper outlines a modeling study in which both the APT sonde and the Litho-Density tool responses are calculated under freshwater invasion into a gas reservoir. Because the two devices have closer depths of investigation, a simple combination of the two responses provides accurate formation porosity estimates without the need to know the details of the formation and invasion properties. These model results are based on the use of both Monte Carlo and Sensitivity Function methods. The Monte Carlo method is a well-known technique for solving nuclear transport problems, and the Sensitivity Function method is analogous to the use of geometric factors in induction modeling in which a change in formation properties is translated into a change in detected signal through the use of spatial sensitivities. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER Y CHARACTERIZATION OF A HETEROGENEOUS FORMATION USING AZIMUTHAL LOGGINGWHILE-DRILLING (LWD) MEASUREMENTS Donald S. Wolcott Colorado School of Mines Dave Shafer ARCO Alaska Inc. Amal Vittachi Schlumberger GeoQuest Inc. ABSTRACT In heterogeneous formations, the distribution of lithology. porosity, fractures and cementation are not laterally or vertically consistent. Conventional wireline and LWD tools measure rock properties in a gross sense. The wireline density tool measures the formation immediately in front of the source and receiver, while the LWI) density is taken as the highest density reading around the circumference of the borehole. In heterogeneous formations, both of these conventional density measurements can give an incomplete picture of lithology and porosity at the wellbore. With the new azimuthal LWD density tool, a more accurate representation of the reservoir is obtained by measuring lithology and porosity in four azimuthal quadrants around the wellbore. A new high resolution laterolog resistivity tool used with the new density tool provides a vertical resolution that enables measurement of fine scale areal and vertical reservoir property distributions, which more closely matches core data. As demonstrated by field examples, we compare and evaluate the response of high resolution LWD logging tools with high resolution wireline resistivity tools. wireline density tools and core data. These comparisons show that bulk density measurements between three different designs of density tools vary by more than 0.22 g/cc because they measure a different volume of rock. High resolution resistivity indicates areas where iron rich minerals are arealy extensive and where they are localized (nodules). Heterogeneities observed with the new LWD measurements can be used to improve reservoir models which leads to better reservoir performance estimation. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER Z ANISOTROPIC SHALE AND INDUCTION LOG SHOULDER BED CORRECTIONS FOR DEVIATED BOREHOLES Teruhiko Hagiwara Halliburton Energy Services ABSTRACT Shoulder bed corrections are commonly applied to induction logs to improve estimates of reservoir-bed resistivity. Conventional corrections account for finite bed thickness and borehole deviation, but not for shoulder-bed anisotropy. Modeling has shown that the application of conventional shoulder-bed corrections can lead to erroneous resistivity estimates when anisotropic shoulder beds are dipping or are penetrated by deviated boreholes. Thus. special correction charts have been developed to account for shoulder-bed anisotropy and for dipping beds or deviated boreholes. To apply the new corrections, the anisotropy parameters of the shoulder beds must be known. Three methods are available at present to estimate these parameters with resistivity tools, and each method requires two independent resistivity measurements: (1) either an electric- or induction- type resistivity tool run in two boreholes with largely different deviation angles; (2) both an electric-type and an induction-type resistivity tools run in one borehole; (3) a higher frequency induction-type resistivity tool, such as a 2 MHz logging-while-drilling (LWD) resistivity tool that provides two resistivity measurements, run in a highly deviated borehole. It has been found that when anisotropy parameters are not known, thin bed resistivity estimates can be improved with the application of conventional shoulder-bed corrections and with the assumption that the horizontal resistivity equals the shoulder-bed resistivity. However, the estimates are approximate, and the method is not applicable when reservoir beds are thin (<=5 ft) and very resistive (>=20 ohm-m). A thin bed can itself be anisotrpic if it consists of thinly laminated sand/shale sequences. Anisotropic shoulder-bed corrections have been generalized to include such situations. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER AA THE RELATIONSHIP BETWEEN POROSITY, MINERALOGY, AND EFFECTIVE STRESS IN GRANULAR SEDIMENTARY ROCKS Phil Holbrook Sperry-Sun Drilling Services, Houston, TX ABSTRACT Sedimentary rocks are constantly under high stress in the subsurface. Within a few years there is a pseudo-static equilibrium between effective stress and the load bearing capacity of the grain matrix framework. Rock porosity remains in pseudo-static equilibrium with effective stress for millions of years until disturbed by rapid reservoir fluid production. Sedimentary rocks are composed principally of single mineral grains. Five minerals: quartz, clay, calcite, anhydrite, and halite compose over 90% of all sedimentary rocks. The physical properties of the mineral grains which compose the sedimentary rock determine grain matrix framework compaction resistance. Grain matrix framework strain is inversely proportional to solidity, the complement of porosity. The porosity of all granular sedimentary rocks is quantitatively related to effective stress by two mineralogic equilibrium compactional parameters in a power law loading limb effective stress-strain relationship. The uppermost compaction limit σmax is a power law zero porosity intercept and a is a compaction slope. The compaction limit (σmax) is sensibly related to other mineral physical properties. The σmax intercept is directly proportional to mineral hardness and inversely proportional to mineral solubility. The σmax of a mixed mineral sedimentary rock is the logarithmically weighted average of the individual mineral σmax coefficients which compose that rock. The observed porosity-mineralogy-effective stress relationships are demonstrated through presentation of several binary mineral-porosity crossplots of subsurface and laboratory petrophysical data at different levels of effective stress. These binary mineralogy-porosity relationships are quantitatively related to average constituent mineralogic physical properties and effective stress to form a general mineralogic compactional relationship. This porosity-mineralogy-effective stress relationship has been applied and tested in over one hundred wells worldwide through continuous comparison of porosity to overburden and pore pressure. This pseudo-static mechanical compactional relationship can also be applied to oil and gas exploration porosity risk assessment where rock samples and logs are not available. Continuous calculated mineralogic - effective stress compactional porosity curves can also support conventional log analysis. When porosity sensor measurements are missing, or a sensor reading is suspect, mineralogic compaction equilibrium porosity can be useful to evaluate petrophysical porosity sensor readings. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER BB AZIMUTHAL POROSITY WHILE DRILLING J. Holenka, D. Best, M. Evans, P. Kurkoski, and W. Sloan Anadrill / Schlumberger ABSTRACT For the first time in formation evaluation, measurement of azimuthal porosity and lithology around the borehole is available with a logging while drilling tool. The ADN (Azimuthal Density Neutron) tool measures the bulk density, photoelectric factor and neutron porosity by quadrants around the borehole. It also measures tool stand-off in each of the quadrants by means of an ultrasonic sensor. The quadrants, generated by a magnetometer, are oriented with the gravity vector. Presently the tool divides the borehole into four quadrants: bottom, up, left and right. Measurement of the formation parameters azimuthally around the borehole provides a quantitative evaluation of the formation heterogeneity in terms of porosity and lithology. This directional information, now available to the reservoir engineer, will allow better evaluation and understanding of complex and heterogeneous formations. This is an improvement over conventional nuclear tools which provide measurements either averaged around the borehole or in only one direction. The measurement per quadrant also improves the tool accuracy in difficult hole conditions, especially in deviated and horizontal wells, where the tool generally makes good contact with the formation on the bottom of the hole. Density, photoelectric factor and neutron porosity computed from the bottom quadrant have minimal tool standoff compared with the other quadrants. This allows two major improvements: first, the tool can be run in enlarged and washed-out boreholes and still provide an accurate measurement; secondly, it can be used with or without a stabilizer, giving more flexibility for the driller to define the bottom hole assembly. Measurements of the minimum, maximum and average standoffs by quadrant provide borehole shape, borehole rugosity and hole size, They also provide neutron porosity borehole correction and quality control for the density measurement. The paper describes the method of the measurement and shows log examples illustrating the tool features. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER CC DIP ESTIMATION FROM AZIMUTHAL LATEROLOG TOOLS Ollivier Faivre and Gerard Catala Schlumberger Wireline & Testing ABSTRACT A new generation of Laterolog tools provides azimuthal resistivity measurements around the borehole. The response of these mandrel tools in dipping beds has been modeled in order to evaluate the viability of computing formation dip from the output azimuthal images. The modeling results demonstrate that dip information is present on the azimuthal resistivity curves. They also show that structural formation dips can be computed with good accuracy although the Azimuthal Laterolog has a much coarser vertical resolution than the standard dipmeter tool. They also show that a dip correction is necessary in order to compensate for the fact that the electrical currents emitted by the device penetrate the formation and make the bore hole diameter appear slightly larger than actual when it comes to computing the dip plane orientation, It is shown that, the classical incremental electrical diameter concept used for all traditional dipmeters is also applicable to the Azimuthal Laterolog although there is no direct link to the depth of investigation of the tool. This important parameter depends mainly on the pre-processing applied to the input curves and the type of algorithm used to compute dip; it is relatively insensitive to formation geometry and to bed resistivity contrast. As a by-product of the dip computation, a better estimation of the formation resistivity is obtained by stacking the azimuthal resistivities along the apparent dip plane. This technique greatly improves the evaluation of laminated reservoirs when apparent dips are greater than 30 degrees. A software package using cross-correlation techniques, or local event detection, has been developed to compute dips for any tool yielding azimuthal data, including the Azimuthal Laterolog. Field examples comparing Azimuthal Laterolog dips with standard Dipmeters confirm the potential of the Azimuthal Laterolog for dip determination. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER DD IMPROVED ENVIRONMENTAL CORRECTIONS FOR MRIL PULSED NMR LOGS RUN IN HIGHSALINITY BOREHOLES D. Mardon, M. G. Prammer, Z. Taicher, It. N. Chandler, and G. R Coates NUMAR Corporation ABSTRACT Introduction of the Series C Magnetic Resonance Imaging Log (MRIL) has significantly reduced the limitations on logging speed and mud resistivity that existed with the earlier Series B tool. However, the high salinity muds in which the tool can now be run present two environmental effects that could have an adverse effect on log quality: I) errors in the internal tool calibration due attenuation of transmitted and received signals in the borehole and 2) an interfering NMR signal from sodium nuclei in the mud. We present an analytical solution for the electromagnetic field propagation problem for the tool which demonstrates that attenuation in both the borehole and the formation have a negligible effect on the accuracy of the calibration. We also present new experimental data for sodium NMR in a variety of commercial salt-saturated muds which shows that 1) the amplitude of the sodium signal is accurately predictable from routine mud composition data, and 2) T2 of sodium in muds is very short (~5 msec) and is nearly invariant with respect to changes in temperature, mud composition, and rheology. These results form the basis of a new software correction for sodium which we illustrate with a log example. Operational strategies that are effective at eliminating the interference also are described. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER EE EXPERIMENTAL EVIDENCE OF RADIAL EFFECTS ON CENTRIFUGE CAPILLARY PRESSURE CURVES. M. Fleury and P. Forbes Institut Français du Pétrole ABSTRACT With the aim of improving the interpretation of centrifuge data, several authors have recently pointed out the theoretical importance of radial effects when calculating centrifuge capillary pressure curves from average saturation data. Capillary pressure curves should shift to higher saturation values and the magnitude of this shift depends on the sample diameter and radius of rotation. In particular for standard high speed centrifuges, a shift of up to 5 % is predicted. We performed centrifuge experiments on outcrop samples using two different geometries to clearly demonstrate the radial effect. The samples were either placed at a large or a small radius of rotation to generate low and high radial effects respectively. For a given sample the two centrifuge data sets (avenge saturation vs speed of rotation) were analysed using a recent and improved interpretation technique available in literature. Two different capillary pressure curves were obtained for the same sample, although the two solutions can reproduce the two data sets after integration (self-consistent procedure). The saturation shift of the two curves due to radial effect was as large as 6 saturation units near the irreducible water saturation. When centrifuge data are interpreted using a new correction accounting for radial effects, an agreement is found between the centrifuge capillary pressure curves resulting from the two kinds of experiment. Using such a correction, a good agreement is also found between centrifuge and porous plate drainage capillary pressure curves obtained from the same samples. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER FF APPLICATION OF A MULTILAYER NEURAL NETWORK AND STATISTICAL TECHNIQUES IN FORMATION CHARACTERISATION C.A. Gonçalves*, P.K. Harvey, and M.A. Lovell Borehole Research, Dept. of Geology, University of Leicester, UK * On leave from NPGP/Federal University of Path - Brazil. ABSTRACT The characterisation of formation heterogeneities requires a multidisciplinary study of data acquired using a large number of numerical geophysical and geological measurements and a rigorous evaluation of the precision and accuracy of the data. Another essential aspect of the appraisal of any measurement is the quality assessment and quality control of the data. In this work multivariate statistical techniques and an artificial neural network (ANN) are used to identify heterogeneities in complex formations and to evaluate the boundaries they generate. The precision and accuracy of the data from different sources are very important and are considered here by using sample support in the integration of measurements at different scales. We use examples from two wells to show the differences in characterisation obtained with each technique. Multivariate statistical analyses are initially used to group the petrophysical, geophysical and geological parameters extracted from the downhole measurements into distinct geologically definable zones. This technique has the advantage of being quasi-independent of any predetermined ideas we have about the whole data set, and has shown to be very reliable in formation characterisation. Thus the result obtained here is a useful basis for comparison with that obtained from the neural network. An artificial neural network is used to characterise the different lithology sequences present in each well. Neural networks are relatively new tools and have proved very useful in applications where conventional computing methods are inadequate. Another application is the possibility of determining quantitative petrophysical parameters from wireline logs and core data in uncored intervals. We also show the ability of the neural network to construct petrophysical logs. The results are presented as a comparison between the two techniques. We show that both methods are very encouraging. When comparing the ANN derived petrophysical parameter logs with actual core measurements we see a good match. Low quality petrophysical measurements can be determined by a mismatch between the responses. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER GG PERMEABILITY FROM MAGNETIC RESONANCE IMAGING LOGS D,W. Curwen Suncor Inc., Resources Group, Calgary, Alberta C. Molaro NUMAR Corporation, Calgary, Alberta ABSTRACT Classic porosity - permeability relationships were developed based on grain size/bound water volume relationships. Because early Nuclear Magnetic Logs indicated relative volumes of free water, and bound water could be calculated, initial permeability relationships were developed using variations of the traditional approach. Newer Resonance tools provide more insight into pore structure. Therefore, better prediction of permeability is possible. The Montney formation is a moderately deep marine deposit of Triassic age. It consists of very fine grained, feldspathic sands and silts with dolomite cementing. Because of the fine grained nature of the rock, permeability is the critical unknown. Extensive core analysis was done on a Suncor well in the Grande Prairie area of Northwestern Alberta. This provided a good dataset from which comparisons could be made between permeability derived from the Nuclear Magnetic Resonance Log and core permeabilities. The more traditional relationships using Free Fluid Index and Total Porosity can be made to fit in one zone, but they give poor answers in another zone. Relationships that account for T2 distributions give a better fit. In fact, a binned T2 spectral approach indicated a strong correlation with formation core measurements in both water saturated and hydrocarbon bearing intervals. This relationship is expected intuitively when comparisons are made between T2 distribution plots and pore throat distributions from mercury injection porosimetry. A statistical approach to T2 analysis should yield relationships that require less local calibration. We conclude that relationships using T2 distributions show the greatest promise. However, variables not accounted for in the data may be affecting some results. The geological environment in this study is relatively consistent. Studies need to be done in other depositional environments with diverse porosity types to check the robustness of the relationships. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER HH THE PETROPHYSICS OF ELECTRICALLY ANISOTROPIC RESERVOIRS J. D. Klein and P. R. Martin ARCO Exploration and Production Technology, Plano, TX D. F. Allen Schlumberger Well Services, Sugar Land, TX ABSTRACT Reservoirs containing thin layers with differing resistivities are electrically anisotropic when considered as a bulk medium. The degree of anisotropy depends on the differences in resistivity between the individual thin beds or laminations, which depends in part on their fluid saturations. Thus in a hydrocarbon reservoir the resistivities of the individual thin beds depend upon capillarity (often controlled by grain size) and the height above the free water level. A water wet formation exhibiting large variability in gain size between the laminations could be highly anisotropic when oil saturated, but nearly isotropic in the water leg. Laboratory measurements and petrophysical models support the above conclusions. The first case simulates the Asands of the Kuparuk reservoir on the north slope of Alaska. The model, based on an FMS interpretation, contains layers of very low permeability mudstone along with layers of permeable sandstone with varying clay content. A second case simulates Gulf Coast sands with clean sand layers of variable permeability and grain size. The calculations predict little anisotropy if the model formations are in the water leg, and large anisotropy when they are in the hydrocarbon column. Pronounced electrical anisotropy, in the presence of porous sediments, is a good indicator of hydrocarbon pay, particularly thin-bedded, low resistivity pay. Unfortunately, at the present time reliable anisotropy measurements can be obtained only with 2 MHz resistivity tools in nearly horizontal wells, thus eliminating most exploration applications of this phenomenon. In addition to predicting electrical properties our models also calculate permeability. An important conclusion from our research is the seemingly paradoxical conclusion that the parallel permeability in an anisotropic formation is proportional to the perpendicular resistivity, not the parallel resistivity. This result provides insight into the applicability of resistivity-based permeability models, and explains why they sometimes fail, since these models usually employ parallel resistivity as measured in vertical wells. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER II INFLUENCE OF FLUID DISTRIBUTION UPON ELECTRICAL RESISTIVITY OF PARTIALLY SATURATED MEDIA Carlos A. Grattoni and Richard A. Dawe Department Mineral Resources Engineering, Imperial College of Science, Technology and Medicine London, U.K. ABSTRACT Water saturation is calculated from well logs using methods based on Archie’s equations. According to his second equation the resistivity will only depend on the water saturation. In laboratory tests to obtain the saturation exponent uniform saturation along the core is assumed but CAT scanning shows that this uniform saturation is seldom achieved. Some of the discrepancies in the literature on resistivity values must be due to this. This paper presents an experimental study of the influence of fluid distribution upon the electrical resistivity of partially saturated porous media. Visual, quasi 2-dimensional models have been used which allow the observation of the phase distribution whilst the resistivity is being measured. A new technique has been developed to produce two liquid phases In-situ with uniform distribution. We have demonstrated that even if the water saturation has a macroscopic uniform distribution within the porous medium it is not always so at the pore level and depends on pore size distribution, orientation of the pores and fluid characteristics. Since the electrical current is transported by the ions in the water the fluid distribution at the pore level will influence the resistivity, In this work the resistivity components for water-wet and oil-wet porous media at the pore level have been identified and their influence on resistivity have been verified through visual observations. We conclude that the influence on resistivity of components (such as partially invaded pores/throats, water films, etc.) plus the wettability of the rock and the method to obtain the saturation in the laboratory all significantly affect the saturation exponent th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER JJ MAGNETIC SUSCEPTIBILITY CONTRAST EFFECTS ON NMR T2 LOGGING G. A. LaTorraca and K. J. Dunn Chevron Petroleum Technology Company D. J. Bergman Tel Aviv University ABSTRACT When a rock sample or formation is subjected to a static magnetic field, magnetic susceptibility contrasts between rock grains and pore fluids induce local magnetic field heterogeneities. These local field fluctuations cause increases in the transverse (T2) relaxation rate. Under specific conditions, the increase in the relaxation rate is proportional to the squares of the averaged field strength, the susceptibility contrast, and the pore dimension. The rate increase is also a function of τ (half the echo spacing), and is proportional to l-(tanh λτ)/λτ. Here λ is an eigenvalue of the diffusion equation which is proportional to the reciprocal of the pore dimension squared for a periodic porous medium, or can be regarded as the dominant relaxation rate in a disordered system. Laboratory NMR core measurements indicate that an apparent l/λ can be determined experimentally by varying the echo spacing, and can be correlated to permeability. Such a correlation may not always be as sensitive as T2G/F, where T2G is the logarithmic average of T2 and F is the formation factor, but it offers additional petrophysical information for formation evaluation. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER KK DERIVATION OF A CORE PERMEABILITY LOG AND EXTRAPOLATION TO UNCORED INTERVALS H. N. Greder, E. A. Mus, A. M. Veillerette, and F. M. Pellerin Elf Aquitaine Production, Pau, France) ABSTRACT Conventional methods to estimate permeability from wireline log responses often consist of building models between regularly spaced core plug measurements and logs without paying much attention to the various investigated scales or to core plug sampling. In order to quantify the influence of these elements, a detailed method of permeability prediction based on a core permeability log was developed and applied to the case of a mixed lithology oil reservoir. The method consists of three stages. Firstly, a continuous high resolution core permeability log representative of in-situ conditions is derived in cored intervals using ONLY core data (conventional and special plug measurements, probe measurements and core lithological information). At this stage conventional permeability measurements need to be validated and corrected, core lithological descriptions are used to restore permeability values in cored intervals for which laboratory measurements could not be performed. Secondly, the core permeability log is normalized in order to obtain a vertical resolution close to that of conventional logs - commonly used for permeability prediction. Thirdly, two different prediction models are constructed in the cored intervals using the normalized core log as the target parameter and conventional logs as prediction parameters. The models, one based on multiple regression techniques and the other on discriminant analysis, were found to be more predictive than the conventional models which use regularly spaced core plug measurements, The second modeling method was found to be more predictive than the multiple regression method. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER LL THE EFFECTS OF PORE STRUCTURE ON ELECTRICAL PROPERTIES OF CORES SAMPLES FROM VARIOUS SANDSTONE RESERVOIRS IN TARIM BASIN Zhi Qiang Mao, Cheng Guang Zhang, Chun Zeng Lin, Jian Ouyang, Qing Wang, and Chun Juan Yan Exploration & Development Research Center, Tarim Petroleum E & D Bureau, CNPC, China ABSTRACT Based on the resistivity measurements, mercury-injection capillary pressure data, casting thin-section analysis, grain size analysis, and other lab measurements on core samples, the effects of pore structure on resistivity characteristics of core samples have been studied for several sandstone reservoirs of Tarim basin, northwest China. It was found that in double logarithmic coordinates, the relationship of formation factor and porosity is a second order function and no longer the linear formula proposed by Archie in 1940’s. The function can be simplified as two linear formulae. They represent different electrical properties which relate to the two predominant pore structures of sandstone-‘network-like’ and ‘pseudo capillary bundle like’. It was also concluded in the paper that the electrical characteristics should be same if the pore structure of the reservoirs is not significantly different. Both the cementation factor m and the constant a are controlled by pore structure, and they relate to each other among the various sandstones with different pore structures. The verified prediction of values of m and a showed that porosity, permeability, and their distributions can be used to characterize qualitatively the pore structure and to determine quantitatively the values of in and a. In this paper, it was also demonstrated that it is the change of pore structure and not the shale effect on resistivity measurements results in the downward curvature of formation factor-porosity relation in low porosity range. The measurements of saturation exponent n showed variety. Different results are obtained when different non-wetting fluids are used. The n value measured by semi-permeable diaphragm is stable and higher than that by conventional techniques. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER MM ELECTRICAL CONDUCTIVITIES IN OIL-BEARING SHALY SAND ACCURATELY DESCRIBED WITH THE SATORI SATURATION MODEL A. de Kuijper, R.K.J, Sandor, J.P. Hofman, J.M.V.A. Koelman, P. Hofstra and J.A. de Waal Shell Research B.V., Koninklijke/Shell Exploratie en Productie laboratorium, Rijswijk, The Netherlands ABSTRACT We present a new class of saturation models based on a Symmetrical Anisotropic Theory Of Resistivity Interpretation that will lead to more accurate hydrocarbon saturation predictions. With SATORI models, the bulk electrical properties of rock can be accurately calculated from the volume fraction, shape, conductivity and connectivity of the minerals and fluids in the rock. The models allow components to remain connected down to low volume fractions and can handle anisotropy in shape and conductivity of the various constituents. There are no adhoc parameters, and the calculated conductivity is independent of mixing order. As a first implementation, we narrowed down the general SATORI framework to describe the low-frequency electrical behaviour of isotropic shaly sands. Honouring the underlying electrochemistry, we obtain a shaly-sand model with three independent parameters: shale conductivity, shale volume and pore-space connectivity. The effects of clay-bound water are accounted for with the equation of Hill, Klein and Shirley; effects of wettability and surface roughness can also be included, albeit at the cost of an additional parameter. We tested this SATORI shaly sand model against: 1) results from numerical random walk calculations on model systems, 2) shaly sand data from the open literature, and 3) new electrical conductivity data of real and artificial shaly sand samples. In all cases, the electrical conductivity and membrane potential as a function of porosity (0.10.. 0.90), salinity (1 .. 300 g NaCl/l), temperature (22..200 °C), and hydrocarbon saturation (0.. 0.85) were calculated with a maximum error of a few percent. This is a major improvement, in particular at low salinities and for high shale fractions. Present computations suggest that hydrocarbon saturations calculated on the basis of SATORI can differ by as much as 5 to 10 saturation percent from those computed on the basis of existing empirical models. SATORI can be incorporated into a saturation model of which the input consists of basic petrophysical measurements; no additional data is required. The parameters in the model are derived from inversion on the measured data. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER NN PERMEABILITY PREDICTION FROM WELL LOGS – A CASE STUDY IN GERMAN ROTLIEGEND SANDSTONES R. van den Bosch Mobil Erdgas-Erdoel GmbH, Celle, Germany A. Fulop RWE-DEA AG, Hamburg, Germany V. Pistre Schlumberger GeoQuest, Hannover, Germany ABSTRACT Permeability prediction in German gas fields has proved difficult in lithologically complex clastic rocks where authigenic illite drastically reduces permeability in relatively high porosity reservoirs. Best results today have been obtained with core laboratory measurements corrected for in-situ conditions and complemented with well test results. However, the cost of coring and testing a complete reservoir section is prohibitive. This study compares two conventional interpretation techniques with permeability predictions derived from Stoneley waves to obtain reliable permeability prediction over the whole reservoir: The two conventional interpretation methods are: I) porosity (O)-permeability (K) correlation II) permeability computed from porosity and grain surface area derived from Thorium logs Empirical φ-K correlations allow permeability estimates which are locally reliable. However, due to the problem complexity, a single variable such as porosity is insufficiently accurate to describe permeability variations. Consequently, the applicability of each regression equation is restricted to a certain lithofacies with more or less homogeneous pore structure type. The second approach is based on a generalized Kozeny-Carman type equation. This theoretically founded methodology overcomes almost all restrictions inherent to regression calculations and has proven a useful, well-behaved permeability predictor. Permeability indications using Stoneley waves are based on the theoretical relationship that permeability changes in a fluid-saturated formation have a measurable effect on low frequency sonic waves. The methods investigated deduce permeability indications from changes in: a) Stoneley slowness b) Stoneley energy c) normalized Stoneley energy d) Stoneley wave attenuation over a frequency range The two conventional correlation techniques have the advantage of providing a quantitative permeability curve whereas the Stoneley wave methods give an indicator that needs additional calibration. All methods were applied in a vertical and a horizontal well in a deep, tight gas field. Results are compared with well performance. The limitations of the Stoneley methods due to low permeability and gas saturation are discussed. An outlook and recommendations for further development work are given. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER OO FORMATION RESISTIVITY MEASUREMENTS THROUGH METAL CASING AT THE MWX-2 WELL IN RIFLE, COLORADO W.B. Vail, S.T. Momii ParaMagnetic Logging, Inc. H. Haines Gas Research Institute J.F, Gould, Jr. Mobil Exploration & Producing U.S. W.D. Kennedy Mobil Research and Development Corporation ABSTRACT A second logging instrument was designed and built by ParaMagnetic Logging, Inc. (PML) to measure the resistivity of geological formations in steel cased wells. In principle, the new logging tool functions like a Laterolog 3 or guarded-electrode tool, but performs measurements in a cased well. Previous measurements made in a research well with a first logging instrument demonstrated the feasibility of the concept. The second logging instrument has demonstrated improved vertical resolution, faster (although still stationary) measurements, and more accurate results than with the first logging instrument. The data obtained to date from the first and second logging instruments show that the methods of analysis published in earlier patents produce logs that agree fairly well with both induction and laterolog open-hole measurements. The data interpretation method uses the single Calibration Constant K and simple mathematics to generate wellsite logs. An approximate algebraic formula shows that K is largely a function of electrode spacing and casing radius. The favorable comparison of these cased-hole resistivity logs with standard open-hole resistivity logs suggests an application of the method to monitor water saturation behind pipe. Two successful tests were performed using this second logging instrument. The first test was performed in a shallow test well drilled in Woodinville, Washington specifically for the purpose of verifying the measurement and comparing the data to open-hole Dual Laterolog and Dual Induction logs. The second test was in the deep MWX-2 well located near Rifle, Colorado. In both wells the cased-hole resistivity compares favorably with open-hole deep induction logs. In many zones the cased-hole apparent resistivity seems to be a better approximation of true resistivity than the deep induction log. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER PP THROUGH-CASING RESISTIVITY: 2-D AND 3-D DISTORTIONS AND CORRECTION TECHNIQUES B. Sh. Singer, O. Fanini, K.-M. Strack, L.A. Tabarovsky Western Atlas Logging Services, Houston, TX X. Zhang Electrical Engineering Dept., University of Houston, Houston, TX ABSTRACT During the past years, through-casing resistivity measurements confirmed the original concept proposed by Alpin (1939) and recently revived and improved by Kaufman (1990) and Vail (1991). A commercial tool can be used for monitoring producing wells. waterflood control, and searching for bypassed hydrocarbons in abandoned and active wells. The wide possible market for the arising technology has inspired a number of authors to investigate its opportunities and limitations. There are several difficulties in the through-casing resistivity technology. The signal level is typically in the nanovolt range. Difficulties in the registration of such signals using the DC-mode made it necessary to apply ACcurrent injection. The operating frequency of the tool must be low enough to guarantee penetration of the field through the casing into the formation. For most of the casings used by the oil industry, this condition is satisfied at the frequencies of about 1Hz. The information on the resistivity of the formation is represented by the second spacial derivative of the casing voltage that is generally small. This makes the evaluation of possible distortions arising from different kinds of casing imperfections important. These imperfections include casing collars, corrosion, perforations, etc. The fact that only low frequencies can be used for the measurement allows us to propose fast and accurate modeling algorithms. A number of different casing imperfections have been considered. Numerical evaluation shows that the distorting effect of casing imperfections is quite moderate. Typically, it does not exceed 10-20% of the measured apparent resistivity, The vertical resolution of the measurements is limited by the spacing between the voltage-sensing electrodes. It can deteriorate due to the cement sheath that always exists around the casing. To improve the readings, a simplifled inversion algorithm has been developed. The most direct application of the algorithm is the monitoring of the pay zone resistivity. A significant distortion of the measurement occurs near the casing bottom. This distortion can reach 50% of the signal. A modification of the above-mentioned inversion algorithm can be used to make the necessary corrections. Overall, through-casing resistivity measurements require an elaborate interpretation technique to improve the accuracy of the estimated formation resistivity. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER QQ NON-CONVENTIONAL APPLICATIONS OF THROUGH-TUBING CARBON-OXYGEN LOGGING TOOLS B. Roscoe, J. Grau Schlumberger Doll Research C. Cao Minh Schlumberger Middle East D. Freeman Saudi Aramco ABSTRACT Through-tubing carbon-oxygen (C/O) tools are commonly used to determine water saturation behind casing in fresh or unknown salinity formation water. In addition to this conventional application, C/O tools offer other useful data that have not, until now, received enough attention. This paper will demonstrate the use of some of this information: 1) to improve the estimation of inelastic yields of elements such as silicon, calcium, iron and magnesium. A new technique determines the abundance of magnesium that may be used to improve the characterization of dolomite in carbonates. These inelastic yields are acquired at the same time as carbon and oxygen yields, but have previously been ignored. 2) to determine both reservoir water volume and borehole holdup from shut-in and flowing passes. Borehole holdup determination has always been difficult in deviated and horizontal wells, in multiple tubing string completions and in high flow rate wells. The technique may also ease difficult coiled-tubing operations by allowing a shorter toolstring combination, saving rig time and reducing hazards. 3) to evaluate Sigma (Σ) from through-tubing C/O tools. The full characterization of the time-decay spectra acquired with C/O tools allows the recording of a quality Σ comparable to that of existing pulsed-neutron tools. Combined C/O and Σ analysis allows the reservoir engineer to distinguish injected water from connate water of different salinity. These non-conventional applications of through-tubing carbon-oxygen logging tools offer a reduction of logging time, while providing the data needed for reservoir evaluation in difficult environments. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER RR A NEW QUANTITATIVE METHODOLOGY FOR GAS SATURATION DETERMINATION AND ITS APPLICATIONS IN PRUDHOE BAY Moustafa E. Oraby ARCO Exploration & Production Technology Jerry Brady ARCO Alaska Incorporation Thu-Thuy Dang ARCO Exploration & Production Technology ABSTRACT Gas saturation, Sg, is one of the most difficult petrophysical parameters to interpret. Also, most of the well known gas detection methods e.g. neutron/density and neutron/sonic are qualitative and can not be easily used to quantify the gas saturation. In this paper, a new methodology is introduced that can be used to quantitatively provide gas saturation in gas or gas/oil reservoirs using the neutron porosity. The new methodology uses the open-hole calculated lithology for a reservoir to forward model the neutron tool response for different assumed gas saturation varying from 0-100%. The number of the forward thermal neutron responses generated at each depth depends on the gas saturation increment. The forward modeling is based on calculating the formation neutron parameters using any of the available neutron diffusion codes which are always fast in computing time and reliable in their results. The calculated formation neutron parameters are then transformed into neutron porosity responses using transformations that most of the service companies have already derived for their tools. To calculate gas saturation, the forward thermal neutron responses are then overlaid with the actual tool response and the intersection at each depth determines the gas saturation. The accuracy of that methodology depends on the accuracy of the open hole formation lithology that is used to forward model the thermal neutron responses, the accuracy of the neutron parameters to the tool response transformation, the quality of the neutron tool data acquisition, and the gas increment used to generate the forward modeled neutron responses. This methodology has been tested in many wells in Prudhoe Bay. The results are more accurate and representative than the conventional gas saturation models. Results of some of these applications and the accuracy of the methodology are presented here. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER SS USING PRODUCTION-LOG DATA FROM HORIZONTAL WELLS G. Brown BP Exploration H. Rees and M. Toro Edinburgh Petroleum Services ABSTRACT The trend for drilling horizontal wells continues but the production-log data from such wells has been presumed unusable in may cases. The possibility of poor log quality (in a conventional sense) and the cost of re-entry into many wells has resulted in the curtailing of some production logging campaigns and only a limited number of sensors being run compared with production-logs in more conventional wells. A study of a large number of horizontal and very-high-deviation wells from the United Kingdom and Alaska was carried out to determine the effects that different completions, geology, deviations, sensor types and production regimes have on a number of conventional production-log sensors. The aim was to identifv any response patterns in the data, to evaluate and define the limitations of conventional sensors in such conditions and to investigate the possibility of improvement to existing techniques of measurement. New and existing ideas were used in analysing the data to determine if the logs would yield more information about flow-profiles and fluid entry-points. Deconvolution and unconventional graphics techniques really do benefit the analyst in the course of evaluating horizontal log data and help to identify zones of static fluid, water production, fluid acceleration gas entry, bore-hole and completion configuration and general log quality control. The ability to predict the type of log responses for a given set of circumstances and the knowledge that seemingly poor data call still be used will assist in the justification and design of a production-log campaign and could influence the type of completion chosen for a well. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER TT STRUCTURAL INTERPRETATION OF RESISTIVITY-AT-THE-BIT IMAGES J. R. Lovell Anadril R. A. Young Schlumberger Wireline & Testing R. A. Rosthal Anadril L. Buffington GeoQuest C. L. Arceneaux, Jr. Anadril ABSTRACT Fullbore resistivity imaging is now available from a logging while drilling (LWD) tool. The RAB (Resistivity-Atthe-Bit) tool is a unique LWD device which acquires fullbore, quantitative, azimuthal resistivity images of the formation while drilling. RAB field logs from offshore California and the Gulf of Mexico, together with a finite element model study, have been analyzed to determine the type and scale of geologic bedding features that can be interpreted from RAB images and dips. Large-scale structural bedding is readily recognized in RAB images as are other geologic events such as faults and unconformities. Hand-picked dips from RAB images are in excellent agreement with those from wireline image logs. Good agreement is also obtained from RAB data processed through commercially available software. Large fractures or swarms of fractures can be identified from RAB images. Fractures of small and medium aperture scales normally seen in reservoirs were more difficult to resolve. For such fractured zones, RAB logs can be used to identify intervals where higher resolution wireline imaging logs are needed to quantify fractures, vugs and fine scale stratigraphic bedding features. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER UU ACCURATE MWD DENSITY MEASUREMENTS WITH VERY LARGE STANDOFFS R. L. Spross, T. M. Burnett, C. A. Golla, and C. Huiszoon Sperty-Sun Drilling Services ABSTRACT The most critical variable in MWD density measurements is standoff. Standoff is usually eliminated as a variable in wireline logging by pressing the tool against the formation with a backup arm. However, it is unavoidable in MWD measurements, although its effects may be substantially reduced through use of a stabilizer blade. Accurate corrections for small or moderate standoffs may be obtained by comparing long and short spaced detector responses. if larger standoffs occur, either because of washout or though the deliberate use of drilling friendly ‘undergauge” tools, corrections obtained from the comparison method become increasingly less accurate. Thus, alternative methodologies are required to obtain accurate density values in enlarged boreholes. To correct MWD density measurements for large standoffs a new statistical method has been developed which uses a time series of many short duration counting samples. Through a statistical analysis of the time series, the method first determines if additional processing for excessive standoff is required. If it is, the short duration samples for both the long and short spaced detectors are sorted according to their count rates, and an improved avenge count rate for each detector is obtained through rejection of those samples taken at larger than acceptable standoffs. The improved count rates are fed back into the correction algorithm to obtain an improved corrected density value. A demonstration test of the method conducted in a water filled, 12 inch diameter limestone test formation (maximum standoff over 3 inches) resulted in improvement from a measurement error of more than 0.5 g/cm3 without the method, to an error of less than 0.03 g/ cm3 with the method invoked. The correction method requires no well-specific input, and is implemented completely in downhole software. The tool must be eccentered and rotating so that minimum standoff is experienced during part of the tool rotation. Since changes in count rate indicate variation in standoff, the effectiveness of the method requires a minimum amount of mud to formation density contrast. However, in situations where the lack of contrast reduces the effectiveness of the method, the algorithm is capable of determining from sample to sample the parameters which optimize the sorting algorithm. Implementation of the method in the downhole software will be described, and a field example employing the method will be presented. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER VV COMPRESSIONAL SLOWNESS MEASUREMENTS WHILE DRILLING John Minear, Robert Birchak, Carl Robbins, Eugene Linyaev, Bruce Mackie, David Young and Robert Malloy Halliburton Energy Services ABSTRACT Research and design of tools for measuring formation compressional wave slowness while drilling have been carried out by a few companies over the past few years. This paper briefly describes the development of one such tool and presents examples of compressional slowness obtained with the tool. Wireline measurements obtained in the same wells are also presented with the logging while drilling (LWD) measurements. Drilling-induced noise decreases with frequency so that transducers can be used that operate in the frequency range of wireline acoustic tools. Tool-borne noise generated by the transmitter and by interaction of the drillstring and the formation must be reduced by attenuating the noise in the drill collar and by isolating the transducers from the tool. Prototype LWD sonic tools have been used to log over 20,000 feet of well during 500 hours of drilling and reaming. Waveforms have been obtained from a four-element receiver array and two transmitters in a reversed profile borehole-compensation configuration. Formation compressional slowness has been obtained from semblance-based processing across the receiver array. A semblance-guided approach has been developed to compute transit times to two receivers from each of the two transmitters to yield a borehole-compensated slowness. Three example logs are presented that compare LWD and wireline slowness for slowness values ranging from 60 µs/ft to 170 µs/ft Several techniques have been applied to reduce non-coherent noise and tool-mode interference. An additional acoustic transducer for measuring the distance from the receivers to the wellbore wall is incorporated into the tool. This standoff measurement may be used for aligning waveforms before stacking and is particularly valuable in rapid-fire mode for acquiring waveform data. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER WW IMPROVING DRILLING PERFORMANCE IN THE CUSIANA AND CUPIAGUA FIELDS BASED ON MEASUREMENT AND INTERPRETATION OF HOLE DETERIORATION USING TIME-LAPSED 2 MHZ ELECTROMAGNETIC CALIPER DEVICES Herve De Naurois and Nigel C. Last BP Exploration Colombia F.J. Husband and Marcus Palstra Anadrill Schiumberger ABSTRACT Ongoing appraisal and development drilling of the Cusiana and Cupiagua fields in the eastern Andes foothills of Colombia has historically met with difficult drilling conditions including wellbore instability, tectonic stress and highly deteriorated boreholes. The search to characterize formation dynamics and employ measurements to improve drilling performance has been compromised by the inability to acquire sufficient volumes of wireline conveyed logging data due to tough hole conditions. Time-lapsed logging measurements now made while drilling and reaming with 2 MHz electromagnetic devices provide a view of the evolution of borehole geometry with time, and evidence of microfractured intermediate formations leading to an accepted model of rock fragility as a major cause of hole failure during the drilling process. Phase derived caliper measurements of borehole compensated 2MHZ resistivity tools are validated with 4-arm wireline caliper devices and core analysis confirming the accuracy of average hole size measurements and the presence of micro-fractured intervals. The interpretation presented includes indices of hole degradation as a function of formation lithological variability. Additionally, time lapsed measurements reveal the influence of various drilling parameters such as rates of penetration. pumping rates, and mud density. With supporting evidence of the near wellbore failing mechanism in Cusiana and Cupiagua wells, operational recommendations have been made including mud density alteration and low energy drilling and reaming techniques. Dynamic hole shape characterization provides new insight into directional well control, vibrational analysis calibration, and the effect of dipping beds on drilling performance. Time lapsed caliper measurements also provide environmental corrections for traditional resistivity and gamma ray logs eliminating erroneous correlations caused by enlarged holes or severe rugosity. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER XX NEW TWO FREQUENCY PROPAGATION RESISTIVITY TOOLS W. Hal Meyer Baker Hughes Inteq ABSTRACT Two propagation resistivity tools for measurement while drilling have been developed recently using two frequencies from the same transmitter-receiver array. Two important reasons for adding a 400 kHz measurement to the traditional 2 MHz propagation resistivity measurement are to double the number of depths of investigation and to increase the depth of investigation. Adding a second frequency results in four depths of investigation (an attenuation and a phase at each frequency) instead of the two depths for a traditional 2 MHz only tool. The extra depths of investigation are significant because invasion correction and analysis require the determination of at least three unknowns (diameter of invasion, invasion resistivity, and the resistivity of the virgin formation). Field data and computer simulations demonstrate that the 400 kHz measurements performed by the new tools have increased depth of investigation in both the vertical hole and geosteering modes. One of these tools is a mud-motorintegrated device specifically designed for geosteering, and the increase in depth of detection is particularly important in horizontal drilling. The 4 depths of investigation of this tool allows for the calculation of distance from an approaching bed in horizontal drilling. The other tool (which is a slimhole tool designed for hole sizes as small as 5 7/8 inches) has two different spacings as well as two different frequencies. This results in 8 depths of investigation. In the vertical hole mode the 400 kHz long-spaced measurement is able to measure the true formation resistivity in all but the most extreme cases. Field data and computer simulations show the advantage of using different frequencies at the same location. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER YY A LABORATORY PROCEDURE FOR ESTIMATING IRREDUCIBLE WATER SATURATION FROM CUTTINGS Rosemary Knight Department of Geological Sciences, University of British Columbia, Vancouver, British Columbia Paulette Tercier and David Goertz Department of Geophysics and Astronomy, University of British Columbia, Vancouver, British Columbia ABSTRACT At low levels of water saturation, the water in the pore space of a rock becomes hydraulically disconnected, and effectively immobile. This level of water saturation is referred to as the irreducible water saturation Swi. The magnitude of Swi is an important consideration in formation evaluation, and is at present determined by laboratory measurements on core samples or plugs using porous plate displacement or centrifuge techniques. We have investigated the use of an alternate laboratory technique that uses a measure of drying rates of samples, and can be conducted with samples of any size - from full size cores to small chips and cuttings. As a fully water-saturated sample dries, the drying rate is initially constant and is controlled by the rate of evaporation at the sample’s outer surfaces. During this constant rate period, the water in the sample has a high degree of hydraulic connectivity and is drawn through capillary transport to the sample surface As the level of water saturation decreases, the water loses hydraulic connectivity, and the drying process enters the falling rate period. The transition from the constant rate period to the falling rate period has been associated with the onset of irreducible saturation in a sample. Due to sample-scale saturation heterogeneity this volume-averaged Swi is considered to be an upper limit for the true Swi for the sample. We have conducted a series of laboratory experiments in which we measured the drying rates of sandstone samples as a function of saturation level. We used two sandstones, Berea sandstone and a tight gas sandstone; centrifuge measurements indicated Swi = 0.16 for Berea; and Swi = 0.43 for the tight gas sandstone. For each sandstone we obtained drying rate data on a range of sample sizes from 1 inch plugs to small chips the size of cuttings. Each sample was fully saturated with water then left to dry on a weigh scale interfaced to a computer to record weight as a function of time; weight was converted to Swi . In the data for each sample we clearly see the transition from the constant rate period to the falling rate period and select the saturation level that marks the end of the constant rate period as the estimate of Swi for that sample. Because a decrease in sample size results in reduced saturation heterogeneity, we see a corresponding decrease in the estimated Swi and conclude that samples the size of cuttings yield the most accurate measure of Swi for the sample. Using the drying rate data, for Berea sandstone we obtain Swii = 0.14 for the tight gas sandstone we obtain Swi = 0.47; both of which agree well with previous laboratory measurements, We conclude that the collection of drying rate data is a rapid and simple laboratory procedure that can be used with cuttings to obtain a good estimate of Swi . th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER ZZ DOWNHOLE CORE PRESERVATION FOR IMPROVED FORMATION EVALUATION R.A. Skopec University of Aberdeen, Aberdeen, Scotland P. E. Collee Baker Hughes Inteq, Houston, Texas, USA L. van Puymbroeck Baker Hughes Inteq, Aberdeen, Scotland L.K. Shallenberger Coring Services Consulting, Houston, Texas, USA ABSTRACT A major issue confronting the coring and core analysis industry has been damage to the core during acquisition and handling. The goal of coring and core preservation should be to obtain rock that is representative of the formation while minimizing physical and chemical alteration of the rock during coring and handling. Standard downhole coring assemblies do not preserve in-situ reservoir properties because no provisions are made for core preservation prior to surfacing. Low invasion coring systems help minimize drilling fluid invasion, but rock wettability and saturation can still be altered by drilling fluid imbibition and/or diffusion before core analysis begins. New technology using high-viscosity gel for downhole core encapsulation and preservation is an alternative to operator-intensive wellsite core preservation. Existing low-invasion anti-whirl coring assemblies can be retrofitted to accommodate use of a simple inner barrel piston for gel distribution and core encapsulation. The viscous core preservation gel is a high molecular weight polypropylene glycol which is non-soluble in water and environmentally safe. Laboratory tests at 200°F and 250 psig on water-saturated gel coated rocks with permeabilities as great as 10 Darcys, indicate no spurt loss and zero water loss. The preservation gel is compatible with most water- and oil-based drilling fluids. Because the preservation gel comes in direct contact with the core during and immediately after it is cut, further exposure to drilling fluid is minimized. The high-viscosity gel stabilizes poorly consolidated rocks with moderate compressive strengths and enhances the mechanical integrity of the core. Surface handling of the core is less time consuming and core damage is reduced during transport to the laboratory. For rocks with little cementation, the encapsulating gel can be replaced with a self-hardening plastic, thus eliminating the need for core resination. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER AAA THE EFFECTS OF TEMPERATURE AND WETTABILITY ON RESISTIVITY INDEX MEASUREMENTS ON ROCK SAMPLES B.M. Elashahab, X.D. Jingg and J.S. Archer Imperial College of Science, Technology and Medicine, London ABSTRACT This paper presents an experimental study of the effects of temperature on the resistivity index for sandstone rock samples with different wettabilities. A multi-electrode technique was used to assess saturation distributions along core length. The saturation exponents of five sandstone samples with different wetting characters were measured during both the heating and the cooling cycles under a net confining pressure of 2500 psi.. During both cycles, the temperature was changed in steps between ambient and about 90 degC while allowing equilibrium at each temperature level. Our results indicated that the saturation exponent of both water-wet and non water-wet rocks decreased by 10-15% with the increase of temperature from ambient to around 90 °C. For the water-wet samples, there was no noticeable difference of the end-point saturation exponents between the heating and cooling cycles, which indicates similar fluid distributions between the two cycles. For the non water-wet duplicate samples, however, a clear difference was observed when comparing the saturation exponents at the same temperatures from the heating cycle against that from the cooling cycle. This hysteresis effect between the heating and cooling cycles suggested a trend of wettability changing towards water wetness due to temperature. As a result of the coupled temperature and wettability effects, the conducting phase distributions at pore scale, and hence the resistivity index differ between the heating and cooling cycles. The experimental results have implications in designing special core analysis procedures for resistivity index versus saturation relationships for both water-wet and non-water-wet rocks particularly for the HPHT reservoirs. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER BBB ADVANCES IN WIRELINE FORMATION TESTING John Michaels, Mike Moody, and Than Shwe Western Atlas Logging Services, Houston, Texas ABSTRACT This paper introduces the Reservoir Characterization Instrument (RCI), a new-generation wireline formation testing tool that provides significant improvements, both in the quality of recovered samples and in the analysis of reservoir parameters. The mechanical and electronic design of the instrument and the manner in which tool controls have been designed to accomplish test objectives in a cost-effective manner are described. Numerous advantages are offered by this new-generation formation tester over existing tools, such as surface control over drawdown volume, rate, and pressure drop. The abilities to inject a known fluid into the formation, to pump fluid from the formation prior to sampling, and to identify the fluid being pumped are also advantages over existing instruments. A new and significant capability for ensuring accurate samples is downhole bubblepoint measurement and the ability to control the pressure during sampling. These controls ensure that a fluid sample is taken without altering the gas-to-oil ratio. In addition, the sample may be overpressured downhole to maintain the sample integrity while returning to the surface. Concurrent data interpretation is done to ensure that valid data are obtained, and that test objectives are met during the logging run. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER CCC X-RAY CT EVALUATION OF POORLY CONSOLIDATED, THIN-BEDDED CORE. A. Brancolini, I.S. Mackenzie, F. Radaelli and F. Rossi AGIP SpA, Milano, Italy ABSTRACT Thin layer sedimentation and poorly consolidated rock are common features of Adriatic reservoirs. In this study we have used X-ray CT on whole core in aluminium sleeves and fibre glass to tackle two significant problems relating to estimation of gas reserves in these reservoirs, namely: 1) to detect and quantify the thin sandstone layers (width of order of 1 cm). 2) to obtain porosity information of adequate resolution to estimate the gas capacity of the layers. X-ray CT is well suited to porosity studies in unconsolidated core where conventional methods can prove problematical. Also the porosity of beds which are thinner than typical plug dimensions cannot be measured unambiguously by conventional plug measurements. Using X-ray CT we were able to determine the porosity variation at high resolution using only a few plug porosity measurements to calibrate the CT scale. We also reviewed the correlation of CT data with laboratory measurements of grain size and mineralogical composition. Core photographs subsequently confirmed the success of the CT log, on the sleeved core, in identifying the position, dimensions and frequency of the thin sandstone layers. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER DDD CHARACTERIZATION OF THINLY-BEDDED RESERVOIRS: A NEW APPROACH IN A SPARSE LOGGED AREA N.I. Frass Schlumberger, Caracas, Venezuela C. CoIl, H. Gamero Maraven S. A.. Caracas, Venezuela I. Blyant Schlumberger Doll Research, Ridgefield, Connecticut ABSTRACT The results of the high resolution logging throughout the Lagunillas Inferior reservoir, in the VLD- 1112 well in the Bloque IV area of Lake Maracaibo, Venezuela. are described in this work. Parts of the reservoir were previously overlooked due to the use of simple petrophysical evaluation using very sparse old logs ( SP. Or. Res.) which show high water saturation’s ( Sw> 0.5) and very low porosity’s (φ<0.1). In the intervals with thin laminations, the use of microresistivity images define the clean sands of the shaly section, where pressure tests were taken. With these logs, was possible to define local vertical permeability barriers, which separate the productive well known high permeability zones from other zones which could represent additional reserves. A new sedimentological model was defined using core information and the high resolution microresistivity images, subdividing the reservoir into 12 layers, replacing the old simple 3 layer model. Open hole evaluation of the new well and cased hole measurements using a saturation log indicate that gas has preferentially moved into high permeability zones, leading to bypassing of oil in moderate to low’ permeability layers. With this study the possible amount of oil which was overlooked by Maraven could be about 23% compared to the estimations derived from the simple petrophysical evaluation, in the shaly sands. which have a big impact in future completion programs. Finally these results were used to locate the first horizontal well in Bloque IV area. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER EEE FORMATION PRESSURE DETERMINATION IN HIGH PRESSURE HIGH TEMPERATURE WELLS IN THE CENTRAL GRABEN AREA OF THE NORTH SEA. .Arild Eldoy, Jan Rafdal and Tim Dodson Statoil a.s ABSTRACT In the past years several wells have been drilled to investigate prospects in the 4-5.5 km depth range in the Norwegian Central Graben area. Formation temperatures of 150 - 190 degree C and formation pressures in the order of 1000 - 1100 bar have been measured. Statistics show that lost time due to kicks in HPHT wells in Norwav is about 3 times greater than that of deep wells with normal pressure and temperature. An HPHT programme comprising several projects, whose objective was to improve the drilling performance for these wells, was initiated in 1992 by Statoil. This paper describes one of these projects, the HPHT prediction project. which deals with prediction of formation pressures. As part of the project a database has been compiled comprising drilling and logging data from 22 wells. together with seismic and geological data. The project comprise several studies: a seismic method which uses a combination of velocities and reflectivity (complex seismic decomposition). basin analysis (using 1-dimensional simulation) and well analysis of drilling and logging data. A comparative study was performed where various techniques, used by different companies, have been applied on the same dataset. The main conclusions of these studies are given in the paper. The results of the project illustrate the degree of uncertainty that can be expected based on the techniques and practices currently used by the industry. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER FFF APPLICATION OF POLARIMETRIC BOREHOLE RADAR TO CRACK CHARACTERIZATION Motoyuki Sato, Takashi Miwa and Hiroaki Niitsuma Faculty of Engineering, Tohoku University, Japan ABSTRACT In this paper we discuss potential of polarimetric borehole radar measurement for formation and reservoir evaluation. Borehole radar have not been used in low resistive formation due to its short penetration depth. However. polarimetric radar measurement has good possibility to give us more information about formation and expands its practical usage. A prototype broad-band (10MHz - 500MHz) polarimetric borehole radar system was made using a network analyzer and an optical signal link. Single-hole and cross-hole measurement using the radar system was carried out in granite rock, The measured signals showed that detection of the short-range reflections can significantly be improved with cross-polarization measurements since direct coupling between the transmitting and receiving antennas is reduced. Applying an inverse filter method to compensate antenna characteristics, polarization dependency of reflectivity from subsurface cracks could be quantitatively evaluated as a scattering matrix. We discuss the polarization dependency of the scattering matrix of cracks for characterization. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER GGG A REVIEW OF GAS HYDRATES AND FORMATION EVALUATION OF HYDRATE-BEARING RESERVOIRS Stephen Prensky U.S. Geological Survey, Denver, CO ABSTRACT Gas hydrates occur within a limited range of pressure and temperature that is found in ocean sediments and in association with permafrost in the Arctic. Gas hydrates represent an enormous potential energy resource, estimates of the methane gas contained in worldwide methane hydrate accumulations exceed estimates of the combined reserves of oil and gas that can be produced by conventional methods. The Messoyakha field, in the west Siberian basin, has produced natural gas from gas hydrates since the early 1970s. Gas hydrates and hydrate-bearing rocks and sediments are characterized by high electrical resistivities, high acoustic velocities, and low bulk densities (in zones of massive hydrates) relative to water or gas-bearing rocks or sediments. In oceanic settings or below the base of ice-bearing permafrost, well-log measurements of these properties can readily identify gas-hydrate zones. However, the presence of ice (which has similar physical properties to gas hydrates) within a permafrost interval complicates the use of logs for qualitative identification and quantitative evaluation. Identification of gas hydrates requires a combination of well logs in conjunction with gas shows on drilling or mud logs. Thermal invasion during drilling or coring can dissociate hydrates and introduce additional complicating factors. Although the physical properties of gas hydrates are well known, little is known about those of hydrate-bearing rocks and sediments. Additional laboratory data are required for calibrating quantitative well-log and seismic techniques for evaluating gas-hydrate-bearing reservoirs and assessing the their methane content. Some of the unique physical properties of gas hydrates can be measured by new wireline logging devices (e.g., array resistivity, dipole acoustic, dielectric, inelastic and capture gamma-ray spectrometry, and borehole imaging) whose application may help distinguish ice from hydrates in permafrost regions, provide improved determinations of porosity, identify type of hydrates present and their distribution within a sediment or rock matrix, and they enable improved determination of the degree of gas-hydrate saturation in hydrate-bearing zones. New and improved measurementwhile-drilling (MWD) logging devices offer the potential for detecting and quantifying gas hydrates prior to any decomposition and may provide early warning of potential drilling hazards. The U.S. Geological Survey, through its Gas Hydrate Research Program, is developing advanced gas-hydrate welllog and seismic models based on (1) laboratory modeling of tool response using actual gas hydrates as well as synthetic samples, and (2) analysis of well-logging data from the Alaskan North Slope and Ocean Drilling Program. This report summarizes the current state of gas-hydrate formation evaluation and discusses work in progress. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER HHH FORWARD MODELING OF DEEP RESISTIVITY RESPONSE IN NORPHLET WELLS Gary F. Beck Chevron Production Co., New Orleans, LA, U.S.A. ABSTRACT The Norphlet Trend is a deeply buried (18,000 -24,000 feet) Upper Jurassic aeolian sandstone that occurs from onshore Mississippi to offshore Florida. Many Norphlet wells exhibit a funnel-shaped resistivity profile (decreasing resistivity with depth) that typically extends over several hundred feet. This response can be attributed to an extremely long transition zone that occurs in the fine grained portion of wind-rippled aeolian strata. The fine grained laminations require an extremely long gas column in order to reach irreducible water saturation while the coarser grained laminations reach a lower irreducible water saturation within several feet of the gas/water contact. The resulting bi-modal distribution of water saturation between the fine grained and coarse grained laminations in the reservoir column results in inaccurate water saturation estimates if standard analytical methods are used. A model is presented to predict the deep resistivity curve based on column height, capillary pressure data. fine and coarse grained lamination porosities, and fine to coarse volumetric ratios. This model yields a theoretical resistivity curve that matches the observed resistivity profile. Application of the model in reverse using observed resistivity enables calculation of more accurate water saturations than those from conventional techniques. The model has also been used as a predictor of gas/water contacts in wells where the free water level is below the total depth of the well. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER III REMOVAL OF BOREHOLE INDUCED NOISE FROM WELL LOGS J.A. Nieto, D.P. Schmitt, and R.G. Keys Mobil Exploration and Producing Technical Center, Dallas, TX, U.S.A ABSTRACT Under certain circumstances (bit whirl in oil based mud environment), a distinctive “sinusoidal” borehole rugosity can be observed on the wireline caliper log. This borehole rugosity can have an adverse effect on the response of logging devices which require borehole contact with the tool. Among such devices are the primary porosity and lithology identification tools, density and neutron. The present method eliminates the sinusoidal noise without affecting the response of the formation. This means that the actual formation responses of the logging tools are recovered, and the logs can be used quantitatively. The removal of sinusoidal noise from a log is accomplished in three steps: 1) the log is reduced to a zero mean stationary series, 2) the wavenumber of the sinusoidal noise is identified by its peak in the Fourier amplitude spectrum, and 3) the noise is removed by applying a zero-phase notch filter. Occasionally, harmonic noise peaks will be present at multiples of the sinusoidal noise wavenumber. In order to preserve the quantitative data integrity, the low wavenumber trend must be kept. This is accomplished by approximating the log with a piece-wise cubic spline (in the least squares sense) which retains the overall log character, ignoring the sinusoidal noise. Standard methods, such as removing the mean, will not preserve the low wavenumber information. A zero mean stationary series is formed by subtracting the least-squares cubic spline from the data. The remaining steps, namely, spectral analysis and notch filtering are performed on the difference series. Recombining the filtered series with the spIne restores the log data - without the sinusoidal noise. An advantage of this procedure is that the results are independent of the depth interval used to perform the analysis. This technique has been applied successfully in several wells world wide. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER JJJ ORIENTATION AND CALIBRATION OF CORE AND BOREHOLE IMAGE DATA Benoit Mathis Elf Aquitaine Production, Pau, France Dents Hailer Elf Petroleum Norge, Stavanger, Norway Herve Ganem Schlumberger GeoQuest, Montrouge, France Eric Standen Schlumberger Wireline and Testing, Montrouge, France ABSTRACT Numerous data sets of differing source and scale are usually used in the detailed sedimentary and structural analysis of a reservoir. Core analysis, when available, is of prime importance in defining the sedimentary structures, critical for the understanding of depositional environments and facies within a reservoir. An added input is the orientation of various bedding features, which complements the vertical facies variations. Core is seldom, however, oriented when it is taken. The orientation of core has been done in the past, by correlating with borehole imagery techniques such as electrical or acoustic images, but, always with a considerable inconsistency. This accuracy problem becomes important when defining the orientation of open, natural fractures for the preferred permeability direction of the reservoir, and detailed cross-bed directions - for paleo-current estimation, depositional environment analysis and overall reservoir geometry. This paper describes a technique for much improved depth-matching of core images to log data and orientation of core features using wellbore images. A workstation system has been developed, which allows the building of a detailed composite display of all the data at the same scale, in order to define the reservoir property variations. The data can include high resolution borehole images, core sedimentological analysis, full circumference core images, standard core photographs of slabbed core and petrophysical formation evaluations as well as well test information and interpretation comments. All core images can be depth matched to the borehole images and full circumference core images, oriented so that features on the borehole images can be calibrated to known events in the core. Natural fractures can be differentiated from drilling-induced fractures, giving a procedure for differentiating fracture types on the images in non-cored sections. With this complete analysis, a detailed stratigraphic model of the reservoir is produced to improve the understanding of the relationships between hydrocarbon saturation (both core and log), grain and pore size variations, fracture cementation, and sedimentary and diagenetic processes. The classical description of sedimentary features in core is complemented by their orientation from both data sets, and leads to a better definition of the nature and geometry of the reservoir. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER KKK CONTRIBUTION OF QUALITATIVE LOG ANALYSIS TO RESERVOIR MANAGEMENT RECENT DEVELOPMENT AND CASE HISTORIES B.C. Duval, P. Imbert, Y. Grosjean and J.J. Lefort TOTAL, Paris, France ABSTRACT The paper will focus on a qualitative type of log analysis, somewhat different from the commonly used quantitative methods. This process makes wireline logs of paramount importance in evaluating reservoir geometry and volume, predicting deliverabilities and more generally contributing to field development and reservoir management. This technique of qualitative log analysis. known as “electrofacies interpretation”, has been in use in the industry for many years. New developments in sedimentology and stratigraphy (sequence stratigraphy in particular) have pushed its predictive character still farther. Three examples from various geological settings will highlight the versatility of the tool and some of its recent contributions. A first example from Indonesia (Tunu gas and condensate field) is a reminder of how this type of technique can discriminate between channel and mouth-bar types of reservoir in a deltaic environment. The former type of reservoir is characterized by excellent petrophysical properties and an average width/depth ratio of about 100, whereas the latter is more extensive laterally, but with poorer reservoirs (different k/phi relationships). Another case study from Argentina (Hidra oil field, Tierra del Fuego) shows the use of electrofacies analysis, properly calibrated on cores, to distinguish laterally extensive permeability barriers linked with regional tectonic activity, from local ones which have litlle influence on reservoir behaviour. A third case illustrates the contribution of log analysis and the sequence stratigraphy approach to understanding the regional cyclicity of deposition within a thick deltaic sedimentary column. This technique, combined with systematic wireline formation testing is determinant in reserve calculation and perforation strategy (Mahakam new gas fields of Kalimantan, Indonesia). th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER LLL SHALES: AN ALTERNATE SOURCE FOR WATER RESISTIVITIES Darrel E. Cannon Schlumberger Wireline & Testing ABSTRACT The accurate determination of hydrocarbons in place is fundamental in both exploration and development wells. Log-based computations of porosity (φ) and water saturation (Sw) are a major input into this determination. Of the two, Sw is the most error prone, often because of inaccurate knowledge of formation water resistivity (Rw). An accurate assessment of Rw is often difficult to obtain because there may be no 100% water zones to back calculate Rw from logs, no applicable Rw catalogs, Rw may vary between water and pay zones, no samples of produced water or contaminated water samples. An untapped source of Rw information can be found in the shale beds adjacent to the zones of interest. Extracting apparent Rw from shales is accomplished by inverting any water saturation equation that compensates for the electrical properties of clay (Qv). The required inputs to calculate Rw are resistivity, total porosity and Qv. Resistivity is estimated from deep resistivity logs. Qv and porosity are estimated from density-neutron logs using a model-based approach. To test this technique, a model for estimating clay volume and Qv was developed. Three water saturation models were evaluated: Waxman-Smits, Dual Water and Sen-Goode. The different models result in similar answers in highsalinity reservoirs but diverge in low salinity reservoirs. To use these models, a reevaluation of the water saturation concept of m and its relationship to Qv is required. The standard assumption that m increases with increasing Qv is found to be valid in normally pressured, consolidated formations but apparently is invalid in overpressured formations. In the latter, a constant value of m for all models is found to be appropriate for both sands and shales. The simple model used in this technique assumes that the shales are in ionic equilibrium with adjacent sands (no active waterfloods), are composed predominantly of clays, quartz and feldspars, and have total porosity that is filled with water not organic matter. This technique has been successfully applied to wells in Canada, the United States, South America and Africa. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER MMM RECONCILIATION OF WATER PRODUCTION FROM RESISTIVE CARBONATES, WESTERN CANADA Tons T. Dziuba and Larry W. Sopiko ABSTRACT Porosity, this apparently simple parameter, requires more attention than it usually is given in the evaluation of carbonates. Often, if two independent porosity sources such as logs and core agree, the petrophysicist assumes that the correct porosity has been established. This assumption can sometimes lead to the incorrect prediction of fluid types as well as an overestimation of hydrocarbon reserves. Total porosity (øt), as measured by porosity logs and routine core analysis, is made up of both effective (φe) and ineffective porosity (φie). For example, in hydrocarbon-productive reservoir rocks φe ≅ φt, but in marginal and nonproductive ones, φe << φt. This change in hydrocarbon productivity can be subtle with few indications from gamma ray, porosity, and resistivity logs. Resistivity logs, responding to the effective porosity (φe), can confuse the petrophysicist because the higher resistivities are due to low effective porosities and not hydrocarbons. From work completed on Canadian carbonates, experience has shown φe will be even lower than the neutron, density, and sonic-predicted porosities, in marginal and non-productive reservoirs. Nuclear magnetic resonance (NMR) logs are showing promise in helping to unravel carbonate pore geometries but nevertheless, calibrated openhole logs should not be neglected. The degree of reduction in φt will be a function of the depositional environment, the corresponding matrix qualities, hydrocarbon fluid types, height of reservoir above the free water level and pore infill cements, An example is cited from Western Canada where high resistivities in vuggy carbonate rocks was interpreted incorrectly as indicative of hydrocarbons in the reservoir. The well produced salt water on test. The contradictions can be explained using computer analyses, capillary pressure curves and geology to show that φe was overestimated. Using the Archie equation and the correct φe value resulted in Sw> 50 percent in the connected pore space and a reconciliation of water production despite high resistivities. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER NNN AN INTEGRATED APPROACH TO SATURATION HEIGHT ANALYSIS Christopher Skelt Scott Pickford Group Bob Harrison Enterprise Oil ABSTRACT One of the principal contributions made by petrophysicists to an understanding of hydrocarbon distribution within the reservoir is the saturation height function. Unfortunately, the shortage of specifically designed commercial software forces many petrophysicists to transform the input data into a domain where it is quasi-linear, so that least squares linear regression techniques can be used to derive coefficients in the equations. This can impose artificial weighting and undesirable constraints on the fitting process. We present a simple, robust, non-linear formulation and optimization method, designed so that each term in the function can be related directly to a physical parameter such as irreducible water saturation, ratio of contact angle and surface tension between laboratory and reservoir conditions, threshold capillary entry pressure, and height difference between free water level and oil water contact. The transformation applied to the function by altering each term is predictable, comprehensible, and independent of the other terms. This property allows petrophysicists to make optimal use of the log, capillary pressure and other special core data at their disposal, capitalising on the relative merits of each type of data. When deriving a saturation height relationship petrophysicists need to be aware of the variation of field area with height above contact. The weighting option of the optimization process recognises this requirement, and fits the data best where each foot of vertical height represents the largest areal extent of the field. Finally, the implications of saturation-related phenomena connected with variations in rock properties observed in the wells on field-wide hydrocarbon distribution are discussed. The optimization method, backed by resistivity profile modelling, is used to distinguish artifacts arising from resolution incompatibilities from real rock property variations. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER OOO VERTICAL RESOLUTION AND SIGNAL-TO-NOISE CHARACTERISTICS OF NUCLEAR MAGNETIC RESONANCE IMAGING LOGS D.T. Georgi, J.C. Zhang, and Y. Y. Huang Western Atlas Logging Services, Houston, Texas R.N. Chandler and M.G. Prammer NUMAR Corp., Malvern, Pennsylvania ABSTRACT The vertical resolution and the signal-to-noise characteristics of the nuclear Magnetic Resonance Image Log (MRIL) are investigated. The theoretical vertical resolution is determined from the physical dimensions of the magnet and the radio frequency antenna. The signal-to-noise (signal power: noise power) characteristics depend on the quality of the echo data (formation signal and system noise). Furthermore, the signal-to-noise characteristics of formation evaluation data derived from the NMR echoes (e.g. MRIL porosity, bulk volume irreducible water, and bulk volume moveable fluid) depend on the details of the time-domain to T2 - domain transformation procedure. The theoretical data are compared with results obtained from time-series analyses of MRIL data from an Austin, Texas test well. The MRIL was logged in combination with conventional density and neutron porosity tools. The vertical resolution and signal-to-noise were determined from multiple passes collected at 3, 10 and 30 ft/minute with a standard 6 inch diameter, 43 inch aperture magnet probe. The analysis quantifies the dependence of signal-to-noise and vertical resolution of the MRIL data on the logging speed. In addition, data collected with a new 24 inch aperture, high resolution magnet probe were also analyzed. As expected, tests showed that the new probe exhibits significantly better vertical resolution with no noticeable decrease in signal-to-noise characteristics compared to the 43 inch aperture probe. At all logging speeds, the vertical resolution compares favorably with conventional neutron and density porosity data. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER PPP ESTIMATION OF CRITICAL FORMATION EVALUATION PARAMETERS USING TECHNIQUES OF NEUROCOMPUTING Jonathan Hall and Livio Scandella AGIP S.p.A., Milano, Italy ABSTRACT Apart from within only a few Oil Companies and Contractors, Neurocomputing (which includes Neural Networks) has made little impact, as yet, on day to day working practices in Formation Evaluation, despite publication in the ‘Literature’ of some practical applications and advantages. Why is this so? Neurocomputing techniques of Estimation differ fundamentally from traditional statistical techniques of regression analysis which require a priori assumption of the functional form of the dependency. Neural Networks learn the nature of this dependency through a carefully selected and representative set of training (and validation) examples. Properly trained Neural Networks also differ from multi-dimensional crossplotting techniques (which exhibit Associative memory) by learning a global dependency and not a local dependency based upon a function of other local examples. By establishing a general dependency, Neural Networks can exhibit remarkable tolerance of noisy and incomplete data sets and demonstrate remarkable speed in processing. Through selected case studies we explore some limitations of conventional statistical estimation techniques, as well as Neural Networks, but also offer some advantages of the use of the latter. We demonstrate some practical applications including; volumetric analysis from logs, treatment of data from the FSU, log to core calibration, and suggest some future applications. Further, we explore issues of generalisation common to both conventional techniques and those of neuro-computing and try to answer the question of how reliable these techniques may be. Hybrid techniques of Fuzzy Logic and Neural Networks as well as techniques of feature analysis of borehole images are introduced. To demonstrate these applications, a prototype interface was developed. The software enables the user to design appropriate training selection strategies, construct Neural Network architectures, process, and analyse results, thereby offering, an integrated tool for Formation Evaluation th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER QQQ FORMATION RESISTIVITY FACTOR AND PERMEABILITY RELATIONSHIPS IN ROCKS CHARACTERIZED BY SECONDARY SOLUTION POROSITY D. C. Herrick and W. D. Kennedy Mobil Exploration and Producing Technical Center, Dallas, Texas ABSTRACT Values of porosity exponent m obtained on a suite of rock samples having varying amounts of secondary porosity tend to be quite variable and unusually high. The variability complicates the prediction of electrical properties and calculated hydrocarbon saturations in these rocks. An understanding of how isolated void space in the form of vugs or moldic pores modifies the electrical behavior of the host rock can obviate problems with prediction and use of porosity exponents in formations with secondary porosity. Since the dominant pore-geometric factor controlling the flow of fluids and electric current in the pore space is the intergranular matrix of the host rock, addition of isolated voids affects the porosity in direct proportion to the void space added; however, the resistivity is much less affected. To correctly interpret the electrical (and, to a lesser extent, hydraulic) behavior of such rocks the two components of porosity must be separated. The Maxwell-Garnett equation can be used to describe F in terms of matrix porosity and a dilute concentration of uniform spherical vugs. This paper demonstrates that the Maxwell-Garnett relationship can be applied with a high degree of accuracy to numerical models which have a realistic concentration of vugs as well as real rocks. The separation of porosity components using the Maxwell-Garnett relationship is accomplished using core and log data. In favorable circumstances, the core data may not be required. The suggested separation of porosity components results in better estimates of water saturation and more reasonable porosity-permeability transforms than are usually obtained in vuggy rocks. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER RRR A KNOWLEDGE BASED SYSTEM FOR LOG QUALITY CONTROL Mauro Gonfalini and Marco Ponzi AGIP S.p.A., Milano, Italy ABSTRACT Log Quality Control (LQC) is one of the most important activities within the Formation Evaluation process adopted by a large number of oil companies. Since an effective LQC requires the setting of rigid QC standards and the use of well defined procedures, a knowledge based (KB) system is an ideal tool for providing computer support to log quality management personnel. AGIP promoted a project aimed at the automatic compilation of an LQC form, starting from the digital data provided on tape and in LIS format by the service companies. In designing the system, the current operational LQC practices were adopted and the visual organisation of the LQC form preserved. The overall validation technique is defined by a set of methods and a set of rules for the logical composition of these methods on the basis of possible log quality problems. The system operates both in automatic and interactive modes. The system also supports an interactive multi-well validation procedure. Using the functionalities provided by the interface, the user can localise the poor quality intervals and assess the severity of the defect. The prototype has been validated on a large number of standard technology services recorded in AGLP wells, allowing the definitions of its advantages and limitations. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER SSS IS THE SATURATION EXPONENT N A CONSTANT? S.O. Stalheim and T. Eidesmo Statoil Norway ABSTRACT A new physical interpretation of the relationship between the resistivity index, RI, and the water saturation of a reservoir rock, Sw is presented. To date the most commonly used method to calculate Sw are based on Archie’s equations, where the saturation exponent n and cementation exponent m are needed. Archie’s model is sensitive to the uncertainty in both n and m, especially n, and the accuracy of water saturation depends strongly on the uncertainly of the estimated n. Conventionally it is assumed to be a constant, independent of water saturation. This is not always true. Several laboratory measurements on core plugs have showed both increasing and decreasing values of n when water saturation is decreasing monotonously. The effect is especially noticeable in the low water saturation region. The results presented in this paper explains this observation by the assumption that the resistivity index is sensitive to the geometry of the nonconducting fluid phase (oil). Different oil geometries show different value of RI for a given Sw. As the oil saturation increases, the geometry assumed to gradually change from approximately spheres to geometries closer to the pore space. The results show that the RI, for a given level of water saturation, might differs according to the geometry of the oil phase. Measurements performed on simple laboratory models verify this geometrical dependence. A new model valid for spherical shaped oil volumes and examples from core measurements where these effects have been observed are presented. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER TTT PRACTICAL ANALYSIS OF FORMATION STRENGTH PROPERTIES IN BOTH CONSOLIDATED AND UNCONSOLIDATED RESERVOIRS E. L Bigelow Western Atlas Logging Services, Houston, Texas E. P. Howell Unocal, Sugar Land, Texas ABSTRACT Engineers need accurate information on rock strength properties to optimize reservoir management, particularly to avoid mechanical failure in wells. This requires proper pressure maintenance during drilling, completion, and production operations. In unconsolidated reservoirs, rock strength information is essential to accurately predict drawdown pressure values where problems with sand production commence. An empirical but practical technique is currently used successfully to predict drawdown pressure limits for such sands in the Gulf of Mexico and elsewhere. The technique utilizes acoustic televiewer images and array monopole/ dipole acoustic data to determine the critical drawdown pressure values. When used in lieu of expensive coring and core analysis, the technique can save several million dollars in unnecessary acquisition and analysis costs. Gravel packs and expensive cleanout operations in producing wells are less frequent. The expense of microfrac stimulation is often avoided. Today, accurate shear and compressional data are commercially available along with acoustic formation imaging so that estimates of in situ properties can be developed with enough accuracy to effectively engineer reservoir performance. Acoustic log data for these successful evaluations are derived from an array of monopole and lowfrequency dipole acoustic measurements that are capable of accurately defining both compressional and shear slowness even in formations with velocities below 2500 ft/sec. Prediction of drawdown limitations are made where acoustic imaging data substantiate occurrences of borehole breakout Major improvements in acoustic logs have made difficult engineering calculations more practical. Field examples of acquired data, interpretation, and results demonstrate the technical and economic success of the practical interpretative technique. Space restrictions limit the article to examples from one well although sufficient evidence has been found in other wells -- in the Gulf of Mexico and the continental U. S. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER UUU BOREHOLE VISUALIZATION IN 3-D Douglas Seller Halliburton, Houston, Texas ABSTRACT The challenge in the interpretation of petrophysical data is to produce a picture which is sufficiently comprehensive to accurately describe complex geology; yet in a format sufficiently simple to be easily assimilated by the human eye-brain network. Most geophysical problems are three dimensional in nature, yet until recently, wireline solutions to these problems have been limited to two dimensions. Current imaging tools such as the Circumferential Acoustic Scanning Tool (CAST), and the Electrical Micro Imaging tool (EMI) describe the formation in much greater detail than conventional wireline tools. It is now possible to produce realistic 3-D colour images and interact with these images in real-time, just as you might manipulate a core sample in your hand. Vast quantities of numerical data gathered by imaging devices can be transformed into easily understood images. Imaging tools represent a new generation of logging tools which can solve problems without the benefit of traditional assumptions. Historically, wireline logging techniques have been premised on two fundamental assumptions: first, the formation is layered and second, the layering is roughly normal to the borehole axis. Correlatable features on seismic sections and wiggle on wireline logs represent the layering phenomena. As a result of the borehole wall mapping approach, these devices provide a much more detailed description of the borehole and geological features. These descriptions are not necessarily bound by a dependence on the layered nature of sedimentary rocks. Three dimensional visualization derived from combining the time-of-flight and amplitude from the CAST, and electrical conductivity from the EMI are highlighted in this paper. A comprehensive set of examples demonstrate various enhancement techniques which the user may apply interactively while viewing the image on a workstation. The accurate description of the borehole wall allows the user to solve petrophysical problems beyond the capability of conventional tools. th SPWLA 36 Annual Logging Symposium, June 26-29, 1995 PAPER VVV STATISTICAL TESTS OF PERMEABILITY ESTIMATES BASED ON NMR MEASUREMENTS P. Kubica Petro- Canada Resources ABSTRACT Permeability estimates are routinely calculated from downhole NMR measurements. Identical NMR parameters can be measured by laboratory NMR spectrometers on core plugs. The purpose of this study is to examine the validity and the statistical significance of presently used permeability correlations. Two sandstone formations of Conoco 33-5 test well were extensively investigated by conventional core measurements (porosity, permeability) as well as by NMR methods. 88 core plugs were measured in low field NMR spectrometer and the NMR data were correlated to the core measured porosity and permeability. Results indicate a high degree of statistical correlation between the core measured permeability and NIMR measurements of free fluid index (FF1) and the transverse relaxation time T2. The multiple regression of permeability on NMR data confirms a strong correlation of permeability to (1-FFI) and NMR porosity. An equally significant statistical correlation was found between the permeability and NMR total porosity combined with T2. While the derived permeability relations may not be universally applicable they provide an important check on permeability models. TRANSACTIONS OFTHE SPWLA THRTY-SEVENTH ANNUALLOGGINGSYMPOSlUM Sponsoredby THE SOCIETY OF PROFESSIONAL WELL LOG ANALYSTS, INC. 8866 Gulf Freeway, Suite 320 Houston, Texas 77017 Presentedat THE MARRIO’IT HOTEL New Orleans, LA June 16-19, 1996 NOTICE TO EDITORS: Permission is hereby granted to publish elsewhere any of these transactions atter June 19, 1996, provided that conspicuousacknowledgementis given to the original presentation of the paper and the authors of the paper have agreedto the republication. (The statements and opinions expressed in these transactions are those of the authors and should not be construed as an official action or opinion of the Society of ProfessionalWell Log Analysts, Inc.) SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER A THE RESPONSE OF MULTIARRAY INDUCTION TOOLS IN HIGHLY DIPPING FORMATIONS WITH INVASION AND IN ARBITRARY 3D GEOMETRIES Barbara Anderson, Thomas Barber, Vladimir Druskin, Ping Lee, and Elizabeth Dussan V. Schlumberger Leonid Knizhnerman and Sofia Davydycheva Central Geophysical Expedition, Moscow ABSTRACT With the rapid expansion of horizontal drilling, the interpretation of logs, especially resistivity logs, has become a serious problem. The proximity of shale layers or of water legs can seriously affect deep resistivity logs, and invasion can seriously affect shallow resistivity logs. The current state of affairs is that determining Rt in a horizontal or very high angle well is often impossible. Modeling techniques are now available for solving the full 3D problem necessary for deviated well interpretation. We have developed such a 3D modeling code and applied it to improve the interpretation of multiarray induction tool response. The code uses the Lanczos spectral-decomposition method to solve Maxwell’s equations on a staggered finite-difference grid. The finite-difference code has been benchmarked against analytical solutions for subsets of the 3D geometry, and agreement is within 3%. When run on a parallel machine, 50 ft of 3D log can be generated in under 6 hours. The code takes into account dipping beds and unsymmetrical invasion at the same time, as well as resistivity anisotropy. Several horizontal well interpretation problems have been investigated with the code. One is the case of axisymmetric cylindrical invasion in a permeable zone below a cap shale interface. In this case modeling shows that for shallow invasion, the deepest Array Induction Imager tool curves can be used to infer and proximity to the shale cap, while the shallowest curve indicates Rxo . If deeper invasion is modeled, only the deepest induction curve indicates Rt,, while several of the shallow curves read Rxo. The code has also been used to model non-circular invasion fronts caused by either permeability anisotropy or buoyancy segregation typical in highly deviated wells. Both cases are characterized by a considerable quantity of filtrate shunted away from the well in preferential directions, resulting in less invading fluid near the wellbore. As a consequence, there is an increase of the influence of Rt on the shallow AlT logs. These cases indicate that induction logs in complex formations still have geometrical interpretations, but that they are different than interpretations used in vertical wells. A log example illustrates the power of 3D modeling in interpreting multiarray induction logs in difficult wells. In a horizontal well with moderately salty invasion, modeling shows that a large separation between the deepest induction curves is caused by a combination of invasion effects and polarization horns near a cap shale. In addition, an annulus is present to complicate interpretation. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER B A NEW, FULLY DIGITAL, FULL-SPECTRUM INDUCTION DEVICE FOR DETERMIMNG ACCURATE RESISTIVITY WITH ENHANCED DIAGNOSTICS AND DATA INTEGRITY VERIFICATION D. L Beard, Q. Zhou, and E. L. Bigelow Western Atlas Logging Services, Houston. Texas ABSTRACT More than 30 years ago, it was recognized that a minimum of four independent resistivity measurements with different depths of investigation is needed to effectively resolve true formation resistivity when deep invasion occurs. The dual-induction and shallow laterolog measurements were combined and run with microresistivity devices: this method increased the accuracy with which true formation resistivities were determined. Later, the dual-phase induction system, with enhanced induction deconvolution, was an attempt to improve vertical resolution. A new induction logging system [High-Definition Induction Log (HDILSM)] effectively addresses both issues. The recently introduced induction tool has seven coil arrays, each making simultaneous measurements at eight different frequencies over a broadband from 10 kHz to 150 kHz. The short-spaced array has an 8-in. (20-cm) median depth of investigation, while the long-spaced array has a much deeper 120-in. (305-cm) median depth of investigation. Using individual coil arrays to try to increase the depth of investigation or improve vertical resolution can do little to eliminate cave effect. Focusing for the system is performed in software so that depth of investigation is optimized and vertical resolution is improved, while cave effect is substantially reduced. There is a trade-off between vertical resolution and sensitivity to near-borehole artifacts when considering measurement responses. A variety of curves with different degrees of trade-off are provided. Matched resolution curves with 1-, 2-, or 4-ft (0.3-, 0.6-, or 1.2-m) vertical resolution are available, as are “true resolution artifactfree” curves. The new instrument is fully digital, so there is a wealth of new signal processing and diagnostic techniques. The multifrequency data set allows the use of a new skin-effect correction scheme. A robust technique free of many problems found in other skin-effect correction methods is implemented. While supplying self-consistency and data quality indicators, the system also provides tool-fault tolerance: e.g., the loss of one receiver channel. Data completeness ensures its usefulness for future developments in processing techniques: e.g., a joint-inversion scheme. Digital acquisition schemes permit received signals to be viewed in the time domain, and thus open a new realm of real-time data-integrity-checking and diagnostics techniques. For SPWLA 37th Annual Logging Symposium, June 16-19, 1996 example, magnetic materials in the borehole or formation can be immediately and distinctly recognized in the time-domain waveform. A number of products can be computed from the more accurate multiple induction measurements. These products include true formation resistivity, flushed zone resistivity, and the volumes and radial limits of both flushing and transitional invasion. The more accurate resistivity values are included in traditional schemes’ calculations of formation water resistivity, water saturation, and residual/movable hydrocarbon values, which are ultimately used for in-place reserve calculations and equity determinations. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER C ELECTRICAL PROPERTIES OF ROCKS: EFFECTS OF SECONDARY POROSITY, LAMINATIONS, AND THIN BEDS D. C. Herrick and W, D. Kennedy: Mobil Exploration and Producing Technical Center, Dallas Texas ABSTRACT Pore geometries can be classified according to their electrical properties. Three main classes of conducting geometries are identified: 1) “Archie” rocks having a simple intergranular pore system similar to the rocks described by Archie (1942); 2) rocks with isolated porosity such as vugs, moldic pores, or microporous grains; and 3) rocks with components which have different capillary properties and which conduct as if the components were electrically parallel conductors. The electrical properties of the third category are often manifested as low resistivity pay. Low resistivity pay can be encountered in a variety of geologic settings such as sandstones with diagenetic clay coatings, thin clean sandstones interbedded with shales, and aeolian sandstones. Any depositional or diagenetic environment which produces a rock fabric with a continuously connected system of small pores in addition to larger pores connected by proportionally large pore throats can result in low resistivity pay. In these types of bi- or polymodal pore systems, the capillary properties of the continuous systems of smaller pores results in a connected pore subsystem which is largely or completely water saturated. A coexisting system with larger pores may contain high hydrocarbon saturation. Such rocks can be modeled and understood using their capillary properties and a parallel conductor resistivity model. Using Archie’s equation to evaluate hydrocarbon reserves in rocks of this class, low values of the saturation exponent n must be used, typically less than 1.5 and sometimes less than one. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER D A THIN BED MODEL FOR THE KUPARUK A SAND KUPARUK RIVER FIELD, NORTH SLOPE, ALASKA Jerry Sovich, Jim Klein, and Neil Gaynor ARCO Exploration and Production Technology ABSTRACT The Kuparuk A sand reservoir has performed beyond the predictions of the original reservoir model. Water breakthrough has been late, original oil volumes appear to have been underestimated, and hydraulic fracture performance has exceeded expectations. The Kuparuk A sand was deposited in a nearshore, storm-dominated marine environment and is characterized by a high degree of lithologic heterogeneity. It has long been recognized from core that numerous thin clean sands (less than 0.5-1 ft thick) comprise a significant portion of the reservoir. The original petrophysical model was not able to resolve these features and computed “pay” intervals were restricted to thicker (greater than 1-2 ft thick) and less shaly sequences. Recent work with micro-scanner resistivity data has shown that the thin sand beds can be resolved and the volume of sand computed. The electrical images can easily segregate thin sand and shale lithologies because of their resistivity contrast This contrast allows the thinlyinterbedded sand-shale sequences to be quantified using a net-to-gross curve. Results were calibrated using whole core to verify the image analysis. Integration of the micro-scanner resistivity data with a conventional clay volume algorithm led to a thin-bed model that was applied field-wide. This new thin-bed model is unique in that it has the ability to distinguish the thicker, water floodable pay sands, from the thinner clean beds which may have a lower recovery factor. The relationship between clay volume and net-to-gross determined from the electrical images furnishes a decreasing thin bed sand volume with increasing clay volume. Application of the updated thin bed model should impact the oil-in-place determination and lead to recognition of possible recompletion opportunities in the “A” sand. Additionally, such opportunities may be recognized in the overlying Kuparuk “B” interval where numerous hydrocarbon-bearing thin sand beds are present . SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER E QUANTITATIVE LITHOLOGY: AN APPLICATION FOR OPEN AND CASED HOLE SPECTROSCOPY Susan L. Herron and Michael M. Herron Schlumberger-Doll Research, Ridgefield, CT ABSTRACT A new quantitative lithology interpretation is based on elemental concentrations available from logs. The concentration logs are obtained from single, induced-neutron gamma ray spectrometers, thus differentiating this work from earlier geochemical interpretations which required additional sondes for the measurement of aluminum and potassium. The new interpretation offers considerable advantages over conventional logs. This new interpretation is based on a recently acquired data base, Fourier transform infrared mineralogy and chemical compositions were measured on over 400 samples to examine the relationships between lithology and a number of geochemical signatures, including gamma ray, the gamma ray components Th, U, and K and other loggable elements. The results show that in individual wells, gamma ray correlates roughly with total clay content, but a closer analysis exposes the inherent weaknesses in clay estimation. Breaking gamma ray down into its individual components brings little improvement to the clay estimation compared to that possible from a different suite of geochemical logs. The study reveals a strong linear relationship between aluminum and total clay concentrations. It is characterized by a near-zero intercept and a common slope. Unfortunately, the measurement of aluminum by logging devices has proven to be difficult and expensive. This study introduces a technique whereby the elements silicon, calcium and iron can be used to produce as accurate an estimation of clay as from aluminum. A general algorithm for predicting clay from these three elements is presented. An example of how the algorithm can be optimized is also provided. Carbonate concentrations are determined from the calcium concentration log with an accuracy that is not available from any other logging devices. Finally, the remainder of the formation is composed of quartz, feldspar, and mica minerals. Examples of the new lithology interpretation are provided for both open and cased hole environments. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER F REAL TIME SPEED CORRECTION OF LOGGING DATA Vincent Belougne, Ollivier Faivre, Laurent Jammes and Stephen Whittaker Schlumberger Wireline & Testing ABSTRACT Data from logging devices is assumed to be acquired at regular depth intervals. ‘Depth’ is taken to be the ‘cable depth’ by the acquisition system. Of great importance to any further type of processing is the actual position of the tool along the borehole axis, i.e. the ‘true depth’. It is well known that the motion of the tool in the well does not match the motion of the cable at the wellhead and therefore the data so acquired is always in error. This is particularly true for high resolution tools with several detectors located at different vertical locations. This paper presents an entirely different way of solving this problem which allows speed corrected logs to be output in real time. Computation of the true depth is obtained via a Kalman filter combined with a fixed lag smoother. The length of the fixed lag is adjusted so as to match the performance of a standard play-back speed correction. The acquired log data is directly re-indexed with respect to this speed corrected depth. This paper also presents a more accurate method of acquiring accelerometer data resulting in increased precision of the depth estimation. Examples of speed corrected high resolution logs are presented which demonstrate considerable improvements in quality, particularly in thinly laminated formations. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER G INTERPRETATION OF RESISTIVITY LOGS IN HORIZONTAL WELLS AN APPLICATION TO COMPLEX RESERVOIRS FROM OMAN J.M.V.A. Koelman, M. van der Horst Shell International Exploration and Production B.V. Research and Technology Services, Rijswijk, The Netherlands AS. Lomas, A.T. Kolemij and JHM. Bonnie Petroleum Development Oman, Muscat, Sultanate of Oman ABSTRACT The design and interpretation of resistivity tools commonly used for quantitative evaluations has been based on the analysis of the response in vertical wells intersecting horizontally layered formations. When these tools are run in highly deviated or horizontal wells, artifacts in the response are encountered that are nor corrected for with the standard processing techniques. Using 3D forward modelling and inversion schemes we are able to recover formation resistivity profiles with associated synthetic resistivity tool responses that are in accordance with field logs obtained in highly deviated wells. Application on horizontal well resistivity logs from carbonate and clastic reservoirs in Oman demonstrate that this is the preferred way to correct laterolog and induction tool readings in highly deviated wells for the combined effects of layering, apparent dip, borehole and invasion. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER H IMPROVED PERMEABILITY PREDICTION IN CARBONATES Taras T. Dziuba ABSTRACT Permeability prediction in carbonates from well logs is a formidable problem to petrophysicists. Clastic-based permeability algorithms have not been successful in carbonates. NMR technology is addressing the challenge of estimating permeability in carbonates but the answers are still forthcoming. A new, empirical method was developed by studying over 15 different carbonate formations for which over 400 FRF and mercury injection capillary pressure measurements were available. Predicted versus core-measured permeabilities using the new relationship were tested for a permeability range of .1 to 500 mD with very good results (r =0.91). Controlling pore throat radius, pore coordination number and geometrical effects such as variations in cross-sectional area affect the permeability of carbonates. The controlling pore throat radius is measured using capillary pressures and occurs at Sw from 75 to 90%. The pore coordination number or connectivity can best be approximated by 1-Swirr, or φe/φt x 100, such that when or φe=φt the resultant Swirr will be low and the connectivity high. Permeability is also dependent on the variations in cross-sectional area as captured by the FRY measurement. The relationship then is: k = (controlling pore throat radius x connectivity)/geometry or k(mD) = (r90 x 1- Swirr )/FRF, where r90 = (pore throat radius at Sw = 90%) The relationship is easily applied to open-hole logs. Obtaining FRF from well logs, using variable ‘m’ is now fairly standard practice in carbonates. Swirr can be obtained from logs or cores but it may also be estimated by φe/φt x 100. The range of r90 values for a particular reservoir can be established by calibrating the new relationship using cores. Using FRF, capillary pressures, facies, depositional environments and Archie rock typing, the r90 value (like the cementation exponent ‘m’) can be predicted. Examples are given for both cored and uncored wells. Using the new relationship the petrophysicist can now generate a reliable, continuous permeability curve for carbonates using open-hole logs. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER I ESTIMATION OF IN-SITU STRESS PROFILES FROM WELL- LOGS Keith W. Kalahara ARCO Exploration and Production Technology ABSTRACT This paper reviews the conceptual basis for some existing methods for estimating in-situ stress profiles from well logs. Many methods make restrictive assumptions about the mechanical state of the rocks. Two common assumptions are that (1) lateral elastic strains are zero in all formations, or (2)the rock is at a yield or failure limit. Logs supply rock mechanical properties, such as Poisson’s ratio or strength, which are then used with the assumptions to compute stress. These methods can fail either because their premises are invalid or because logs are not good predictors of the relevant properties. Alternatively, there are methods that infer stress levels from breakouts or fractures at the borehole surface. Because in-situ stress directly controls the mechanical failure of boreholes, this class of methods can require assumptions that are less restrictive than the first class. These methods are best used together with additional geologic and engineering data. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER J CAPILLARY PRESSURE: THE KEY TO PRODUCIBLE POROSITY W. Scott Dodge SR ESSO Australia Ltd., Melbourne, Victoria, Australia John L. Shafer and Robert E. Klimentidis EXXON Production Research Company, Houston, Texas, USA ABSTRACT Producible porosity, defined as the pore volume available to hydrocarbon emplacement, has been computed from log measurements by modelling capillary pressure irreducible water saturation as a function of permeability and maximum hydrocarbon column height. Producible porosity has been also measured directly by NMR when the T2 relaxation time distribution cut-off is calibrated to the maximum capillary pressure in the reservoir. The producible pore volume imposes a calibration constraint on the maximum hydrocarbon pore volume that can be computed from logs. Producible porosity contains no immobile or irreducible water. Total, effective, isolated, macro and micro pore volumes are all used to characterise porosity based on specific definitions, criteria and measurement techniques. Total porosity computed from logs should match core porosity where core porosity represents the total interconnected pore volume, however, total porosity in shaley sandstone reservoirs computed from the crossplot of bulk density and neutron porosity logs has been shown to overestimate core porosity. By modelling formation mineralogy based on a calibration set and solving the log response equations through least squares inversion, total porosity from logs accurately matches core porosity. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER K EXPERIMENTAL STUDY OF DIFFUSION AND RELAXATION OF OIL WATER MIXTURES IN MODEL POROUS MEDIA D. Mardon, J. S. Gardner, G. R Coates NUMAR Corporation, Houston, TX H.J. Vinegar Shell Exploration and Production Technology Co., Houston, TX ABSTRACT Fixed gradient NMR diffusion measurements have been performed on a series of porous ceramic samples containing mixtures of water and a refined oil (dodecane) using a laboratory instrument that emulates the operating characteristics of a commercial gradient NMR well logging tool. Diffusion coefficients were computed using a standard log interpretation procedure that is based on shifts in the distribution of transverse (T2) relaxation times measured as a function of inter-echo time. An average pore fluid diffusion coefficient for partially oil-saturated samples is defined and shown to be a useful estimator of oil saturation. This saturation estimator has a potential advantage over standard NMR methods that are based on integrating relaxation time distributions because it works best at large pore sizes where light oil and water signals may not be resolved in the relaxation time distributions. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER L IDENTIFICATION OF GAS WITH THE COMBINABLE MAGNETIC RESONANCE TOOL (CMR) Charles Flaum, Robert L. Kleinberg, and Martin D. Hurlimann Schlumberger-Doll Research ABSTRACT Nuclear Magnetic Resonance (NMR) tools are sensitive to proton density of the formation, and thus to the formation hydrogen index. Therefore, the measurement can be used to detect gas, in a manner analogous to the neutron log. The first part of this paper derives the Amplitude Method to evaluate formation gas volume. It quantifies the reduction of the apparent formation hydrogen index and relates it to gas volume. It has been successfully applied in the field, and an example is presented here. In addition, presence of gas has more subtle effects on the physics of NMR measurement. A recent paper has described how a pulsed NMR measurement can detect gas in the formation, by taking advantage of the static field gradient to measure the effect of diffusion on the observed T2 relaxation time. The authors proposed the “Shifted Spectrum” method to observe diffusion from data at two different echo spacings. The method involves taking a difference between the two T2 distribution spectra, whereby the water and oil signals would cancel out, and the remaining signal would have a characteristic gas signature. It can be shown, however, that the Shifted Spectrum method can be unreliable, because of secondary distortions in the shapes of the T2 distributions. The second part of this paper describes a robust method for identification of gas via detection and measurement of diffusion, as applied to the data from Schlumberger’s Combinable Magnetic Resonance (CMR) tool. The method, called “Echo Ratio Method”, uses the ratio of smoothed measured echo trains for two passes (at widely differing echo spacing) in a fit to an analytical expression. This expression is a function of the formation apparent diffusion coefficient and the CMR field gradient distribution. The Echo Ratio method has been successfully applied to CMR data from two commercial wells. The advantage of this method over the Shifted Spectrum method can be easily seen from these results. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER M A NEW PETROPHYSICAL INTERPRETATION MODEL FOR CLASTIC ROCKS BASED ON NMR, EPITHERMAL NEUTRON AND ELECTROMAGNETIC LOGS Patrizio Gossenberg, Giuseppe Galli Agip spa Milan, Italy Marc Andreani, Werner Klopf Schlumberger Milan, Italy ABSTRACT A combination of high technology wireline logging instruments, run in an exploration well in the North Adriatic Sea, was used to develop a new petrophysical model which allows to optimize well completion for gas production. Turbiditic deposition along two directions of transport (lateral appenine and longitudinal alpine sediments) in an open piggy back basin has produced a complex sand reservoir, interbedded with silty shale beds. Petrophysical interpretation using a traditional set of logging tools gives unsatisfactory results when applied to these particular formations which comprise complex mineralogy and a changing grain size distribution. A new approach, based on a model, strongly taking into account the sedimentological environment, has the advantages of having less unknowns with respect to a classical mineralogical model and of addressing itself to the fluid distribution inside the pore system. Integration of Nuclear Magnetic, Epithermal Neutron and Electromagnetic tool responses allows the design of a model where clay, silt and sand are the matrix volumetric unknowns. Also clay bound water, capillary bound water, free water and gas volumes are computed. The resulting interpretation provides a reliable profile of potential hydrocarbon production, also in case of low resistivity environment. Silt and clay volumes are more accurately computed in gas bearing levels, where the uncertainties, using traditional log interpretation methods, are too high. Shale sequences in the reservoir can be better described by characterizing their sealing potential. Moreover, grainsize distribution and therefore bound fluid volumes can be used to discriminate between the two different turbiditic contributions. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 Production tests, repeat formation tester data and geological correlations confirm the validity of the model. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER N THE MRIL LOG IN A LOW POROSITY AND PERMEABILITY FORMATION J.R. Hook Mobil North Sea Ltd. P.B. Basan and B.D. Lowden Applied Reservoir Technology Ltd. ABSTRACT NUMAR’s Magnetic Resonance Imaging Log (MRIL) was run in a well in a Rotliegend gas reservoir in the UK Southern North Sea province. The objectives of running the log were twofold: 1. To assist in identifying the gas-water contact. 2. To investigate the reliability and accuracy of permeability determination using the MRIL log. The paper describes the results of the logging run and highlights some of the problems encountered in interpreting the data. The wireline data significantly underestimated the porosity and bulk volume irreducible water saturation in the gas-bearing section of the well. Despite this a valid gas-water contact was identified. The Rotliegend Formation in this well is composed of quartz- and dolomite-cemented aeolian dune facies. These cements, together with grain-lining chlorite and small amounts of illite have created the low permeability (rarely exceeding 1 mD) and porosity (<10%). Was the problem one of tool limitations where porosity is low or was the response affected by a gas effect or did the iron-bearing chlorite affect the signal? To assist in understanding the response of the logging tool a programme of core analysis was performed with mercury injection pore size and NMR T2 distributions forming the bulk of the data. The pore size and T2 distributions are unimodal and agree well with each other indicating that in this reservoir the concept of a genuine irreducible water saturation is not valid. The results of the core analysis are integrated with the log data to enhance the determination of the gas-water contact. Permeability estimates based on the Coates equation did not provide a good quantitative relation with core permeabilities, especially in the gas-bearing section. The geometric mean T2 from the laboratory data was found to correlate well with the core permeability. Addition 2 m of a porosity term in an equation of the form kg = AT2 φ enabled improved permeability predictions to be achieved. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER O NEW METHODS IMPROVE INTERPRETATION OF PROPAGATION RESISTIVITY DATA W. Hal Meyer Baker Hughes INTEQ Tom Maher Shell Offshore Inc. David J. McLean Shell Offshore Inc. ABSTRACT New propagation resistivity tools have better accuracy, more independent measurements (up to 32), and greater depths of investigation. However, the data processing has not advanced significantly beyond the presentation of multiple attenuation and phase difference based resistivities, As a result, the log analyst is faced with eight or more curves instead of two, which often adds to the confusion rather than reducing it. New processing schemes have now been developed to help eliminate this confusion. As none of these new schemes refer to attenuation or phase difference based resistivities. they avoid the controversy about their relative depth of investigation. The new processing routines start with the real and imaginary parts of the voltage from each transmitter to each receiver, and then proceed to one or more different output data sets, If the application requires geosteering with resistivity, then depth to the nearest bed can be displayed along with the resistivity of the zone the tool is in and the resistivity of the nearest bed. This geosteering calculation can be performed by the downhole computer to reduce the amount of data transmitted uphole. If the purpose of the application is formation evaluation, then the data can be displayed with four different depths of investigation. The four depths of investigation will be at specific depths to eliminate variable depth of investigation as a function of varying resistivity, a problem that has plagued the interpretation of propagation resistivity data in the past Another scheme calculates both resistivity and dielectric permittivity at each frequency. This method not only produces low frequency dielectric measurements, but it also eliminates the uncertainty of the dielectric constant as a source of error in the resistivity measurements. Dielectric error is often the greatest source of error in measurement of the resistivity in highly resistive zones. Using this technique, reliable resistivities have been calculated to 5,000 ohmmeters. In all cases. the raw data are stored, which allows any of these methods to be used in post-processing regardless of which method was used for the real-time data. These new processing routines have been tested using both field data and computer simulations. The tests show that these new methods can significantly improve the quality of SPWLA 37th Annual Logging Symposium, June 16-19, 1996 propagation resistivity interpretation. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER P HIGH-SPEED 2-D INVERSION OF INDUCTION LOGGING DATA A.Tabarovsky and M.B.Rabinovich Western Atlas Logging Services, Houston, Texas ABSTRACT In this real-time inversion for an array of induction sensors, high performance is achieved with an ultrafast approximation of Doll’s geometric factor and a special inversion strategy. The fast-forward modeling employs a spline approximation and a specific technique of computing the formation blocks’ responses. An iterative inversion process combines linear and nonlinear optimization. Linear optimization is used to determine formation conductivity. Nonlinear optimization is applied to adjust the geometry of formation zones. An adaptive moving window used during optimization consists of three subwindows the predictor, corrector, and upgrader. The layers are dynamically introduced, if necessary. The resolution is improved in sequential iterations because increasingly finer details are added to the previously obtained models. The final selection of parameters satisfies a priori constraints. The resulting distribution of conductivity in both radial and vertical directions produces a formation resistivity image. Interpretations of synthetic and real data confirm the viability of the method. — SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER Q A NEW METHOD TO DETERMINE HORIZONTAL-RESISTIVITY IN ANISOTROPIC FORMATIONS WITHOUT PRIOR KNOWLEDGE OF RELATIVE DIP T. Hagiwara Halliburton Energy Services ABSTRACT Anisotropic resistivity is observed in many logging environments. Although an anisotropic formation is characterized by both the vertical and the horizontal resistivities, it is often the horizontal resistivity that is most desired in log interpretation. The horizontal resistivity is the only resistivity that is measured in vertical holes. We propose a new method in which the horizontal resistivity is determined in deviated boreholes or in dipping formations through the use of induction-type resistivity measurements. The advantage of the new method is that it requires no knowledge of hole deviation or formation dip angles. The method is applicable to all induction-type resistivity measurements, including 2-MHz LWD induction-type resistivity data. The horizontal resistivity and an anisotropy response factor, which is a simple function of both the formation anisotropy and the relative dip/deviation angle, are determined from two induction-type measurements, such as R and X signals from conventional (20-kHz) wireline induction logs, or from the phase-derived and attenuation-derived resistivities from 2-MHz LWD resistivity devices. In addition, if the relative dip/deviation angle is already known, the new method is much more efficient than prior methods in determining anisotropy. Only two nomograms or computer look-up tables are needed in this method since the formation anisotropy is determined from the anisotropy response factor. The prior methods require many look-up tables. Alternatively, if the formation anisotropy is known, the relative dip/deviation angle can be derived from the anisotropy response factor. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER R A PRACTICAL DIPPING-EFFECT CORRECTION FOR MULTLARRAY INDUCTION TOOLS IN DEVIATED WELLS J. Xiao, D. Beard, and Q. Zhou Western Atlas Logging Services, Houston, Texas ABSTRACT A data processing technique including a method, database, and programmed algorithm, corrects the array induction tool logging data for dipping bed effects in the real time. The processing technique is achieved by fully utilizing the measured data of the multifrequency, multi-receiver features, which are not available with other conventional induction tools. The objective of this correction is to make the corrected logs from deviated wells have the same responses as those from vertical wells through the same formation. In deviated wells, the electric currents induced by the transmitter flow across formation boundaries, causing electric charge buildup on the interfaces. In addition, the currents pass through inhomogeneous ground loops due to the formation layers of different conductivities. The electric charge buildup influences the induction logs in both the absolute reading and the curve shape. Furthermore. the inhomogeneous ground loops deteriorate the log resolution and the log amplitude. These two types of effects are tangled with the formation conductivities, making an accurate deviated well log interpretation very difficult. The method we describe separates the dipping effects from the formation conductivities with a two-dimensional adaptive filtering algorithm, and the unwanted effects are eliminated. The filters are designed at different dipping angles for different formation conductivity contrasts and are stored in the form of a database. A special optimization algorithm is also developed for designing the filters. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER S A NEW MODEL FOR CALCULATING WATER SATURATION IN CLEAN SAND RESERVOIRS S. O. Stalheim and T. Eidesmo Statoil Research Centre, Trondheim, Norway ABSTRACT In order to calculate water saturation from Archie’s equation, the cementation factor (m) and saturation exponent (n) have to be determined. These are normally obtained from electrical measurements made on core plugs (i.e. from the formation factor and resistivity index). However, considerable scatter in the measured data leads to much uncertainty. A method is provided here for describing uncertainty attached to the determination of the cementation factor (m) from core measurements. Based on core analysis data from several North Sea wells, it is shown that m largely depends on the average area of the pore throats an observation which led to a relationship between m and permeability being established. By combining this relationship with Archie’s equation, a new model is developed which reduces uncertainty in the calculation of water saturation in clean sand reservoirs. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER T UNRAVELING THE DIFFERENCES BETWEEN LWD AND WIRELINE MEASUREMENTS Pia Hansen Maersk Olie og Gas AS, Copenhagen K, Denmark Frank Shray Schlumberger, Sugar Land, Texas, USA ABSTRACT In January 1995, a pilot hole and a horizontal well were drilled in a chalk oil reservoir in the Danish part of the North Sea. A combination of wireline services and logging while drilling (LWD) services were used. Differences in density, neutron porosity, and resistivity measurements were observed between the two logging methods. A study was conducted to investigate the measurements, Maersk Olie og Gas AS (as the operator on behalf of the Danish Underground Consortium, a joint venture between Shell, Texaco, and A. P. Møller) needed to know the true porosity and true resistivity of the reservoir. The study was also to determine the source of the differences between the two sets of measurements, and to rate the validity of the LWD measurements for formation evaluation purposes and future cost efficiencies. The differences between the LWD and the wireline density and neutron porosity measurements were the result of changes in density of the pore fluids. There had been approximately six to ten days between the LWD logging and the wireline logging. The fluid density derived by the LWD density-neutron tool was approximately 0,78 g/cm3. The fluid density derived by the wireline density-neutron tool was approximately 0.90 g/cm3. It is also shown that it was necessary to apply a standard, chartbook hydrocarbon correction to both sets of data which resulted in exactly the same true porosity from both methods. The combination of neutron and density data yielded good evaluation results, without any special empirically-derived equations, or data-value shifting. A method of moved-fluid analysis is presented, based on a comparison of fluid densities as derived from the two measurement methods. It is independent of the resistivity measurements, which are shown in this ease study to be difficult or not possible to use. It was generally observed that the deepest reading wireline induction measurement matched the shallowest LWD resistivity measurement. But, the shallower wireline induction measurement matched the LWD-obtained horizontal resistivity (corrected for anisotropy). Late time (wireline) induction measurements are shown to be affected by the combined effects of anisotropy, shoulder beds, and noncircular and varying invasion patterns. Through the use of several well examples and a review of research associated with anisotropy and migration patterns of invasion fronts, it is shown that variations in permeability anisotropy SPWLA 37th Annual Logging Symposium, June 16-19, 1996 lead to variations in invasion patterns, which in turn lead to varying effects on the wireline induction measurements. Correlation is shown between the electrical anisotropy ratio, Rv / Rh, and permeability anisotropy ratio, kh /kv. Deviated well LWD 2-MHz and wireline induction measurements must be carefully examined and corrected, if possible, for anisotropy prior to theft use for saturation calculations. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER U EXAMINATION OF MWD WIRELINE REPLACEMENT BY DECISION ANALYSIS METHODS: TWO CASE HISTORIES Gary F. Beck Chevron Production Co., New Orleans, LA, USA ABSTRACT Decision analysis (DA) methods were used to analyze the costs vs. benefits for MWD wireline replacement programs on two 6-well platform projects in the Gulf of Mexico, The DA on the first project indicated a cost savings between $0-$50,000 per well could be expected, with a median probability of saving $21,000. The wells were drilled with MWD replacing wireline with total cost savings of over $240,000 for the project. The DA on the second project indicated that there was virtually a 100% chance of not realizing any cost savings, and that we could incur additional costs of up to $70,000 per well with a median probability of an additional $29,000 cost per well, It was decided to not use MWD to replace wireline on this project, which turned out to be a prudent decision as drill pipe was stuck in several of the wells and a bottom hole assembly (BHA) was lost in one well. The common cost analysis on MWD wireline replacement is usually a simple comparison of invoice costs and anticipated rig time savings. While this type of method may yield useful information regarding cost benefits, it can be misleading as it does not take into account all of the risks and diverse factors that need be considered to evaluate the economic benefits of running MWD. Decision analysis (DA) programs are capable of incorporating variable costs, risks, and diverse factors in evaluating the possible economic benefits of running MWD. They accomplish this by performing a Monte-Carlo simulation on a range of possible outcomes and their associated costs. Comparison of one set of outcomes and associated costs (wireline logging) to another set of outcomes and costs (MWD replacement) is then possible. Since risk is incorporated into this type of analysis, a more accurate picture can be obtained regarding the possible economic benefits of MWD wireline replacement. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER V REDUCTION OF STANDOFF EFFECTS ON LWD DENSITY AND NEUTRON MEASUREMENTS G.L. Moake, LA. Beals, and W.E. Schultz Halliburton Energy Services, Houston, Texas ABSTRACT In wireline logging, standoff effects on density, photoelectric factor (Pe), and neutron measurements are reduced by forcing the density pad and neutron assembly against the borehole wall. However, this technique is not practical for logging-while-drilling devices. Consequently, a processing method has been developed that utilizes a tool-standoff measurement to emphasize small-standoff data. The standoff is measured with an ultrasonic transducer that is located on the same side of the tool as the nuclear sources and detectors. The nuclear counts are acquired in 20-millisecond intervals. The counts are immediately multiplied by a weighting factor that is appropriate for the standoff measured during the same time interval. The weight is large when there is no standoff, and it decreases rapidly with standoff. After a predefined time interval (e.g., 10 seconds), the weighted averages of the nuclear data are normalized and stored downhole in nonvolatile memory. The data is also processed down-hole to transmit real-time results, which include density, Pc, and neutron values, all with reduced standoff effects. Although standoff weighting greatly improves the accuracy of the nuclear measurements, it also increases their statistical uncertainties. These uncertainties are strongly affected by the drilling rate of penetration, as well. In order to monitor the overall statistical quality of the measurements, uncertainty curves are calculated for the nuclear measurements. In general, the increase in the statistical uncertainties of the measurements due to weighting is small compared to the reduction in systematic error that is achieved. After the data has been weighted, it is processed similar to wireline data, although several enhancements are made to the neutron porosity. The weighted standoff is used to dynamically correct the neutron log. Furthermore, measurements made with the standoff transducer and two other ultrasonic transducers, located at 120-degree intervals around the tool, are combined to make a caliper measurement while drilling. The caliper is used to apply borehole-size corrections to the nuclear measurements. As additional quality monitors, a neutron-porosity correction, density and Fe corrections are computed. The weighting technique significantly reduces these corrections, and consequently reduces potential systematic errors. Simulations based on test-formation data are used to compare this technique to unweighted, statistical, and four-quadrant processing. In addition, log examples illustrate the effectiveness of this new technique. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER W ENHANCED RESOLUTION LWD RESISTIVITY LOGS USING A NEW INVERSION TECHNIQUE G. A. Merchant Halliburton Energy Services R. W. Strickland Independent consultant C. E. Jackson Halliburton Energy Services ABSTRACT Wave propagation resistivity is the primary resistivity measurement used in logging while drilling. In high resistivity environments, the LWD logs frequently have poorer vertical resolution and more adjacent bed effect than their wireline induction or resistivity counterparts. This paper presents a new method for enhancing the resolution of wave propagation resistivity logs. The resulting logs are fully corrected for adjacent bed effects and have a vertical response as good as wireline high-resolution induction logs, Typically, wave propagation LWD tools measure the attenuation and phase shift across a pair of receivers for one or more transmitter-to-receiver spacings. The phase shift and amplitude ratios across the receiver pair are converted to apparent resistivity logs. These apparent resistivity measurements are affected not only by the formation between the receivers but also the formation between the transmitters and the receivers, and the by the beds above or below the tool. This adjacent bed effect varies with the actual resistivity value as well as the contrast. The attenuation resistivity logs have more adjacent bed effects and poorer vertical resolution than the phase resistivity logs. The method described in this paper begins by picking significant bed boundaries from the actual log. It then performs inversion by repeatedly adjusting the resistivities in a model formation until the simulated log matches the actual log. Additionally, the technique can enhance the resolution of the attenuation resistivity logs to match the resolution of the phase resistivity logs. The result is a suite of resistivity logs with matched vertical responses but different radial depths of investigation. The resulting vertical response is virtually independent of resistivity, The technique can also be applied to dipping beds and deviated boreholes. The method is illustrated with field logs. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER X THE USE OF PETROPHYSICAL DATA FOR WELL PLANNING, DRILLING SAFETY AND EFFICIENCY Phil Holbrook Sperry-Sun Drilling Services, Houston, Texas ABSTRACT The two open borehole fluid pressure limits, pore pressure (p) and fracture gradient (pf) are represented by linked stress/strain relationships which are best obtained directly from in situ petrophysical data. The key to in situ petrophysical determination of these stress fluid pressure relationships is that rock solidity (1.0-φ) is an absolute measure of granular matrix strain. - A new rock mechanics system has been developed from and for use with downhole petrophysical data related to in situ borehole fluid pressure measurements. Using gravitational force balance (σv = Sv - p), two new in situ mineralogic stress/strain (1.0-φ) relationships were derived directly from subsurface measurements of porosity on granular sedimentary rocks. These in situ compactional relationships vary with average mineral ionic bond strength and are independent of any particular material response law. Only two compaction coefficients Smax and α, are used to relate vertical stress (σv) to in situ grain matrix compactional strain (1.0-φ) in Normal Fault Regime ~biaxial basins. The compaction coefficients are weighted average mineralogic constants in a general compactional power law linear stress/in situ strain (1.0-φ) relationship; σv = σmax ((1.0-φ)α In NFR ~biaxial basins; the horizontal/vertical stress ratio (σh/σv) increases in direct proportion to in situ compactional strain ((1.0-φ) following the relationship; σh/σv = (1.0-φ) Fracture propagation pressure (fp = σh + p) is therefore also linked to in situ compactional strain (1.0-φ) and average sedimentary rock mineralogy. This new compactional strain fracture pressure relationship has been shown to be very accurate (4% SD) in 5 separate statistical studies. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER Y DIMENSIONLESS PARAMETERS FOR INTERPRETATION OF WFT DATA: SIMULATIONS AND EXPERIMENTS K. Huang, A. Samaha and E. Kasap The University of Tulsa ABSTRACT Wireline Formation Test (WFT) can provide a valuable, cost effective source of near wellbore formation data. The analysis and interpretation of WFT data are hampered by flow geometry, formation damage, lamination and anisotropy. This study introduces the use of dimensionless parameters and a 3-D near wellbore numerical simulator for a comprehensive study of problems associated with WFT data interpretation. Case studies include sensitivity to formation damage, internal probe diameter, fluid compressibility and viscosity, formation permeability, drawdown tank volume, formation heterogeneity and anisotropy. The results of numerical simulations indicate that the dimensionless parameters of WFT can be very useful to identify lamination and anisotropy. In addition to the numerical simulations, an experimental apparatus which models downhole measurements of a damaged formation was built and attached to a probe permeameter equipment. These experiments enabled us determine the dependence of the measured permeability of the damaged formations on the severity and the depth of damage. The results of the experiments are in good agreement with the 3-D numerical simulator results. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER Z SUPERCHARGE PRESSURE COMPENSATION WITH NEW WIRELINE FORMATION TESTING METHOD Mark A. Proett and Wilson C. Chin Halliburton Energy Services ABSTRACT Wireline formation testers are increasingly being used to obtain formation pressure and permeability measurements, which in turn are used in geological evaluations and in designing well completions. It is generally assumed that the pressure measured by the tester probe is near formation pressure. Unfortunately, in many instances, such as in low permeability formations, the mudcake does not adequately isolate the hydrostatic pressure in the wellbore from the formation. As a result, the sandface pressure measured with the formation tester is “supercharged” and not the formation pressure after all. This sandface pressure may be as much as 1,000 psi over the actual formation pressure of interest. This paper demonstrates that a new analytical model can be used along with a new testing technique to correct the measured pressure to actual formation pressure. A high accuracy finite element near-wellbore simulator (NEWS) is used to test the sensitivity of the analytical model. Log examples are presented to demonstrate the success of the new testing technique. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER AA A NEW, SIMPLIFIED, UNIFIED TECHNIQUE FOR THE ANALYSIS OF WIRELINE FORMATION TEST DATA K Kasap University of Tulsa, Tulsa, Oklahoma D. Georgi, J. Michacis and T. Shwe Western Atlas Logging Services, Houston, Texas ABSTRACT Wireline Formation Test (WFT) data routinely are used to determine the undisturbed reservoir pressure, P*, and its vertical gradient, and to estimate a formation’s near-wellbore permeability. Different techniques for analysis have been borrowed from the well testing arena and adapted to WFT drawdown and buildup data. Spherical-flow analysis utilizes early-time data and usually gives a reliable estimate of permeability. For P* determination, cylindrical-flow analysis is preferred because it focuses on late-time buildup data. However, the cylindrical-flow analysis has its drawbacks, Late-time data is crucial for cylindrical-flow analysis, especially in low permeability formations, but long testing periods are not desirable because of potential tool ‘sticking” problems. Even on long tests, the cylindrical-flow period may not occur or may not be detectable on WFT’s. Even when it does occur, permeability estimates derived from the cylindrical-flow period may be incorrect and their validity difficult to judge. We introduce a new analysis technique that simplifies the interpretation of WFT pressuretransient data. It utilizes the geometric factor concept for modified hemispherical flow, which has been used successfully in probe permeability measurements. The technique is based on a pressure vs. pressure-time derivative analysis of WFT buildup data. The technique is less sensitive to data quality than other methods and can be implemented with a simple graph from which both near-wellbore permeability and P* are readily determined. We applied conventional WFT analysis (pseudosteady-state drawdown, spherical and cylindrical buildup), and the new buildup analysis techniques to six sets of field data. Calculated permeabilities varied widely among the various techniques in some of the tests. To better understand and evaluate these results, we conducted sensitivity studies with a 3-D near wellbore, numerical-flow simulator. Simulation results indicate that the new technique consistently provides better estimates of permeability than the conventional methods. P*, which can be derived from less data with this technique, is as good or better than P* from other analysis techniques. It matches the field-measured formation pressure and does not require late-time buildup data. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER BB SANDWICH: LOG EVALUATION IN LAMINATED SHALY SANDS Frans G. van den Berg Sarawak Shell Berhad Wim J. Looyestijn and Robert K.J. Sandor Shell International Exploration and Production B.V. SUMMARY Formation evaluation from well logs in laminated shaly sands has often suffered from lack of proper models, and in many cases led to underestimation of oil or gas in place. Shell Research have developed a technique that allows consistent analysis of formations with laminated, dispersed and structural shale, leading to better quantification of oil and gas volumes in place. Present research is concentrating on further calibration of the model input parameters by core and high-resolution logs such as borehole imaging logs. The present paper shows results of application of the model to the oil fields in Sarawak Shell Berhad (SSB) in Malaysia. The results show increases in the hydrocarbon volumes of up to 40%, as compared to the previous conventional analysis. Higher increases were obtained in environments with higher amount of laminated shale. The main benefit of application of the SANDWICH model was the increased confidence in the volumes of oil and gas in place in laminated shaly sands. This was achieved through reduction of the uncertainty in the log derived inputs to volumetrics. As a result, the reduced uncertainty has helped to reconcile reservoir production with oil in place, and it will improve further development planning such as re-developments and re-completions. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER CC ESTIMATION OF GAS/OIL RATIOS AND DETECTION OF UNUSUAL FORMATION FLUIDS FROM MUD LOGGING GAS DATA Alan C. Wright Texaco Exploration and Production Technology Department, Bellaire, Texas, USA ABSTRACT Equations for predicting gas/oil ratios (GOR) from gas compositions measured during mud logging were developed by use of regression techniques applied to a data base of total reservoir fluid analyses. Based on the equation standard deviations, GOR can be predicted to within a factor of 3.0 with 90% confidence when methane through pentane are accurately known, and to within a factor of 4.6 when methane through butane are known. Application of four published fluid typing methods to the same data base resulted in GOR correlations of lesser accuracy. Because of variations in reservoir genesis, it is likely that the equations presented here provide close to the best GOR predictive accuracies which can be expected when gas composition alone is considered. Correlations were also developed which permit flagging of unusual gas compositions or analytic errors, Oil gravity was found to be only loosely correlated with gas composition; no useful predictive equation was the purpose of reservoir fingerprinting and/or detection of analytic errors. For all of the developed equations, an important goal was the determination of reliability factors as well as predicted values so that the results might be realistically applied. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER DD NEW PRODUCTION LOGGING TOOL ENABLES PROBLEM WELL DIAGNOSIS: A CASE STUDY Michel Didek and Bruno Pedron Elf Aquitaine Production Fadhel Rezgui, André Abriol, Jean Pierre Yver, and Bruno Deruyck Schlumberger ABSTRACT The new production logging tool discriminates between water and hydrocarbons (oil and gas) by means of four identical probes symmetrically positioned across the wellbore. Each probe is a bubble detector that yields one of two outputs, depending on whether it “sees” water or hydrocarbons. Because of its binary nature, the measurement readily provides water holdup, without prior knowledge of water, oil or gas densities required for the calibration of a standard gradiomanometer. Unlike the gradiomanometer, the new measurement does not have to be corrected for friction effect or well deviation. Because of its local nature, the new measurement provides an image of the distribution of segregated fluids across the wellbore. This feature can help in understanding segregated flow regimes, especially in deviated wells. The bubble count (number of dispersed fluid bubbles observed per second) is very sensitive in detecting first oil entries and individual perforation contributions. The use of the new measurement is illustrated by a field example where a leaking bottom plug was suspected to be responsible for the well’s high water cut. It unequivocally identifies the source of water production where conventional sensors would have been inconclusive and provides the justification for a workover that resulted in blocking water entries and increasing oil production by a factor of three. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER EE FIELD TESTS OF AN ACOUSTIC LOGGING-WHILE-DRILLING TOOL IN VARIOUS BOREHOLE ENVIRONMENTS Dale R. Heysse, Carl Robbins, and John Minear Halliburton Company ABSTRACT Acoustic logging-while-drilling represents very recent advances in formation evaluation, and the industry is still learning the limitations of this new service. Field tests began in the third quarter of 1993, and have just been completed on a sonic LWD tool, during which over 60,000 ft of hole were drilled and logged. These tests were in boreholes drilled with 8.5-inch and 12.25-inch bits, in the United States and in the North Sea. These tests illustrate the effects of various borehole environments, including water-based and oil-based muds, borehole enlargement and rugosity, and various lithologies. Formation slowness ranged from 55 µsec/ft to 170 µsec/ft. Some formations also exhibited shear waves, which raises the possibility of shear wave logging and rock mechanics evaluation while drilling. This paper will show several examples of acoustic LWD waveforms and processed logs, in a range of environments. It will also discuss limitations in the measurement and tool reliability during the field tests. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER FF HOSTILE ENVIRONMENT FULL-WAVE SONIC LOGGING PERMITS FRACTURE EVALUATION IN HIGH-TEMPERATURE SLIMHOLE WELLS Paul Boonen and Steve Flowers Halliburton Energy Services ABSTRACT Slimhole drilling and deep-hole drilling have become increasingly important methods in the petroleum industry. Small-diameter, 2 3/4-in., full-wave sonic tools can record acoustic waveform information under hostile downhole conditions up to 500°F (260°C) and 25,000 psi in deep boreholes. All applications of acoustic tools, based on the availability of compressional, shear and Stoneley waves are now possible in hostile conditions. Acoustic waveforms are excellent natural fracture indicators. Instantaneous Waveform Characteristics (IWC) and Stoneley wave reflection processing are techniques to enhance the identification of fractured intervals. IWC is a complex trace analysis system inherited from surface seismic processing. This processing separates the acoustic energy down into its transmissivity, phase, and frequency components. Stoneley waves are strongly attenuated by fractures intersecting the borehole. Stoneley wave reflection logs show these fractures as chevron patterns in the filtered waveforms. Log examples compare natural fracture identification logs from the highly fractured reservoirs of the bay of Campeche (Mexico) to the unfractured carbonates of the similar formations in deep wells in the state of Tabasco (Mexico). Instantaneous Waveform Characteristics (IWC) and Stoneley wave reflection processing are used to enhance the identification of naturally fractured intervals. Several different full-wave fracture identification methods are compared. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER GG PERMEABILITY FROM WAVEFORM SONIC DATA IN THE OTWAY BASIN Andrew Buffin Santos Ltd., Adelaide, Australia ABSTRACT The early Cretaceous Pretty Hill sandstone reservoir of the onshore Otway Basin, Australia, is characterised by facies changes due to the depositional environment of the Pretty Hill. This leads to compartmentalisation and hence low contrast pay, with zones of low permeability gas and high Swirr in the different facies, in addition, bore holes generally suffer from breakout and well bore ovality. Waveform sonic data are processed to provide compressional slowness, shear slowness and Stoneley slowness. Past studies have shown that the use of compressional slowness and shear slowness in the form of a Vp/Vs ratio can identity gas bearing sands in the Pretty Hill reservoirs, now it can be shown that a good correlation between permeability measured in core data and a permeability index computed from the Stoneley component of the full waveform sonic data is also possible. The measured Stoneley slowness is known to be sensitive to permeability whilst a synthetic Stoneley slowness assumes no permeability, therefore the ratio between the two generally indicates permeability. The technique used to obtain the synthetic Stoneley slowness is a simple cross plot correlation of the Stoneley with shear and density derived within the impermeable units, however when deriving a Stoneley permeability index it should also be considered that the Stoneley waves are affected by hole size, mud slowness, the formation and toot characteristics. This case study evaluates the simple Stoneley correlation technique and its application. Core, Formation Tests and Drill Stem Test data were used firstly to calibrate the ratios from Stoneley and secondly as ground truth. The permeability of the sandstone reservoir units vary greatly over very short vertical and areal distances, knowledge of their permeability can affect the completion strategy. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER HH FRACTURE HYDRAULIC CONDUCTIVITY ESTIMATION FROM BOREHOLE STONELEY WAVE TRANSMISSION AND REFLECTION DATA X.M. Tang Western Atlas Logging Services, Houston, Texas ABSTRACT Permeable fractures in reservoirs are important structures in the exploration and production of hydrocarbons, and hydraulic conductivity is of primary importance in the characterization of fractures. The hydraulic conductivity is defined as κ0H/µ, which is integrated fluid mobility (ratio of permeability κ0 to viscosity µ) over a fracture zone of thickness H. A method for data processing and analysis is developed that obtains the conductivity of borehole fractures from array acoustic waveform data containing low-frequency Stoneley waves. This method consists of three major procedures: wave separation, wave modeling, and fracture hydraulic conductivity estimation. The first procedure separates Stoneley waves into transmitted and reflected wavefields. The second procedure corrects the effects of borehole irregularity (e.g.. enlargement, washout, etc.) on the Stoneley waves; it accomplishes this by using caliper and Stoneley slowness logs to numerically model Stoneley wave propagation. The final procedure uses both the measured and synthetic Stoneley wave transmission and reflection data to estimate the fracture’s hydraulic conductivity. A theory for Stoneley wave propagation across fractures and washouts is used in the estimation. In this new technique, the correction for borehole irregularity removes the effects unrelated to fracture permeability and the use of both transmission and reflection data constrains the estimation to yield valid conductivity values, so that the process is effective in assessing borehole fracture fluid-transport properties from Stoneley wave data. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER II HYDROCARBON EVALUATION THROUGH MODULUS DECOMPOSITION OF SONIC VELOCITIES IN SHALY SANDS Vimal Saxena Oil & Natural Gas Corporation, India ABSTRACT Borehole shear waves are known to provide lithology calibration of formation by combining its velocity attribute with compressional signals. Acoustic velocities are affected by formation porosity, fluid saturation and clay presence. While the porosity effect is well understood in clean sands, efforts have been made in past to analyze sonic velocities through modulus decomposition. It has shown promising prospects for gas evaluation in predominantly clean sands. The analysis of oil saturation is not as promising, as the effect of oil is more subdued in nature and is sometimes negated by shale influence. Additionally, little work is available for modulus decomposition in shaly sands. There is a need to interpret elastic moduli in oil bearing shaly sands for petrophysical understanding. The present study incorporates exact results of Generalized Gassmann‘s Theory for the combination of two porous constituents. This allows to compute the dry frame bulk modulus (Kd) for dispersed shale in sandstone matrix. Its interpretation with Vp/Vs clearly indicates under-estimation of dry frame moduli in hydrocarbon bearing zone. The study establishes a general relationship between shear and bulk grain conmpressibilities, which permits to estimate corrected Kd. Although this in-turn allows to compute pore fluid modulus from Gassmann‘s Theory, the gas saturated live oil indicates the restriction for using simple mixing law for elastic moduli. However a comparison of computed and corrected grain bulk compressibilities, provides a sonic derived saturation parameter. It creates an exciting prospect for resistivity independent oil evaluation in shaly sands. The technique has been successfully applied in freshwater shaly sands reservoirs of Upper Assam. Results have been compared with testing data and computed oil saturation from resistivity based interpretation, and are found consistent. The technique even allows to separate gas and oil bearing zones, The complexity of actual mineralogy, assumed elastic properties of shales, non-applicability of simple mixing law, depth of investigation of sonic signal and permeability are the major influencing factors for hydrocarbon detection through this technique. However, it has been successful in identifying oil zones in shaly sands and separating gas zones. The study opens fresh possibility of hydrocarbon evaluation in cased holes. With the availability of more laboratory studies for elastic moduli in shaly sands and validation of basic results using dead oil, further improvement is expected in sonic derived saturation modeling. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER JJ DEVELOPMENTS IN UNDERSTANDING THE PHYSICAL FOUNDATIONS OF FORMATION DENSITY AND LITHOLOGY LOGGING Daniel C. Minette Baker Hughes INTEQ, Houston, TX ABSTRACT Over the last few years, the fundamentals of formation density measurement have been questioned. Theoretical papers have reported finding substantial errors in the rib-spine technique. While these theoretical arguments appear to be compelling, they are not reconciled with experimental measurements. Indeed, numerous logs have been run by many different service companies utilizing the rib-spine technique, providing measurement accuracy that contradicts these theoretical predictions. This variance between theoretical predictions and experimental results is reconciled through an improved understanding of the density measurement. The foundation of this understanding is the basic physics underlying gamma-gamma density and lithology measurement. Two fundamental papers exploring the physical basis for density measurement will be reviewed (Tittman,1965)(Bertozi,1981). The conclusions arrived at in these works will be expanded and explored through the use of Monte Carlo modeling. After the basic physics is considered, parameter affecting density logs will be addressed. The parameters listed by several authors will be explored in some depth. This will allow us to translate what we know about the basic physics to a practical logging environment. Having done this, the basic density response, in the absence of mudcake will be defined. The equations used by a number of authors to obtain formation density from detector count rates will be presented and understood in terms of the basic physics and the parameters that affect the log. The Pe measurement will then be considered. Both the standard measurement of Pe and a proposed Pem will be presented. Finally, density compensation will be explored. Newly developed techniques, as well as the standard rib-spine technique will be explored. The emphasis in this review will be on experimental validation of the methods as well as theoretical soundness. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER KK MORE ANSWERS FROM PRODUCTION LOGGING THAN JUST FLOW PROFILES DeWayne R. Schnorr Schlumberger Wireline & Testing ABSTRACT Production Logging for the most part is used to determine the production profile of a well that has multiple zones producing. This production profile is for a single surface production rate and down hole flowing pressure. But much more than that can be gained from the data if the procedure for acquiring data is altered. The additional information gained is productive index, PI, and reservoir pressure for each of the producing layers. This information cannot be determined from a pressure buildup or drawdown test when more than one zone is commingled in a single production string. The production log data, spinner and pressure, are acquired at several different surface production rates. This data is then used to determine the PI and reservoir pressure for each producing layer. This can be done with or without shutting in the well. The PI of a zone determines how much production will be produced for every pound that the pressure is reduced. It also determines the reservoir pressure for each layer. First, the PI is directly related to the deliverability of the zone and also the damage factor skin effect. It is a quick way to determine if a zone is damaged due to drilling, completion, or over time due to production. Second, the reservoir pressure determination of each producing zone will determine if different reservoirs are in communication with each other. It can also determine if parts of the producing intervals are being supported by some type of pressure support, water drive, while other producing intervals are not being supported, only a solution gas drive. Four case histories are presented where the PI and reservoir pressures were determined from production logs. These four examples each determined information about the reservoirs that was not known or not determined from other evaluation methods. This information was then used to improve the recovery factor for each of the wells. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER LL A NEW LOW-ENERGY GAMMA RAY TOOL FOR FULLBORE MEASUREMENT OF GAS HOLDUP IN A CASED WELL Margaret C. Waid, William P. Madigan, Harry D. Smith Jr., and Rafael B. Vasquez Halliburton Energy Services AB STRACT The measurement of gas holdup in a cased-hole environment is a fundamentally difficult problem. Traditionally, an estimate of the percentage of gas in a cross section of a wellbore has been computed from fluid density measurements. These estimates are often inadequate for gas holdup in horizontal or highly deviated wells since the fluid density measurements are not fullbore measurements. To accommodate the need for increased fullbore accuracy over a greater range of applications, a new Gas Holdup Tool (GHT) provides a more accurate technique of obtaining these gas holdup values directly during logging. The GHT tool is a 111 /16-inch-OD through-tubing production logging device used to determine the volumetric fraction of gas over a cross-sectional volume element of the wellbore. The new tool operates in horizontal, highly deviated, and vertical cased wells and generates a 0-to-100% gas holdup log in stratified or uniform flows. The tool requires a low-energy cobalt-57 source and a sodium iodide detector located a short distance from the source and separated from the source by a tungsten shield. A backscatter technique is used to accurately measure density differences between fluids and gases. The measurement consists of counting gamma rays scattered back from the production fluid to the detector, the count rate being associated with the gas volume fraction (holdup). The measurement is not affected by the composition and density of materials outside the casing. Monte Carlo modeling (using MCNP) and experimental data collected over a range of casing sizes at standard temperature and pressure conditions are used to validate the model and establish empirical relationships leading to the association of count rates with the gas holdup in different fluids and casing sizes. The sensitivity of the measurement to other parameters such as fluid type (oil, fresh water, salt water) and casing size are investigated. Correction is made for dead time, and compensation is made for downhole PVT conditions. A calibration algorithm and design of a calibration fixture and portable wellsite verifier are based on the relationships obtained from MCNP modeling and data collected. An algorithm based on these relationships results in a log that reports gas holdup and does not require the traditional post processing of fluid density logs. A log example, generated in a controlled experiment simulating dynamic conditions of horizontal stratified flow, is given to illustrate the application of the new tool as a fullbore gas holdup tool for logging horizontal wells. A log example from a vertical well logged by the initial prototype of the GHT tool is discussed. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER MM FIELD EXPERIENCES WITH A NEW COMPENSATED NEUTRON TOOL R-C. Freitag, M.W. Mickael, J.C. Koudelka, and W.A. Gilchrist, Jr. Western Atlas Logging Services, Houston, Texas ABSTRACT A new compensated neutron porosity instrument has been developed, and is currently under field test. The design of the new instrument was optimized to obtain high statistical precision while maintaining good porosity sensitivity and small environmental effects. The tool was designed and characterized with high-precision state-of-the-art Monte Carlo modeling. The tool response and environmental corrections were then verified by laboratory and field data. This approach represents a significant change in tool-design philosophy, allowing more costeffective and stream-lined research. The high counting rates of the detectors and the enhanced ratio-to-porosity sensitivity of this instrument provide accurate porosity response over the entire porosity range. especially in shales and shaly formations. The large set of modeling data used to characterize the tool also makes all environmental corrections more accurate than those produced for older-generation tools. Specifically, the small borehole salinity effect on this measurement makes interpretation easier and more accurate in high-porosity, high-salinity environments. To document the extent and accuracy of the tool characterization, examples from various logging environments are presented. Openhole data, acquired in boreholes from 6 in. (15 cm) to 12.25 in. (31 cm), covering a wide porosity range of both carbonate and clastic formations, are compared to core data with good results. Lithology information derived from neutron/density and neutron/acoustic crossplots compared favorably with core mineralogical information. These results show the effectiveness of the design technique and also help validate the accuracy of the tool response. Cased hole data for single- and dual-casing strings are also presented. Comparisons of these measurements with other available data show the consistency and applicability of the casing corrections derived through numerical modeling. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER NN COMPARATIVE EVALUATION OF CORE COMPACTION CORRECTIONS FOR CLASTIC RESERVOIRS Paul F Worthington Gaffney, Cline & Associates, UK Roll K Bratli, Rune Nicolaysen Saga Petroleum as, Sandvika, Norway Jeremy M Daines Gaffney, Cline & Associates, UK ABSTRACT Traditional methods for effecting core compaction corrections are compared with an iterative procedure that characterises the reservoir stress equation through convergence of the Biot alpha parameter. The iterative approach requires a modest database of elastic properties, a knowledge of pore pressure, and an insight into the prevailing major components of stress in the reservoir, The output is a mean effective stress that is based on a substantive description of the stress regime in situ and is therefore considered to be more definitive than corresponding data derived from regional approximations. The mean effective stress is used as a basis for establishing improved algorithms to correct petrophysical core data from ambient to reservoir conditions. Against this information, it has been possible to benchmark the more traditional methods of overburden correction and to assess the degree to which these can be accepted. The benchmarking draws upon data from a North Sea field. on the basis of which a generic methodology is proposed for implementing core compaction corrections. This method is demonstrably superior to conventional approaches, some of whose underlying assumptions are shown to be unsupported. Once established for a given field situation. the method has a straightforward application to diverse petrophysical properties during the course of integrated reservoir studies. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER OO ACCURACY OF SORW AND SORG MONITORING IN A MIDDLE EAST CARBONATE RESERVOIR Nick A. Wiltgen, Sultan A1-Merikhi and Ibrahim J. Al-Othman QGPC ABSTRACT Accurate knowledge of oil saturation and its distribution is essential for any reservoir being considered for redevelopment or for secondary and tertiary operations. Determination of oil saturation forms the only rational basis for prediction of recoverable oil reserves and it is critical in the selection, design and economic evaluation of an oil-recovery process. As technology advances, more of this oil will come within our reach. However, it is essential that we understand the reservoir involved so that proper, economically viable technology can be applied. Thus, the importance of residual oil saturation (SOR) which in turn is the basis for the estimation of remaining oil-in-place. Five techniques to determine remaining oil saturation are currently used. They are coring, logging, chemical tracers, reservoir engineering studies and pressure transient tests. Normally, two or more of these methods are used together and the data evaluated as a whole since each technique has its own advantages and limitations. When evaluated together, a clear understanding of the total reservoir can be obtained. Conventional logging tools and procedures have been used extensively for many years for determination of reservoir and fluid properties. Although SOR determinations normally require a higher degree of accuracy than is commonly achieved by conventional logging methods, the incentive to use logs for SOR measurements remains high. They are often cheaper and easier to use than is the case for other methods which are currently available for measuring SOR. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER PP NMR LOGGING OF MIXED-WET NORTH SEA SANDSTONE RESERVOIRS T. Eidesmo, O.M. Relling and H. Rueslåtten Statoil Research Centre, Trondheim ABSTRACT During the last year, a number of wells in the North Sea have been logged with NMR logging tools (NML) and a large database has been established to evaluate the potential of NML for interpretation of permeability, porosity and fluid saturation and their distribution in reservoir zones. NML data from three wells in a mixed wet sandstone reservoir, from the prolific Brent Group in the Northern North Sea, have been evaluated and compared with conventional logs. Furthermore, a comprehensive experimental programme was set up and selected core samples from the same logged wells were analysed for calibration purposes. Despite the rather low signal-to-noise ratio of the NML data, the result demonstrate that it is possible to obtain rather good porosity and permeability estimates for brine saturated rocks. On the other hand, in the oil saturated zones it is more difficult to estimate these parameters. This result is in accordance with earlier laboratory studies of mixed-wet rocks where it has proved difficult to identify the various fluid phases and to estimate permeability from NMR relaxation data. This problem has been addressed, particularly bearing in mind the rather low signal-to-noise ratio of the NML data. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER QQ INVESTIGATING THE UTILITY OF NMR SPECTROSCOPIC TECHNIQUES FOR FRACTURE CHARACTERIZATION C.T. Philip Chang, Jinli Qiao, Songhua Chen and A. Ted Watson Texas A&M University ABSTRACT Methods for identifying and characterizing fractures in reservoirs by well logging can be very valuable. In this paper, we explore the feasibility for using various nuclear magnetic resonance (NMR) methods for fracture characterization. We present analyses of experimental data obtained with a variety of samples. We show that relaxation contrasts are useful for separating the signal contributions from fluids in the fractures and the porous matrix, and describe how relaxation weighting can be used in combination with other NMR techniques for enhancing fracture characterization. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER RR IMPROVED NMR WELL LOGS FROM TIME-DEPENDENT ECHO FILTERING Carl M. Edwards and Songhua Chen Western Atlas Logging Services, Houston, Texas ABSTRACT NMR well logging using permanent magnets, although still in its infancy, is rapidly becoming a standard service of most logging programs. Estimates of effective porosity, and the fractions of bound and movable fluids are derived from NMR relaxation measurements. Recent advances in acquisition and interpretation techniques have demonstrated that the presence of hydrocarbons can be detected, and, in some instances, quantified when relaxation distributions from NMR data acquired with different logging parameters are compared. An accurate representation of relaxation distributions is the key to reliably identifying and quantifying hydrocarbons. The required accuracy can only be achieved through the optimization of the signal-to-noise ratio (SNR) of the spin-echo data; a poor SNR smears out or blurs estimates of the relaxation distributions, as well as making them unstable. To improve the SNR, a running average is typically employed, but it adversely affects the depth resolution of the log. Regularization is employed to improve the stability of the estimates, but it also adversely affects the resolution in relaxation time of the distributions. A filtering technique has been developed that improves the SNR of the slowly relaxing components in the spin-echo data. This technique also reduces the effects of noise rectification when phase-insensitive relaxation data are used. The filter is implemented in the time domain on the unprocessed phase-sensitive data. It takes advantage of the fact that the signal is the sum of decaying exponentials and that components with short relaxation times have decayed away after a time greater then 3T2. Thus, the filter’s bandwidth decreases with increasing time. Early-echo, large-signal data are minimally filtered while late-echo, smallsignal data are filtered more heavily. A proper choice of filter bandwidth as a function of time within the relaxation decay maximizes noise reduction and minimizes distortion of the underlying decay curve. Using a simple boxcar-averaging technique. signal distortion is less than 2% of echo amplitude. We demonstrate with both synthetic and field data the benefits of echo filtering. The inversion of filtered data increases the resolution of calculated T2 distributions. At the same time, it reduces the effect of noise rectification that is characteristically observed on NIMR logs as an excess of the bulk-volume movable fluid. Fourier analysis demonstrates that the bandwidth of the noise power spectrum is reduced by a factor of four. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER SS DETERMINATION OF OIL SATURATION FROM DIFFUSION NMR LOGS W.J. Looyestijn, Shell International Exploration and Production B.V, The Netherlands ABSTRACT Diffusion NMR on fluids contained in porous rock can be interpreted quantitatively in terms of oil saturation. This is a new application, which has proven to be very useful to supplement conventional saturation evaluations. This new method utilises the differences in molecular self diffusion between oil and water. The application requires at least two T2 decay measurements taken with different interecho spacings in a gradient magnetic field. The method has been applied successfully to the data recorded in wells that contain heavy oil. The NMR-derived saturations are in good agreement with core and log data, and have the advantage that core calibration is not required for each field; in this case available core data verified the method. A sensitivity analysis showed that saturations can be derived with an uncertainty of some 5 to 7 percent pore volume, provided that the oil and water diffusion coefficients differ by at least a factor of five. The uncertainty in saturation is mainly due to the noise level of the data, and may thus be further reduced by improvements in hardware. The NMR measured saturation is at a very shallow depth (typically 10 cm from the borehole) and is only representative of the virgin saturation if no, or only very shallow, invasion has taken place, or if the interval is at “residual” oil saturation, e.g. after water flood. Interpretation of the ‘standard’ T2 data for distinguishing between bound water and movable fluid (water and/or oil) failed in the presented cases because of the short relaxation time of the viscous oil. Numerical simulation has shown that under most conditions encountered in sandstones the Free Fluid Index concept remains valid for oils with T2 greater than 50 ms, which corresponds to in-situ viscosities less than 20 cP. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER TT SELECTION OF OPTIMAL ACQUISITION PARAMETERS FOR MRIL LOGS Ridvan Akkurt NUMAR, New Orleans, LA Manfred G. Prammer NUMAR, Malvern, PA M. Andrei Moore Shell Offshore Inc., New Orleans, LA ABSTRACT The recently introduced NMR-based logging techniques, such as the Differential and Shifted Spectrum Methods and the Time Domain Matched Filter Analysis, have added hydrocarbon typing and calculation of near-borehole water saturation to the suite of available NMR applications. The key to the new applications are the exploitation of the relaxation and diffusion properties of reservoir fluids by utilizing the single-valued magnetic field gradient and depth-of-investigation of the Magnetic Resonance Imaging Logging tool (MRIL). Reservoir, fluid and borehole properties determine the optimum mode of operation for the MRIL. Temperature, pressure, hydrogen index, oil viscosity mud type and invasion characteristics can impact the information available from the log. This complexity, considering the variety of the answer products available from the MRIL log, requires the careful selection of optimal acquisition parameters based on expected logging conditions. The objective of this paper is to establish the background to develop basic guidelines that can be used to identify and screen particular applications for reliable and robust MRIL-only interpretation. The mechanics of selecting the optimal acquisition parameters is demonstrated for a DSM application in the Gulf of Mexico, where the primary objective is the detection and quantification of free gas by relying on MRIL as the primary log. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER UU AN ACCURATE METHOD FOR WATER SATURATION EVALUATION BASED ON ADVANCED THEORY OF ELECTRICAL CONDUCTIVITY OF THE TERRIGENOUS ROCK VS. Afanasyev and S.V. Afanasyev Research and Production Center for Geophysical Operations, Tver, Russia ABSTRACT In practical well log interpretation of terrigenous sections, Waxman-Smits and dual-water models are used to evaluate water saturation (Sw) of the rock. These approaches need a whole range of corrections in the model parameters, first of all the structural coefficient m. Different researchers use different approximations of m as a function of Qv. The authors of this paper, on the basis of detailed studying sandstone and shale conductivity experiments published in the world, as well as their own data, found out new, never established earlier regularities in forming the relationship C = f(Cw). Taking these data into consideration, a new model of terrigenous rock conductivity was developed and presented at the 15th European SPWLA Symposium in 1993. This paper provides new data for theoretical validation of the electrical conductivity model. It is compared with the Waxman-Smits and dual-water models. The study showed that the new model more accurately describes the electrical conductivity of electrolyte, Cel, in the pore space. That gives an opportunity to assume m = const ~1.7 in the model. It was also established that the Waxman-Smits and dual-water models gave the very Sw values obtained with the new model if their m values were properly adjusted. However, when processing specific reservoirs, the latter procedure appears to be a source of uncertainty and excessive errors in Sw evaluation. Applying the new model forms a basis for developing a reliable method for the evaluation of water saturation of the terrigenous rock. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER VV APPLICATION OF MAGNETIC RESONANCE LOGGING IN SAN JORGE BASIN (ARGENTINA) A. Corbelleri Western Atlas Logging Services, Canadon Seco, Argentina D. Lenge, J. Ortea, and E. Breda YPF S.A., Argentina ABSTRACT Magnetic Resonance Imaging Log (MRIL) data are providing formation evaluation information heretofore not available in the Golfo San Jorge basin. Argentina, and are contributing to increased well productivity. This paper deals with log interpretation problems associated with conventional tools, and the solutions provided by the MRIL service. We briefly explain the geology of the basin and the principal problems encountered in the evaluation of the conventional logs. We then describe an empirical MRIL interpretation methodology and compare the predictions with production results. In the San Jorge basin, the sandstone reservoirs contain a high percentage of diagenetic clays, volcanic lithic fragments in a tuffaceous and argillaceous matrix. This complex lithology influences conventional porosity and resistivity tools, so despite extensive analysis with conventional logs, completion strategies remain costly and uncertain. Conventional porosity and resistivity tools are heavily influenced by the complex lithology. The MRIL service, based on the principle of nuclear magnetic resonance and relaxation, provides critical petrophysical formation evaluation data. Without relying on radioactive sources, it offers a mineralogy independent, effective porosity. It provides a measure of free fluid, irreducible water saturation, and pore-size distribution. From these data, permeability is derived. The MRIL permeabilities are compared with core and production data. Finally, the MRIL data have permitted us to identify sands suitable for fracture stimulation. The predicted results compared well with actual production data. The MRIL service has improved the accuracy of productivity predictions and increased hydrocarbon production. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER WW IMPROVED RESERVOIR CHARACTERIZATION THROUGH CROSS- DISCIPLINE MULTIWELL PETROPHYSICAL INTERPRETATION C. Corbett GeoQuest, Houston, Texas, USA I. S. Plato GeoQuest, Stavanger, Norway G. F Chalupsky GeoQuest, Dubai, United Arab Emirates Robert J. Finley Bureau of Economic Geology, The University of Texas at Austin, Texas, USA ABSTRACT Multiwell petrophysicai interpretation provides improved consistency in the evaluation of log-derived reservoir characteristics. This principle is demonstrated by comparing the results of two independent petrophysical interpretations to a common set of seismic attributes. Seismic atuibutes have recently been employed to distribute log properties measured at the wellbore to the farthest extent of a seismic survey. This distribution is based on a relationship between the seismic attributes and the log properties for which a special emphasis may be placed on a petrophysically derived effective porosity. This paper demonstrates the critical importance of an accurate petrophysical interpretation of the available log data to derive such a relationship. A dataset covering the Frio formation in part of the Stratton Field, onshore South Texas, included 3D seismic data and well logs for several wells. The petrophysical data from these wells were subjected to two independent interpretations; one used a standard single-well deterministic dual-water model, whereas the other employed a multiwell multi-mineral analysis in an optimizing routine. The resulting effective porosity was then averaged over a geologically determined zone of interest. The averaged effective porosity was then compared to the available seismic attributes in both cases. In the single-well petrophysical method the relationship between seismic attributes and effective porosity was obscured. However, the multi-mineral petrophysical model provides a clearly defined relationship to the seismic attributes. Such cross-discipline evaluation of the reservoir serves to validate the petrophysical interpretation. The results provide a mechanism for seismic-guided or geostatistical reservoir property distribution, with an enhanced confidence in the final 3D fluid-flow model. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER XX ESTIMATING PERMEABILITY OF RESERVOIR ROCKS FROM COMPLEX RESISTIVITY DATA P.S. Denicol and X.D.Jing Imperial College of Science, Technology and Medicine, London ABSTRACT Complex resistivity measurements were carried out on sandstone samples with a wide variation in permeability in the frequency range from 20 Hz to 2 MHz. Experiments were conducted in a multi-sample rig at effective confining pressures ranging from 400 psi to 4000 psi for water-wet samples saturated with 50,000 ppm NaCl brine. The measurements were made using a two-electrode technique. As a result, electrode polarization was observed in the low frequency range (20Hz - 10 kHz) and separated from the bulk sample response when data were plotted on an Argand diagram. The bulk sample response shows a nearly constant phase angle behaviour in the range from 10kHz to 100 kHz and the classical Cole-Cole behaviour with a depressed semi-circle from 100 kHz to 2 MHz. The frequency response of the complex impedance presents two regions of distinctive frequency dependence: the first one is located in the intermediate frequency range (10 kHz - 100 kHz) and has a relatively flat slope that correlates with permeability. The second region is located in the high frequency range (100 kHz to 2 MHz) and shows a more pronounced slope. The slope of the intermediate frequency range is practically independent of confining pressure for the high permeability samples and clearly pressure dependent for the low permeability samples. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER YY HIGH-RESOLUTION RESERVOIR CHARACTERIZATION OF MIDCONTINENT SANDSTONES USING WIRELINE RESISTIVITY IMAGING, BOONSVILLE (BEND CONGLOMERATE) GAS FIELD, FORT WORTH BASIN, TEXAS David L. Carr and Ronald A. Johns Bureau of Economic Geology. University of Texas at Austin, Austin, Texas Robert Y. Elphick Scientific Software-Intercomp, Denver, Colorado Laura S. Foulk Schlumberger Wireline & Testing, Englewood, Colorado ABSTRACT In the absence of abundant core data, Formation MicroScanner (FMS) and Fullbore Formation Microlmager (FMI) wireline logs from three wells in the Boonsville (Bend Conglomerate) Gas Field, Fort Worth Basin, Texas, provided valuable geologic information in a thick Pennsylvanian interval (900 to 1300 ft). It was critical to identify and trace key chronostratigraphic surfaces in these Mid-continent rocks to accurately capture the reservoir architecture in well-log correlations and 3-D seismic calibrations such that infill drilling strategies could be developed. Although cores provided premium stratigraphic information, only 358 feet of core from 4 wells were available in 30 sq. mi project area. Our interpretations were supplemented by the 1316 feet (total footage from 3 wells) of FMS and FMI wireline log data. FMS and FMI tools generated oriented, high-resolution resistivity images which served as continuous, ‘virtual’ cores that effectively tripled our high-res. stratigraphic database. After calibration with core, high-resolution geologic interpretations of the FMS and FMI data were made. We were able to routinely identify lithofacies, sedimentary structures, faults and fractures and key chronostratigraphic surfaces in the wireline resistivity images. Since image features are oriented in space, quantitative paleocurrent estimates were inferred from crossbed dips, which support the hypothesis that most Boonsville Bend Conglomerate reservoir sandstones in the project area represent lowstand valley-fill deposits derived from the Muenster and Red River Uplifts rather than from Ouachita-derived delta progradation. Combined analysis of cores and core-calibrated wireline resistivity images enabled the development of a fine-scale sequence stratigraphic framework which formed the basis for correlation and mapping of the major reservoir zones in the Bend Conglomerate interval, and also helped us to determine compartmentalization mechanisms. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER ZZ INTEGRATING IMAGING LOGS IN FORMATION EVALUATION M.Y. Fam, M. Haugland, K. Chemali, D. Seller, Halliburton Energy Services W.F. Stewart Chevron U.S.A. Inc. ABSTRACT Micro-imaging devices that use pad-mounted electrode arrays provide high-quality formation images. Formation evaluation is improved by using electrical micro-images as an aid in integrating logs and core data. In fact, direct comparison of electrical micro-images and core photographs illustrate the value of this technique and provide additional understanding of the nature and type of formation being investigated. The central electrode on each pad of a recently introduced six-arm imaging device is configured to measure the absolute survey current, from which a quantitative formation resistivity measurement can be derived. Such high-resolution quantitative resistivity data can be used for improving log analysis results in thinly laminated reservoirs. Three-dimensional modeling of the quantitative resistivity measurement shows that its depth of investigation is comparable to that of short-guard and other equivalent resistivity devices. Integrating electrical micro-image data with standard openhole logs through a special algorithm provides considerable improvements in the final log analysis results. Application of this technique to laminated shaly sand examples from the Gulf of Mexico has noticeably improved net-pay estimates, which have been in good agreement with production testing results. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER AAA ACOUSTIC POROSITY CORRECTION FOR GAS AND LIGHT HYDROCARBONBEARING SANDSTONES J. E. Thomas Smith, T. K. Fishburn, and E.L. Bigelow Western Atlas Logging Services, Houston, Texas ABSTRACT Gas-saturated sandstones slow the propagation of acoustic compressional waves through the formation. The slowing of the signal can be identified and corrected back to the proper velocity of a sandstone that has equal porosity, but occupied only by fluid. The ability to detect gas using acoustic velocities has been documented and demonstrated. The gas correction is performed using measured shear slowness (DTS), which is not affected by the contents of the pores, the measured ratio of cornpressional velocity (vp) to shear velocity (vs); (i.e., vp/vs), and the volume of shale throughout the interval of interest. The ratio vp/vs is crossplotted vs. DTS (1/vs). The crossplots will display a linear trend through fluid-saturated sand-shale sequences, while gas-affected intervals plot outside of the linear trend (Castagna et al., 1985). The correction method depicted in this article migrates the outlying data points back to the linear fluid-saturated trend by varying the vp component of the vp/vs ratio. The technique discussed here predicts the velocity that would have been obtained in a wet interval. Cutoffs are placed in the procedure, and they will vary with respect to the geological age and vertical depth of the sediments under study. The slope of the data trend is also noted as a variable, based on shale volume (Vsh), and is considered in the computation of gas-corrected velocity, Benefits are numerous, including: • • • • • • • More valid porosity values in gas-bearing zones More accurate lithology determination from crossplots with other logs Integration with seismic improved More accurate depictions of the geothermal pressure gradient (using acoustic log data) More accurate analysis of rock mechanical properties More accurate acoustic data to resolve special interpretative problems such as gypsum-bearing intervals or micaceous sands Cased hole porosity. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER BBB DETERMINATION OF PERMEABILITY DISTRIBUTION AT LOG SCALE IN VUGGY CARBONATES H. N. Greder, P. Y. Diver, S. Danquigny, F. M. Pellerin Elf Aquitaine Production, Pau, France ABSTRACT Vuggy carbonate reservoirs are characterized by poor permeability/porosity correlations and the assignment of permeability to each gridblock in reservoir modeling requires an accurate knowledge of the permeability distribution for each major facies at the appropriate scale. The thickness of individual gridblocks in reservoir modelling usually ranges from 0.3 meter (one foot) to several meters, meaning that permeability distribution has to be provided by core analysis at least at a 0.3m vertical scale (or log scale). In the industry, the conventional approach consists in sampling small cylindrical core plugs (2.5 cm * 2.5 cm) every 0.3m along the cores on the assumption that permeability distribution measured on those small samples is representative of the permeability distribution at log scale. Although this may be valid in many facies, this paper proves that this method leads to an overestimation of permeability dispersion at log scale in vuggy carbonates. This study focused on two types of vuggy carbonate facies. Both types of facies can be considered homogeneous on a geological basis for reservoir modeling purposes. Small plugs were sampled at one-foot spaced regular intervals along cores representative of these facies in order to perform gas permeability measurements. Four 20 cm-long full size samples representative of the two facies were saved to perform extensive gas permeability measurements at different scale: full-size scale which is close to log scale, large plug scale (40mm * 50mm). small plug scale (25mm*25mm) and minipermeameter scale (0.2 cm3). More than 200 gas permeability measurements were performed on each full-size sample leading to their total destruction. These measurements show that permeability distribution within these full size samples strongly depends on the scale of measurement. The value in the center of the range is consistent from one type of measurement to the other but the spread about the center significantly increases as the scale decreases. The results show that permeability dispersion measured at small plug scale within one foot long intervals is significant when compared to the dispersion measured on one foot spaced small plugs. The true permeability dispersion at log scale is therefore inferior to that estimated by the conventional approach. - The results prove that in order to estimate the distribution of permeability at log scale in vuggy carbonates, large plug sampling every foot should replace small plug sampling every foot. Extensive probe permeameter measurements along the cored intervals is a good alternative to properly estimate permeability distribution at log scale. A recommended number of probe measurements per foot is provided in the paper. This figure was obtained by Bootstrap techniques and compared to the one calculated with the “No testing method” which was recently published. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER CCC AUTOMATIC EXTRACTION AND CHARACTERISATION OF GEOLOGICAL FEATURES AND TEXTURES FRONT BOREHOLE IMAGES AND CORE PHOTOGRAPHS Jonathan Hall, Marco Ponzi, Mauro Gonfalini and Giorgio Maletti Agip S.p.A., Milan, Italy ABSTRACT A methodology is proposed for complementing the skills of reservoir characterisation professionals engaged in the interpretation of wireline images and core photographs. This, largely automatic, methodology recognises strengths and weaknesses of human cognition in the context of analysing images to extract sedimentological, structural, petrophysical and reservoir description information. Certain challenges must be overcome if automation of feature recognition (planes, lenses etc.) and texture segmentation from images is to be realized. These arise from the consequences of divers and extreme borehole environments, registration problems and limitations in logging tool design and instrumentation. Some shortfalls are commented upon and their relative impacts assessed. Most borehole image and core photograph interpretation packages available today consist mainly of utilities for image display of small intervals of data, interactive picking of planar events, computation of plane orientations and the presentation of summary statistics. These capabilities appear relatively immature when compared with advances made in S.E,M., satellite remote sensing and medical imaging. Using techniques established in image processing, such as edge detection and mathematical morphology and partitioning afforded by co-occurrence matrices and artificial neural networks, the extent to which human acquired experience and inherited skills can be ameliorated by automatic procedures is explored. The particular applicability of Hough transform for characterising certain geometrical shapes expressing geological features is examined. A comparison is made of an automatic image analysis method and the dipmeter correlation method of planar events picking. An alternative and largely innovative approach is offered, employing a consistent and automatic reduction of voluminous borehole image and core photograph data to pertinent image texture attributes. These may be compared to analytical models of particular geological structures, prior to human interpretation. Optimisation techniques such as simulated annealing are used to implement optimal parameter search routines in three dimensional problems, offering improvement in computational performance. Limitations to the automatic procedures tested are identified. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 Recognition of expressions of geological processes at all scales requires human mental abstraction to some known model. The proposed procedures should ensure that this will be more successfully, more consistently and more cost-effectively realised. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER DDD RESERVOIR CHARACTERIZATION BY COKRIGING PETROPHYSICAL WELL DATA AND SEISMIC ATTRIBUTES Tomas Hansen Western Atlas Logging Services, London, U.K. ABSTRACT Describing the heterogeneity of a reservoir is a challenge to all geoscience disciplines. A full understanding of the depositional environment and the well performances can be used to generate a reservoir model that provides the most likely hydraulic response of the reservoir. Such a model may still be a poor predictor of inter-well reservoir properties, giving high uncertainty in development drilling and reservoir simulations. This uncertainty may be minimized if the seismic response is used to guide the interpolation of petrophysical properties and lithology changes between the wells. A brief introduction to geostatistics, including regression and correlograms, is provided. Two practical applications using this approach are then presented: • Petrophysical properties from 17 wells and two seismic attribute maps are combined to generate cokriged porosity and productivity index maps of a producing field, explaining interference testing results. • Electric logs and vertical seismic profile from a single well are used to produce sand volume cross section. indicating location for an appraisal well. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER EEE VISUALIZING PRODUCTION IN FLOWING OIL WELLS Steve Maddox Halliburton Energy Services ABSTRACT Recent video recordings of oil and gas production into wellbores through perforations and of the resulting flow up production tubulars are enhancing the industry’s ability to measure and model such production flow. This paper and accompanying video presents visual examples of oil and gas production at various depths, pressures, temperatures, water cut percentages, inclination angles, and tubular sizes. A brief overview of the current state of the art of downhole video technology and of the history of such technology is followed in this paper by some cases of video production logging. These cases include analysis of flow patterns observed and the factors that contribute to those patterns. Some of the challenges involved in designing flow meters to accurately measure oil and gas production are presented and made visually apparent from the actual footage of oil and gas production through the inclined flow paths normally found in the oil field. The insight gained from viewing production flow as it happens in the oil field will ultimately result in more accurate production logging tools and techniques. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER FFF THE 3-DIMENSIONAL MODELLING OF LARGE-SCALE ISOLATED INHOMOGENIEITIES IN A NORTH SEA RESERVOIR SAND BODY John Foot British Petroleum, Aberdeen, Scotland Geoff Page Western Atlas Logging Services, Aberdeen, Scotland Qiang Zhou Western Atlas Logging Services, Houston, Texas ABSTRACT The response of resistivity measurements to a traditional ‘layer cake” formation with variations in target bedding thickness, wellbore deviation, and invasion profiles has been well documented, and can now be quickly modelled with a variety of 1-D and 2-D computer codes. This paper describes the modelling of truly 3-dimensional inhomogeneities, in the form of boulder-sized concretions within otherwise homogenous bedding, and the anomalous effects that these inhomogeneities can have on measurements made from a wellbore that does or does not intersect with them. As discussed in previous papers (Hardman and Shen, 1986), other possible causes (not discussed here) of induction anomalies include washouts or caves on the borehole wall filled with conductive mud, and a highly deviated borehole with the tool crossing a bed boundary. These should also be considered as options in any analysis. The project in this study was defined by a plan to drill a nearly horizontal wellbore through a thick reservoir sand body, in which layers of high-resistivity concretions sometimes occur. The locations of these layers are difficult to predict, and the well could be drilled along or close to one of these layers. Although the total volume of the concretions is a small percentage of the reservoir’s actual volume, the effect of these concretions on the recorded measurements, the calculated saturations, and net/gross was unknown. The target formations are described with reference to surface outcrops of similar character several hundred miles to the southwest. These similarities helped define the model parameters. The 3-D modelling process is briefly discussed, and the results of several modelled “representative cases” are presented. Some of the resulting “anomalies” may be useful in explaining strange log responses from historical wells, because the new modelling method is applicable to wellbores of any deviation. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 At the time of writing this paper, the target well has not yet been drilled, but may be completed by the time of publication. So, there could be an update to report whether a layer of this type was actually encountered. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER GGG NMR LOGGING IN A NORTH SEA GAS WELL O.M. Relling, T. Eidesmo, L.H. Flølo and H. Rueslåtten Statoil Research Centre, Trondheim ABSTRACT The new generation of Nuclear Magnetic Resonance (NMR) logging tools provides a way of obtaining the volumes and relaxation behaviour of Hydrogen-bearing fluids confined in the pore spaces of reservoir rocks. Based on the relaxation behaviour of the fluids, it is possible to estimate parameters such as effective porosity, fraction of movable and irreducible fluids and permeability. This ability to obtain several petrophysical parameters simultaneously makes NMR logging an attractive addition or alternative to the traditional suite of well logging tools. This paper reviews some experiences gained from a comparative study of a North Sea gas well using a traditional suite of well logging tools, NUMAR’s MRIL-C tool and laboratory experiments on core plugs including standard petrophysical analysis and NMR measurements. The main purpose of this study was to investigate if the Bound Volume Index (BVI) represents the irreducible water saturation (Swirr) calculated from resistivity logs. To match the irreducible water saturation from cores (based on Dean & Stark titrations), the four first bins in the NMR T2 spectra were used. By defining BVI as the four first bins of the MRIL-C T2 spectra it was not possible to obtain a good match with Swirr from resistivity logs in all zones. However, the MRIL-C BVI data clearly confirm the qualitative changes in Swirr. The main uncertainties related to the MRIL-C BVI parameter is due to a shift in the T2 spectra towards longer T2 components; most likely caused by dissolved gas (carbon dioxide) in the formation water and high temperature. This shift makes it difficult to obtain reliable estimates of Swirr and permeability, and clearly demonstrates the need for developing a further understanding and definition of the BVI for quantitative use in gas wells. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER HHH IMPROVING MWD SURVEY ACCURACY IN DEVIATED WELLS BY USE OF A NEW TRIAXIAL MAGNETIC AZIMUTH CORRECTION METHOD Roar Sognnes Halliburton Energy Services Bob Smith BP Norge U.A. Graham McElhinney Halliburton Energy Services Dr. Michael A. Yuratich TSL Technology Ltd. ABSTRACT Inaccuracies in measurement while drilling (MWD) survey data produce uncertainties in the position and true vertical depth of every point in the wellbore. Magnetic azimuths from MWD directional sensors are affected by magnetic field distortion from many sources, including the drillstring, even when nonmagnetic collars are used. These errors accumulate over the length of the well and may affect calculations of bottomhole location by more than 10 m displacement over a 1000 m well section. Current magnetic azimuth correction methods attempt to remove axial magnetic field distortion (known as z-axis) by initially trying to solve the error in that axis, Some then use ad hoc methods to remove the distortion along the other two orthogonal axes. A new triaxial magnetic correction method has been developed that uses standard. raw MWD survey data as inputs. The results obtained in a recent extended reach well in the North Sea show considerably less azimuth error than only zaxis-corrected MWD surveys and north-seeking gyro surveys, when compared results from a wireline inertial guidance system. The new method thus reduces the need for expensive and time-consuming gyro surveys in directional wells. This paper describes the new correction method, then demonstrates the method on data obtained from sections of this North Sea well, with comparisons to uncorrected MWD azimuth readings, a conventional magnetic azimuth correction model, and two different highaccuracy gyro surveys. The new method was found to produce significant reductions in the ellipse of uncertainty compared to the uncorrected MV/I) data. In addition, the azimuths from the new method were far more consistent in direction than the z-axis-corrected azimuths, and compared more favorably with the gyro azimuths. Additional verification of the method is provided by calculation of the earth’s magnetic field. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER III THEORETICAL AND EXPERIMENTAL STUDIES ON THE DETERMINATION OF THE AVERAGE FLUID VELOCITY FROM SPINNER FLOWMETER RESPONSES S. Song, C. Jordan, and D. Georgi Western Atlas Logging Services, Houston, Texas ABSTRACT To accurately determine the average fluid velocity in a wellbore from spinner flowmeter responses it is necessary to take the radial fluid velocity profile into account. This paper presents theoretical and experimental data for the velocity profile correction factor for spinner flowmeters. The theoretical studies show that the correction factor is a function of both the Reynolds number and the ratio of the tool diameter to the wellbore diameter. The correction factor can vary between 0.8 and 0.95, depending upon the flowrate, wellbore diameter, and spinner size. The 0.83 correction factor widely used by the industry is only applicable to spinner flowmeters with very small blade sizes compared to the wellbore diameter (about 5% of the wellbore size). Extensive water and oil-flow experiments conducted in a flowloop consisting of two pipes having 6.5-in. (16.5-cm) II) and 8.5-in. (21.6-cm) ID verified the theory at flowrates higher than 159m3/d or 1,000 b/d for water flow and 427m31d or 3,000 b/d for oil flow. At lower flowrates, the experimental data deviated from the theoretical predictions. This is attributable to the fluid friction force acting on the spinner blades, which is significant at low velocities compared to the fluid momentum. It is concluded from this study that the minimum fluid velocity to which the spinner will respond linearly is a function of the flowing fluid properties. Also, the minimum velocity is higher for oil flow than for water flow. This study provides the theoretical equations that govern the determination of the spinner flowmeter correction factor. These equations can be easily implemented and the computed flowrate may increase by as much as 15%. Various charts that can be used to determine the correction factor, based on the spinner size and wellbore diameter, are also provided in this paper. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER JJJ PERMEABILITY AND PERMEABILITY ANISOTROPY CHARACTERISATION IN THE NEAR WELLBORE: A NUMERICAL MODEL USING THE PROBE PERMEAMETER AND MICRO-RESISTIVITY IMAGE DATA Simon Thomas, Patrick Corbett & Jerry Jensen Department of Petroleum Engineering, Heriot Watt University, Edinburgh, Scotland, UK ABSTRACT The Morecambe gas reservoirs are located within Block 110 in the East Irish Sea Basin. The reservoir interval lies within the Ormskirk Sandstone Formation of the Sherwood Sandstone Group. This is lithostratigraphically assumed to be Triassic (Scythian) in age. The producing horizon comprises of mixed sabkha and fluvial deposits with subordinate aeolian facies. Within this area gas recovery is highly sensitive to the kv/kh, the ratio between vertical and horizontal permeabilities, especially in those zones affected by illite cementation. This determines production inflow into horizontal wells and the rate of recharge from underlying gas sources into the main gas reservoirs. At present, kv/kh is either inferred from coarse numerical methods or measured in the well-bore by the Multi-Probe Formation Dynamics Tester (MDT). This paper proposes an improved numerical technique for estimation of kv/kh in the near well-bore region. The model is built around core derived probe permeability data measured at the lamina scale. Assuming these lamina to have effectively isotropic permeability, a numerical model is then utilised to estimate kv/kh up to the bed-scale. The Formation MicroImager (FMI) has mainly been used as a sedimentological tool. However, resistivity data, which may be derived from the tool, are at a similar measurement scale to that of the probe permeameter. Correlation of these data has allowed a prediction of permeability into uncored intervals. The relationship between the FM! data anisotropy and permeability anisotropy has allowed kv/kh to be predicted in the near well-bore. Results are presented demonstrating agreement between estimates using this technique and estimates derived from the MDT. SPWLA 37th Annual Logging Symposium, June 16-19, 1996 PAPER KKK APPLICATION OF AN ACOUSTIC IMAGE DEVICE TO OBTAIN FULL STRUCTURAL INFORMATION IN HALITE E. Voigt Western Atlas Logging Services, Den Helder, The Netherlands J. Haberland Pipeline Engineering GmbH, Germany ABSTRACT This paper describes the ability of the acoustic Circumferential Borehole Imaging Log (CBIL) to provide structural information in a highly plastic deformed evaporite sequence and its potential to reduce the need for coring. In the past. RUHRGAS AG’s storage cavern wells have been cored extensively, and the oriented cores have been used to assess the deformation of evaporites present and to recover samples for laboratory analysis. The CBIL service has identified both bedding features (formed by interbedded anhydrite and polyhalite layers) and diagenetic features such as recrystallization boundaries. With cored material, the latter features cannot be identified without the aid of transmitted light. Recrystallisation boundaries could, however, be directly identified from the CBIL-image data, which also gives full radial and vertical coverage of the borehole. The subsequent structural analysis can therefore separate the diagenetic features on the basis of their different acoustic character. Research has shown that the acoustic CBIL service is a reliable means of acquiring accurate image data in structurally complex evaporite sequences and that it can provide a low-cost alternative (up to 70% cost reduction in the cored interval) to expensive conventional coring programs. TRANSACTIONS OFTHE SPWLA THIRTY-EIGHTH ANNUAL LOGGING SYMPOSIUM Sponsoredby THE SOCIETY OF PROFESSIONAL WELL LOG ANALYSTS, INC. 8866 Gulf Freeway, Suite 320 Houston, Texas 770 17 Presentedat THE ADAM’S MARK HOTEL Houston, Texas June 15-18, 1997 NOTICE TO EDITORS: Permission is hereby granted to publish elsewhere any of these transactions after June 18, 1997, provided that conspicuousacknowledgementis given to the original presentation of the paper and the authors of the paper have agreedto the republication. (The statements and opinions expressed in these transactions are those of the authors and should not be construed as an official action or opinion of the Society of ProfessionalWell Log Analysts, Inc.) SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER A COMPLEXITIES OF THE LATERAL RESPONSE AND TECHNIQUES THAT FACILITATE ES-LOG INTERPRETATION Pedro Anguiano-Rojas and James H. Spurlin Instituto Mexicano del Petroleo ABSTRACT Modern log analysts confronted with electrical survey (ES) logs often tend to use the normal curves because they understand them much better than the lateral tool. Even depth-matching of the lateral response proves to be a challenging task. The lateral tool is valuable because it reads deep into the formation and yields information about bed boundaries. The lateral tool is the only widely used asymmetric sensor array. Modeling and inversion software can handle laterals and help log analysts, but this kind of software may not be available to most log analysts and is computer-intensive. Furthermore, modeling software might not help the log analyst fully understand the response of the tool. We show that the lateral response results from the addition of two components, namely a normal response, with AO spacing (typically 18 ft 8 in.; 5.7 m), plus a derivative. The presence of these two components together with a shift in depth between them is what makes the appearance of the lateral response confusing. The normal component behaves in the known standard way for any normal tool, including its reversal for layers thinner than AO. The depth shift between the normal and the derivative is half AO. The derivative of the normal apparent resistivity is with respect to the tool spacing and not to the depth coordinate; this makes it behave in unexpected ways. The derivative is asymmetric in two ways: given a step response the derivative is mainly above the interface and the amplitude is very different depending on the direction of the step. As beds get thinner than AO, the amplitude of the derivative remains almost constant. Nevertheless, the peaks of the derivative indicate the location of the bed boundaries. The study of the different behaviors of these two components, together with the depth shifting between them, is key to understanding the complexities of the lateral tool response. We can split the apparent resistivity curve of the lateral into normal and derivative components. This facilitates the interpretation and correlation of the asymmetric lateral curve with other logs. The derivative component also can aid in picking bed boundaries for an initial guess in the inversion process and both components help assess the quality control of the whole interpretation process. We apply the splitting of the lateral response to both synthetic and field data. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER B STRANGE INVASION PROFILES: WHAT MUILTIARRAY INDUCTION LOGS CAN TELL US ABOUT HOW OIL-BASED MUD AFFECTS THE INVASION PROCESS AND WELLBORE STABILITY Jack La Vigne, Tom Barber and Tom Bratton Schlumberger ABSTRACT Invasion is considered a nuisance in interpreting resistivity logs. No resistivity tool can “see past” invasion without being affected by it. The problem has been attacked over the years by creating tools with multiple depths of investigation such as the dual induction and dual laterolog tools. The dual induction tool combined deep and medium depth induction arrays with a shallow laterolog measurement. The interpretation of a dual induction or dual laterolog tool requires a simplistic model of invasion which often does not fit what is actually happening in the earth. The AIT Array Induction Imager Tool produces five logs carefully designed to maximize radial information content. This has allowed inferences concerning invasion profiles that could not be obtained with conventional technology. Oil-based mud (OBM) invasion has been particularly poorly understood because the shallow measurements in conventional tools are laterolog devices that cannot operate in the resistive mud. This limits the usable logs to just the deep and medium induction measurement, from which it is difficult to infer quantitative information about the invasion process in OBM. By looking carefully at AIT field log examples with OBM invasion, several observations can be made. First, contrary to conventional wisdom, OBM can invade deeply. Second, OBM is a complex mixture of oil, salt water, and the surfactants necessary to keep the oil-water mixture in emulsion. Invasion can involve the oil phase, the water phase, or both. The surfactants can have profound effects on the formation connate water in the invaded zone and greatly alter residual water content by reducing the surface tension of the water. Logs with both R~,> R1 and <R1 profiles have been observed with OBM, indicating oil-phase and water-phase invasion, respectively. Other effects besides invasion can produce apparent “invasion profiles” on AIT logs. These effects include shale fracturing and alteration. Geomechanical modeling of the stresses involved (from overburden pressure, well-bore pressure, and pore pressure) indicates that weak rocks such as shales fracture in complicated patterns. By modeling the AIT logs with the assumed fracture patterns using 3D forward modeling codes, one can show that the geomechanical model is consistent with the fracture pattern needed to produce the observed log effects. All of these “strange invasion profiles” contribute to our understanding of how choices made during drilling and mud selection can affect the invasion process. This illustrates how we can infer much more than resistivity from resistivity logs. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER C MEASUREMENT OF DIP ANGLE AND HORIZONTAL AND VERTICAL RESISTIVITIES USING MULTIPLE FREQUENCY PROPAGATION RESISTIVITY TOOTS Jian-Qun Wu, M. M. Wisler, and W. H. Meyer Baker Hughes INTEQ Houston, Texas ABSTRACT It has been shown that using resistivity tools with axial-dipole antennas one can measure horizontal resistivity and a function of vertical resistivity and relative dip angle. Without additional information, the vertical resistivity and relative dip angle can not be uniquely determined no matter how many antenna spacings and/or frequencies are used in the measurement. The determination of horizontal resistivity and the function of vertical resistivity and relative dip angle using resistivity tool measurements is a nonlinear inversion process. Stability and uniqueness of the solution are two problems associated with the nonlinear inversion process. To accurately and uniquely determine the horizontal resistivity and the function more than two measurements are required At high frequencies where the dielectric effect becomes important, the horizontal resistivity can not be determined with only two measurement in deviated wells. When dielectric effect is significant, the antenna readings are affected by five formation parameters: two resistivities, two dielectric constants and the dip angle. Four independent parameters derived from the five formation parameters completely determine the antenna readings. Therefore, all five formation parameters can not be determined without using a piece of information from sources other than the tool readings. Factors causing the resistivity anisotropies necessarily cause dielectric anisotropies. The resistivities and dielectric constants of a given formation are not independent of each other. Once the relationship between them is established for a given formation type, one can use the four parameters measured plus the formation anisotropy model to uniquely determine all five formation parameters. If the relative dip angle is known, the resistivities and dielectric constants can be determined without any formation models. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER D SHEAR WAVE VELOCITIES IN VERY SLOW GAS-BEARING SANDS IN THE OFFSHORE NILE DELTA, EGYPT Mohammed Badri and Alain Brie Schlumberger Wireline & Testing Saad Hassan Belayiem Petroleum Company ABSTRACT In the offshore Nile Delta of Egypt, several significant gas discoveries have been made based on amplitude anomalies observed on three-dimensional surface seismic data. Accurate measurements of compressional and shear sonic velocities are essential to verify and calibrate seismic amplitudes associated with gas-bearing sand reservoirs. The reservoir is at a depth of about 2000 m and consists of poorly unconsolidated Pliocene age sand. Recent acquisition of shear sonic velocities using a dipole shear source over the gasbearing sand showed not only very low shear wave velocities but also the presence of two distinct shear wave arrivals with a significant slowness difference between them. Although the P-wave slowness ranged between 150 to 170 µsec/ft the shear wave slowness was observed between 280 to 475 µsec/ft. Over the reservoir section. the presence of two shear wave arrivals produced two significantly different shear wave velocity values that resulted in two different Poisson’s ratio measurements. Amplitude Variation with Offset (AVO) analysis is strongly affected by the value of Poisson’s ratio, making the appropriate choice important. Spectral frequency analyses of the shear waves recorded over the reservoir section showed that the two shear arrivals are at two different frequencies. Two distinct spectral peaks at 1 and 2 kHz were observed. The lower frequency of 1 kHz corresponds to the slowest arrival at about 450 µsec/ft whereas the higher frequency peak corresponds to the faster arrival at about 280 µsec/ft. This suggests that the two arrivals travel in different parts of the formation. This paper investigates and discusses two hypotheses that have been proposed to explain the presence of dual shear arrivals in the gas-bearing sand reservoir. The first hypothesis is the effect of borehole fluid invasion; however, the effect is significantly larger than that predicted by the Biot-Gassmann theory. The other explanation is the presence of a mechanically altered zone, associated with the release of rock stresses in the sand around the wellbore. Shear wave anisotropy measurements showed little anisotropy effect around the wellbore over the gas-bearing sand section. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER E INTEGRATION OF CROSS-DIPOLE ACOUSTICS FOR IMPROVED FORMATION EVALUATION Doug Patterson Western Atlas Logging Services, Houston, Texas Gabriela Schell Phillips Petroleum Company, Houston, Texas ABSTRACT Uses for dipole acoustic measurements have steadily increased in recent years. These uses have involved combining shear slowness with other wireline data for various well-documented applications. The newest development for dipole acoustics is the use of an orthogonal source and receiver system (cross dipole) to measure azimuthal anisotropy. The cross-dipole development has its background in the shear seismic domain. A review of this background and the basic theory provides valuable insight into the proper application of this developing technology. This technology, along with the dipole’s ability to investigate three to five borehole diameters, allows for evaluating azimuthal features beyond the borehole wall. Application of the cross dipole with actual field data is presented. The critical issue of proper quality control is covered since the rotational analysis puts additional demands on dipole systems. The repeatability of the azimuthal anisotropy results is shown by utilizing data obtained from multiple passes that exhibit varying tool orientations. The results are compared to conventional analysis using a standard suite of openhole logs along with Stoneley analysis from full waveform monopole data. The comparison demonstrates that the integration of cross-dipole analysis does allow for improved formation evaluation. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER F SHEAR SLOWNESS DETERMINATION FROM DIPOLE MEASUREMENTS Alain Brie Schlumberger K.K., Wireline & Testing, Fuchinobe, Japan Christopher V. Kimball Schlumberger-Doll Research, Ridgefield, Connecticut, USA J. Pabon, Yoshiyuki Saiki Schlumberger K.K., Wireline & Testing, Fuchinobe, Japan ABSTRACT Acoustic modes that are linked to the presence of the borehole are dispersive. That is, their phase slowness (the inverse of velocity) varies with frequency. Equivalently, the pulse wave shape changes as it propagates across the array. The Stoneley wave generated by a monopole source and the flexural wave generated by a dipole source are examples of dispersive borehole modes. The flexural wave is important because it provides shear in slow formations (i.e., formations in which the shear slowness, the inverse of shear velocity, is slower than the compressional slowness of the borehole fluid). Flexural mode dispersion makes shear slowness determination more difficult. The dipole flexural wave propagates at the shear slowness at zero frequency, but above this frequency, its phase slowness increases by an amount called the dispersion bias. The dispersion bias usually represents a small fraction of the measurement One approach would be to measure at very low frequency where the dispersion bias is negligible; this is, however, very difficult to achieve because of the extremely low excitation of the flexural mode at low frequency. In practice, dispersion of the flexural wave cannot be avoided and must be taken into account. Various methods can be used to measure shear from the flexural wave in frequency bands where the flexural energy is sufficient for an accurate slowness measurement, but where dispersion bias is not negligible: • • Process the waveforms with traditional non-dispersive techniques in a narrowfrequency band to minimize frequency dispersion and then apply a correction for dispersion bias. The slowness time coherence (STC) processing in a narrow frequency band followed by dispersion bias correction is a robust technique, but it lacks flexibility. In some instances, the correction may be inaccurate when the model used to generate the correction tables does not fit the actual conditions. Process the waveforms dispersively, making full use of the flexural wave dispersion and eliminating the requirement for dispersion bias correction. Dispersive processing is the maximum likelihood or least-mean-squared error SPWLA 38th Annual Logging Symposium, June 15-18, 1997 solution and provides better results for uncertain wave spectra. It also allows a more accurate representation of the borehole condition (for example, the presence of very slow oil-base mud in the borehole). In this paper, dispersion curves obtained from mode tracking are presented. The different processing techniques are discussed and compared. Examples are given in formations of different types and discussed. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER G COMPUTING BOREHOLE GEOMETRY AND RELATED PARAMETERS FROM ACOUSTIC CALIPER DATA John F. Priest Western Atlas Logging Services, Western Atlas International, Inc., Houston, Texas ABSTRACT A robust method for calculating borehole eccentricity, major and minor diameters, elliptical orientation, eccentering radius, and direction (position of tool relative to center of the borehole) is presented. Applications of the geometrical outputs are presented for both openhole and cased hole wells. Even in casing, the ellipse direction and the eccentering direction can be oriented to the high side or the low side, and if valid orientation data are available, to magnetic or geographic north. Data defects in the acoustic caliper data, such as early picks and missing data, are corrected using elliptical (or circular) models for the wellbore or casing. The eccentering radius and orientation are used to correct for image distortion caused by eccentering and to compute the borehole bend radius. Acoustic amplitude processing using the computed geometrical parameters includes corrections for mud attenuation, reflection angle, and transducer radial properties. Acoustic radius processing (or traveltime) corrections include centering the radius image so that the image appears as if the tool were perfectly centered. Eccentered radius data appear on an image plot as gradually changing light/dark bands running parallel to the borehole, while ellipticity will appear as two light/dark bands. The data can be presented without the image defects caused by eccentering and ellipticity by producing a ‘flat’ image of the borehole or casing. The centered and flattened image provides a ‘residual’ image presented as a standard radius image by adding the geometrical mean or arithmetic mean of the data to the residual. This image significantly enhances the visualization of small-scale features in the radius image that could have been obscured by either ellipticity or eccentering or both. Three applications are presented. A method for open-hole wells will illustrate the possibility of relating rock properties to radius data, stress orientation, and borehole erosion. Casing analysis illustrates the flattening of the radius, which can be routinely plotted with a 0.050 to 0.100-in, full-scale range, and show various forms of casing distortion. The geometrical outputs can be processed to provide borehole curvature, torsion, casing strain, and orientations. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER H INVESTIGATIONS INTO ANOMALOUS RESPONSES OF GAMMA DENSITY AND NEUTRON POROSITY TOOLS IN HORIZONTAL GAS WELLS P.Cowan and G.A.Wright AEA Technology plc, UK ABSTRACT Observations in horizontal wells in Southern North Sea gas fields (Cuddy 1994) show that gamma density and neutron porosity tools give anomalous responses compared to vertical wells in the same formation. One hypothesis given for these observations was nonuniform invasion around the horizontal borehole due to different horizontal and vertical permeabilities and fractures at the sides of the borehole caused by stress overburden. Several companies were interested in understanding and quantifying these observations, including Amoco, BP Exploration, Saga and Statoil who went on to sponsor a programme of work to investigate the hypothesis of non-uniform invasion. This was carried out by modelling the responses of laboratory gamma density and neutron porosity tools with the Monte Carlo code MCBEND. The Monte Carlo method is now a well established technique for predicting the response of nuclear logging tools. These calculations include explicit representation of the geometry, source and materials of the problem and of the physics of particle transport. They are thus numerical experiments, equivalent to a range of well controlled measurements in which all of the relevant parameters are known. This paper presents the results of these investigations. The hypothesis of a nonuniform invasion profile around the horizontal borehole does explain the observed density tool response. As a side benefit of this work, information on the depth of investigation of the tools and sensitivity of the tools to different formation regions were obtained. These results highlighted the differences between the gamma density and neutron porosity tool and the difficulty in determining porosity in gas wells with invasion. The results showed that logs need to be corrected or compensated for this gas effect Therefore a number of methods of determining formation porosity in horizontal gas wells were studied as part of this project. These methods included the square root method and an iterative method which is used in practice. This project helped to identify the methods which give the most accurate determination of porosity in horizontal gas wells. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER I THE EMPIRICAL INVESTIGATION OF DENSITY ANISOTROPY IN HORIZONTAL GAS WELLS Jonathan Bedford Schlumberger Wireline & Testing, Aberdeen, UK Steve Cuddy BP Exploration, Aberdeen, UK Jim White Schlumberger Wireline & Testing, Aberdeen, UK ABSTRACT Following the introduction of horizontal wells into Southern UK gas fields, density readings were observed to be significantly lower than those measured in nearby vertical wells. A model relying on asymmetric invasion was proposed to explain this anomaly, but further support for this approach was needed. This took the form of numerical modeling, which supported the model, and in-situ readings from around the well-bore. The overall objective was to be able to quantify and use these anomalous density readings to obtain true formation porosities. To test our model, two density tools were run sequentially in a horizontal well drilled in BP’s Newsham field. The two densities were oriented at 90 degrees to each other, one focused down and the other to the side of the borehole. Based on the modeled invasion profile, the side density was expected to measure density values similar to the offset vertical wells. The down facing densities measured relatively low density, as predicted by the model. However, the side facing density log gave far lower densities than predicted, and intermediate density tool positions confirmed that the measured density varied dramatically around the borehole. Ultrasonic borehole images showed microfracturing along the side of this well. It is thought that these micro-fractures are caused by loop stresses that form around horizontal wells. These fractures contribute to the very low orthogonal density values seen in the well. This innovative application of density tools demonstrated two insights into horizontal well logging. First, unusual invasion profiles can be formed in horizontal wells, by formation anisotropy, causing the density log to read an apparent too-high porosity if normal filtrate invasion is assumed. Secondly, stresses around horizontal wells cause micro-fracturing at the side of the borehole which results in real increases in density porosity. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER J INTERPRETATION OF THE RESPONSE OF A NEW THROUGH-TUBING CARBON/OXYGEN INSTRUMENT USING NUMERICAL MODELLING TECHNIQUES M.W. Mickael, W.A. Gilchrist, Jr., and R.J. Mirzwinski Western Atlas Logging Services, Houston, Texas G.N. Salaita Chevron Petroleum Technology Company La Habra, California R.T. Rajasingam Saudi Aramco, Dhahran, Saudi Arabia ABSTRACT Carbon/oxygen measurements are unique in their ability to measure formation saturation in fresh, brackish, or unknown water salinity. Progress in numerical modeling has enhanced both the design and interpretation of carbon/oxygen instruments which improves the overall accuracy of the saturation measurement Modeling the energy response of gamma-rays produced by inelastic and capture neutron collisions has traditionally suffered from the lack of accurate models or from long computational times required to achieve acceptable statistical precision. A new computer model based on the numerical solution of the neutron and photon transport equations makes it possible to obtain accurate non-statistical estimates of the response of carbon/oxygen instruments in a short computer time. The model was used to fully characterize the response of a new instrument to changes in formation saturation and the effects of porosity and lithology, borehole size and composition, and casing and cement Formation saturation is directly calculated by comparing the measured carbon/oxygen ratio to the estimated response based on modeling data of the formation and borehole conditions. Good results were obtained by applying the new interpretation method to measurements made with a new 1.69inch (43 mm) diameter pulsed neutron instrument under a wide range of formation and borehole conditions. Examples are presented to show the carbon/oxygen interpretation in carbonate and silicate mineralogy and in shut-in and flowing conditions. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER K A NEW APPROACH TO COMPUTING FORMATION DENSITY AND Pe FREE OF MUDCAKE EFFECTS F. Allioli, O. Faivre, L. Jammes Schlumberger Wireline & Testing M. Evans Anadrill ABSTRACT Density measurements are generally affected by the presence of mudcake against the formation. In order to correct for such environmental effects, a density device comprises several detectors having different depths of investigation. Algorithms in current use combine the different sensor measurements in order to derive corrected formation density as a function of the detector count rates. In this paper, we discuss a measurement analysis technique based on physical response models of the sensors. These models predict the tool response as a function of various physical and geometrical parameters affecting the measurement. Given a set of measurements with different depths of investigation, the inversion of the response equations using iterative techniques will allow the computation of formation parameters free of mudcake effects. The use of a forward model-based inversion brings the following advantages compared to previous analysis techniques; optimal estimation, error characterization and solution control. The first part of the paper details how the parametric forward model is established and calibrated, using reference experimental data and a priori knowledge of the physics of Compton scattering and of the photoelectric effect. In the second part, the inversion algorithm based on the minimization of a cost function is presented. The minimization technique, the computation of errors in formation parameters, and the constraints and control of the solution are discussed and illustrated by several examples. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER L REVISITING THE NEUTRON CAPTURE EQUATION: A NEW ROCK PERSPECTIVE VS. THE TRADITIONAL FLUID PERSPECTIVE Varouj Y. Kotchounian Schlumberger Wireline & Testing ABSTRACT Pulsed neutron capture technology is a well-established methodology in the oil and gas industry. Measurement of formation capture cross section (sigma) has been used for decades in the estimation of fluid saturation, primarily in cased hole environments. The stand-alone applications have varying success, depending on the complexity of the environment, with most cases mandating a time-lapse approach. One of the main causes for the failure of the stand-alone applications has been the presence of neutronabsorbing minerals other than chlorine in the formation. To capitalize on this fact, the sigma equation is revisited from a different perspective: a “rock” perspective rather than the traditional “fluid” perspective. An innovative technique is presented for quantification of the neutron-absorbing properties of the formation rock in a manner that is independent from the other formation properties such as porosity, fluid content, saturation and salinity. The technique is based on the unique combination of openhole capture cross section (sigma) and electromagnetic wave attenuation measurements. Full error analysis is performed to quantify the errors resulting from the assumptions made and any possible inaccuracies in the knowledge of the requited environmental and formation properties, including incomplete flushing, variations in the salinity of the invaded zone, effect of shales and clay minerals, temperature, heterogeneous mineralogy and carbonate impurities. The results of the analyses indicate that the errors at minimal. In most cases, the effect is less than 1 capture unit. The technique has several promising applications, including clay analysis and stratigraphic correlation. The approach was applied in several wells drilled through Cenomanian and Albian formations in the Egyptian Western Desert. Because both groups are mainly “shaly sand” sequences, an insight into the neutron-absorbing minerals in the sand reservoirs added a valuable dimension in reservoir characterization. Field examples are presented. The technique proved to be a valuable aid in stratigraphic correlation and formation zonation. Furthermore, because of the relatively simple nature of the mathematics involved and the current capability of logging technology to acquire the required measurements in one descent in the well, this technique can be implemented at the wellsite, providing the information in real time during logging. Information available at such an early SPWLA 38th Annual Logging Symposium, June 15-18, 1997 stage increases the efficiency of the decision-making process. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER M SIMULATION OF INDUCTION AND MWD RESISTIVITY TOOLS IN ANISOTROPIC DIPPING BEDS Stephane Graciet and Liang C. Shen University of Houston ABSTRACT Shale formations as well as thinly laminated sequences of shale and sand are known to be anisotropic formations often encountered in electromagnetic logging. Also, many wells are deviated from the direction perpendicular to the formation beds. In this paper, the combined effect of dipping and anisotropic bed on the readings of induction and measurement-while-drilling (MWD) tools is studied. Solutions of Maxwell’s equations in anisotropic media are obtained and computer codes are developed to model responses of induction and MWD resistivity tools. The following conclusions are obtained based on simulation results. (1) The induction and MWD tool responses in thin sand-shale sequence and in anisotropic beds show different characteristics even when the sand-shale layers are as thin as 6 inches. Of course, the difference disappears when the sand-shale layer thickness is much thinner than 6 inches. (2) In the case of an MWD tool, the anisotropy can be identified by the difference between the phase-based and the amplitude-based apparent resistivities (R~ and R3 and the difference increases with the dip angle. (3) Although a simple equation can be used in order-to take anisotropy into account in thick beds, in thin anisotropic beds, it becomes necessary to make the correction by using a correction chart. A typical correction chart is presented to demonstrate that the magnitude of the correction is very significant (4) A set of conversion charts has been prepared to convert MWD logs (Rp and Ra) to Rh and Rv, the horizontal and vertical resistivities of an anisotropic bed. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER N MODELING 3D EFFECTS ON 2-MHZ LWD RESISTIVITY LOGS B. I. Anderson, V. Druskin, F. Lee, M. G. Luling, F. Schoen, J. Tabanou, F. Wu Schlumberger S. Davydycheva, L. Knizhnerman Central Geophysical Expedition, Moscow ABSTRACT Steering directional wells through complex reservoirs frequently requires the use of 2MHz logging-while-drilling (LWD) resistivity tools. In such situations, tools can encounter formation heterogeneities such a-s faults, asymmetric invasion and skewed fractures. Because vertical-well interpretation techniques assume azimuthal symmetry, they can result in significant errors when extended to high-angle wells. Azimuthal symmetry breaks down in deviated wells, where nearby beds parallel to the tool can cause polarization horns, and triaxial anisotropy has an increasing effect. In complex, deviated wells, 3D forward log simulation offers a reliable interpretation method for estimating formation resistivity We introduce a new modeling program that computes 2-MHz tool response to fully 3D formation geometries with triaxial anisotropy. This program uses the spectral Lanczos decomposition method to solve Maxwell’s equations on a staggered finitedifference grid. The program is similar to a program recently introduced for the interpretation of wireline induction logs in horizontal wells. A workstation-based graphical interface simplifies the description and visualization of 3D geometries. This interface allows the user to assemble a formation model from elementary building blocks, such as layers, boreholes, arbitrarily shaped invasion fronts and skewed fractures. The graphical interface translates this formation model into an input file read by the modeling program. This graphical interface will become a module in an integrated reservoir description and analysis package. Effects of 3D formation features on 2-MHz field logs have been reproduced by modeling. Effects studied include invasion in horizontal wells, oil-base-mud-filled fractures, noncircular invasion fronts and dipping invaded beds with lamination anisotropy. A study of invasion in horizontal wells shows that while the deepestreading curves are usually unaffected by invasion, they are influenced by proximity to adjacent layers located as far as 15 ft from the wellbore. At the same time, the shallowest-reading curves give a reliable interpretation of invasion until the tool is within 2 ft of an adjacent layer. For non-circular invasion fronts, modeling shows an increasing influence of the formation beyond the invaded zone on the shallowestreading logs. However, the effect is so small that it can be ignored to first order in invasion interpretation. Drilling-induced fractures filled with oil-base mud can cause curve separations that resemble invasion profiles. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 Complex 3D resistivity interpretation problems can now be resolved with modeling, both for LWD and wireline tools. In addition, 3D modeling allows a better understanding of tool physics and provides the basis for the design of new tools and interpretation methods. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER O AUTOMATIC HIGH RESOLUTION SEDIMENTARY DIP DETECTION ON BOREHOLE IMAGERY Shin-Ju Ye Image Laboratory, Institute of Geodynamics, Talence, France Philippe Rabiller and Noomane Keskes CSTJF, ElfAquitaine Production, Pau, France ABSTRACT This paper details a new automatic processing methodology to extract the bed boundaries and the sedimentary dips from the borehole images. Instead of using side by side and pad to pad correlation, the proposed methodology is based on edgematching and statistical analysis of the full image, thus it takes full advantages of the high resolution of the tools and of the possibility to discriminate between different families of planar features and marks (vacuoles, nodules. etc.). In the first stage, an original filtering technique is used to generate connective component images, from which edges of bed boundaries can be efficiently detected in each pad. Several criteria are then developed to consolidate the edge-matching process for plane construction. Compared with methods using Hough Transform, which is time consuming, this method works much faster. In the second stage, high frequency stratification planes are detected using statistic methods. Local orientations are computed from pad images by Minimum-Variance method to determine the sinusoids. Lamination or stratification planes are then fitted on maximum local contrasts in order to simulate the interactive hand-picking. Together with the dips, several associated “quality indexes” are computed. They take into account the correlation between pads, the contrast along the bed boundary or stratification on all pads, and the roughness of the plane approximating the bed boundary or stratification. Those indexes bear some geologic meaning complementary to that of dip values and trends. Thus they can be used to help characterize the sedimentary processes and depositional environments. These methods for bed boundary and lamination detection are robust and easy to operate. They are not sensitive to occurrences of fractures (both open and sealed) and vacuoles or nodules, which is an important advantage for the geologic interpretation of the dip trends. The reliability and the consistency of our methodology are demonstrated by the operational work carried on more than 20 runs of microresistivity imagery. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER P PARAMETRIC STUDY OF CEMENT BOND EVALUATION USING EARLY REFRACTED ARRIVALS R. Rao V. N. Earth Resources Laboratory, Massachusetts Institute of Technology, Cambridge, MA B. Mandal Halliburton Energy Services, Houston, TX C. H. Cheng and M. N. Toksöz Earth Resources Laboratory, Massachusetts Institute of Technology, Cambridge, MA ABSTRACT The cement bond evaluation tool is a device used to examine the integrity of cement bonding to the casing. A conceptual tool operating between 80 - 200 kHz is considered here, with a transmitter and two receivers, oriented parallel to the axis of the borehole and next to the casing. The compressional head wave in the casing, excited by the transmitter, will be the first arrival to be measured by the two receivers in most situations. With both receivers on the same side of the transmitter, the attenuation of this wave in traveling between the two receivers is dependent on the properties of the medium immediately outside the casing. The radially layered borehole was modeled as a layered plane medium for large operating frequency. A spectral integral approach (complete wave synthesis) was used to compute the response at the receiver locations, which then provided attenuations. Different parameters, such as, transducer separation (1 - 12 inch), annulus thickness (0.6 inch), annulus impedance (free pipe to good cement), casing thickness (0.25 - 0.45 inch), standoff distance (0.5 - 1 inch) and source frequency (80 - 200 kHz) were varied in the evaluation of the operation of the tool. The parameter studies based on the theoretical computations revealed that free pipe could be distinguished from the presence of cement in a variety of situations. Additionally, lower bounds on receiver separations are given for reliable operation of the tool. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER Q EXTRAPOLATION OF CORE PETROPHYSICAL GROUPS TO UNCORED WELL INTERVALS READY TO USE SAS PROGRAMS H. N. Greder Elf Exploration Production, CSTJF, Pau Cédex, France F. Genissel Cap Gemini, Pau, France D. Samba and I. Branger University of Pau, Pau, France Ph. Lefebvre and F. M. Pellerin Elf Exploration Production, CSTJF, Pau Cédex, France ABSTRACT In complex reservoirs, where petrophysical properties are related to both sedimentary and diagenetic events, reservoir modelers are more and more inclined to define Rock Types by grouping core analysis data rather than using geological facies that are not closely related to flow properties. However, prior to the 3D mapping of such RockTypes, a continuous log of core petrophysical groups is needed on both cored and uncored wells. This paper presents three programs which were written with the increasingly popular statistical SAS software and which enable such extrapolation. The construction of core petrophysical groups, based on core analysis data, is explained and referenced. Their practical use in reservoir modeling is also referenced. The paper focuses on the method to extrapolate these groups to uncored well intervals. The method which uses both interpreted and non interpreted logs as predictors is based on discriminant analysis and respects the following steps: Firstly, all available logs are ordered in the cored intervals according to their correlation with the petrophysical groups. The more correlated logs are placed in first positions. Then a criterion, based on the statistical << Wilks’ lambda>> is provided in order to help the log analyst to retain a limited number of logs among the list. Reducing the number of predicting logs has two advantages: 1- it avoids redundant information in the model and 2- it facilitates the quality control of the prediction in non-cored intervals as the log analyst only focuses on a limited number of logs. Secondly, the mathematical model is built between the retained predicting logs and the actual petrophysical groups in the cored intervals. This model is based on discriminant analysis. The quality of the prediction is evaluated both by crossvalidation and simulation techniques which are commonly used in geostatistical tools. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 Thirdly, when predicting a petrophysical group in an uncored interval, the program provides two criterions: one indicates the probability that the predicted group is the actual one. The other indicates if the cores are representative of the logged intervals where the prediction is performed. Those criterions are very useful to evaluate the confidence of the extrapolation. All the steps and criterions are detailed on a mathematical point of view and three examples of such extrapolation are provided. Examples originate from three different reservoirs: a mixed lithology reservoir, a vuggy carbonate reservoir and a shaly/sandstone reservoir. Listings of the three programs are provided in the Appendix. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER R FRACTURE AND PERMEABILITY EVALUATION IN A FAULT ZONE FROM SONIC WAVEFORM DATA Takeshi Endo Schlumberger K.K., Wireline & Testing, Fuchinobe, Japan Hisao Ito Geological Survey of Japan, Tsukuba, Japan Alain Brie Schlumberger K.K., Wireline & Testing, Fuchinobe, Japan Mohammed Badri Schlumberger Logelco Inc., Wireline & Testing, Cairo, Egypt Mohamed El Sheikh Agiba Petroleum Company, Cairo, Egypt ABSTRACT Identification and evaluation of fracture systems are important in oil and gas exploration in hard-rock areas and for scientific drilling. In fracture systems, faults are major events that impact not only the fracture distribution but also the rock structure and stresses. Near faults are significant rock deformation, fracturing and variations of the stress field. Faults have, therefore, large effects on the producibility and stability of a reservoir and must be accounted for when completing the well. We integrated various sonic interpretation techniques to evaluate these effects. Stoneley wave reflections and attenuation analysis are known techniques for fracture evaluation. They have the advantage of showing open permeable fractures but also are sensitive to borehole irregularities. Stoneley modeling, when added to these techniques, estimates the effect of the borehole and improves the reliability of Stoneley fracture evaluation. Stoneley permeability analysis evaluates the slowdown of the Stoneley wave to indicate fluid mobility in the pore space, both from fractures and porosity. The recently introduced dipole shear anisotropy evaluation provides information on aligned cracks and stress directions. The technique evaluates shear wave splitting resulting from acoustic anisotropy. These techniques react in different ways to the presence of fractures in the formation. Combining these indications provides additional information on the reservoir characteristics, especially the location of fault zones. We applied these techniques in two wells. The first well was drilled for scientific purposes to intercept a known large active fault in a granodiorite formation. In this data set, near the fault the Stoneley wave shows evidence of strong fracturing and deformed zones. Acoustic anisotropy indicates significant perturbations of the stress field. Stoneley permeability analysis detects high fluid mobility. The second well was SPWLA 38th Annual Logging Symposium, June 15-18, 1997 drilled in a granite formation in an oil field. We observed similar signatures in the Stoneley and anisotropy results that strongly suggest the presence of a fault intercepting the wellbore. In this data set, systematic variations of the fracture’s dip and strike, consistent with sonic evaluation results, are also observed on the microresistivity images. High mobility indications from Stoneley analysis are confined by the production data. Integrating anisotropy information with conventional fracture evaluation techniques uncovers new possibilities for reservoir evaluation. Fractures can be identified and better understood with the resulting fracturing and rock alteration to give new insights on reservoir properties and producibility. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER S THE APPLICATION OF THE MATHEMATICS OF FUZZY LOGIC TO PETROPHYSICS Steve Cuddy BP Exploration, Aberdeen, UK ABSTRACT Certain areas of petrophysics have benefitted from the application of the new mathematics of “fuzzy logic”. BE’ Exploration routinely uses new interpretation techniques, based on fuzzy logic, to predict permeability and litho-facies in uncored wells. Lithofacies and permeability prediction has presented a challenge to Petrophysics due to the lack of tools that measure them directly and because of the inherent errors in any measurement. Fuzzy logic is simply an application of recognized statistical techniques. Whereas conventional techniques deal with absolutes, the new methods carry the inherent error term through the calculation rather than ignoring or minimizing it. This retains the information associated with the error and gives surprisingly better results. The first application is to litho-facies determination. Litho-facies typing is used in well correlation and is important for building a 3D model of the field. The technique makes no assumptions and retains the possibility that a particular litho-facies type can give any log reading although some are more likely than others. This error or fuzziness has been measured and used to improve the facies prediction in several North Sea fields. In one study, descriptions from 10 cored wells were used to derive facies descriptions in 30 uncored wells. This technique gave near perfect differentiation between aeolian, fluvial and sabkha rock types. In addition, it went some way towards differentiating between sandy, mixed and muddy sabkhas. Using the fuzzy logic technique gives better predictions than more complicated methods. Other applications include permeability calculation, thin bed resolution and fracture prediction. Knowledge of permeability is important in determining well completion strategy and the resulting productivity. The problem with permeability prediction is derived from the fact that permeability is related more to the aperture of pore throats rather than pore size, which logging tools find difficult to measure. Determining permeability from logs is further complicated by the problem of scale; logs having a vertical resolution of typically 2 feet compared to the 2 inches of the core plug. The new techniques quantify these errors and use them, together with the measurement, to improve the prediction. In several fields in the North Sea, this new approach has given better estimates of permeability compared to conventional techniques such as multiple linear regression. In addition, the method relies on basic log data sets such as gammaray and porosity rather than depending on new logging technology. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER T REAL TIME INTERPRETATION OF MWD ANISOTROPY IN HIGH ANGLE WELLS, OFFSHORE GULF OF MEXICO Victor Rosato Spirit Energy 76, Lafayette, Louisiana Joseph Beck Sperry-Sun Drilling Services, Lafayette, Louisiana ABSTRACT The purpose of this article is to tabulate the range of anisotropy observed on real time MWD logs, offshore Gulf of Mexico, in order to aid in the interpretation of the data. A table is presented which includes age, depositional environment, and anisotropy ratios, at various relative dip angles. Log examples and their models are also presented. The table can be used to predict the anisotropy ratio a formation may exhibit. Anisotropy is a condition where sensors exhibit different results when measuring a unit volume from different directions. This effect has been observed on numerous real time MWD logs with high relative dip angles. Predicting anisotropy effects in a formation can be extremely important in determining zone tops and geosteering within the objective zone at high relative dip angles. There is not an easy way to predict the anisotropy ratio (Rv/Rh) without a high relative dip angle hole, extensive core analysis, or forward modeling. In older fields without modern logs or core material, pre-drill models require numerous assumptions. References cite observed Rv/Rh ratios as great as 50 to 1 (Lesso and Kashikar, 1996). A rule of thumb for pre-drill models in low resistivity pay zones is to begin with ratios of 3 for shales and 1 for sands. In higher resistivity pay zones a ratio of 2 for shales and 6 for sands is used. Observed anisotropic effects vary by field, formation age, depositional environment depth of investigation, and apparent dip angle. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER U PETROPHYSICAL INPUT TO SIMULATION: DERIVE ROCK TYPE MAPS FROM LOGS WATER SATURATIONS AND CAPILLARY PRESSURE DATA. A. G. Jacques Elf Exploration UK PLC, Aberdeen, UK ABSTRACT Rock types in reservoir simulators are usually defined to describe different geological facies of a similar dynamic behaviour. A capillary pressure versus water saturation curve and a set of rel-perm curves will be allocated to each rock type. For both types of curves the irreducible water saturation will be used to define the curves’ endpoints. Reservoirs are usually composed of a wide variety of rock types varying both vertically and laterally. While it is difficult to predict in advance how such a complex distribution should be modelled, the common approach is to group the individual sets of lab capillary data by rock types, combine rock types when required for calculation constraints and eventually compare the computed water saturations with the by-layer water saturation values derived from open hole log data Rather than group rock types by permeability ranges, porosity ranges or geological facies, the basic idea of the proposed method is to obtain rock type maps, consistent with the dynamic definition of a rock type in a reservoir simulator, by mapping irreducible water saturation for each reservoir layer. Logs of Irreducible Water Saturation, similar to Water saturation logs regardless of the height above the contact, are mathematically derived from log and core data. Correlations of the shape and the displacement pressure of the associated capillary pressure curves at the reservoir model layer scale with permeability, porosity and clay volume enable the mapping of the rock types for each reservoir layer. Finally rock types maps with their associated capillary curves will be used to calculate water saturation by grid block with respect to their height above the contact. The proposed method has the advantage of being able to define and correlate the different rock types in a reservoir regardless of the height above the contact. The paper details one field case where this method was successfully applied from the log to the reservoir scale. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER V ACCURACY — ESSENTIAL INFORMATION FOR A LOG MEASUREMENT Philippe P. Theys Anadrill ABSTRACT Two previous papers have expanded on the need for the rigorous quantification of the radius of investigation, vertical resolution and precision of a log measurement. The trilogy is completed by this paper investigating the measurement attributes that make a log accurate. Log accuracy depends on three factors: the robustness of the tool response definition, the way the measurement is calibrated and the impact of environmental (and other) corrections. This paper describes how these effects can be quantified and entered into a specification to provide traceability and substance to the claims made about a measurement. This paper also investigates what accuracy brings to the log user and why it may be sacrificed to other measurement attributes such as precision and vertical resolution. Guidelines are presented for the selection of logging measurements that minimize the uncertainty of the petrophysical parameters in different conditions. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER W COMPLEX RESERVOIR EVALUATION IN OPEN AND CASED WELLS John P. Horkowitz Schlumberger Wireline & Testing, Midland, Texas Darrel E. Cannon Schlumberger Wireline & Testing, Houston Product Center, Houston, Texas ABSTRACT Hydrocarbon-bearing formations in West Texas commonly contain radioactive sands and carbonates as well as varying amounts of anhydrite and gypsum, making the distinction between productive and nonproductive intervals difficult, especially through casing. The ability to identify and quantify hydrocarbons in these mixedlithology, low-porosity reservoirs has changed with the exciting development of spectral tools that accurately determine lithology in both open- and cased holes. The lithology determined is based on elemental concentration logs obtained from inducedneutron gamma ray spectroscopy measurements. When this lithologic information is combined with other measurements, definite improvements in the accuracy of porosity, water saturation and productivity predictions are observed. Stratigraphic interpretation is also significantly enhanced, especially in cased wells. SpectoLith lithology algorithms were developed using a core database of more than 400 samples characterized by dual-range Fourier transform infrared mineralogy and chemical analyses. Total clay dry weight percent is derived from elemental concentrations of silicon, calcium, and iron; anhydrite and gypsum dry weight percents are determined from sulfur and calcium concentrations; and carbonate dry weight percent is determined by calcium concentrations. The remainder is assumed to be composed of the minerals quartz, feldspar and mica. Applications of this technology have successfully 1) quantified clay volumes independent of gamma ray, spontaneous potential and density-neutron, 2) determined permeability and bound fluid based on mineralogy, 3) derived spectral-based foot-byfoot sigma matrix values for accurate sigma based saturation analyses behind pipe, 4) lead to the modification of completion techniques in clastics based on improved clay and carbonate volumes and 5) provided chemostratigraphy information for enhanced geologic correlation and reservoir modeling. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER X PRODUCIBILITY ESTIMATION FROM ARRAY-INDUCTION LOGS AND COMPARISON WITH MEASUREMENTS—A CASE STUDY T. S. Ramakrishnan Schlumberger-Doll Research, Ridgefield, CT J Al-Khalifa and H. H. Al-Waheed Saudi ARAMCO, Saudi Arabia C. Cao-Minh Schlumberger Middle East, Al-Khobar, Saudi Arabia ABSTRACT A method to deduce two-phase mobilities from an array of resistivity logs with different depths of investigation has been proposed (Ramakrishnan and Wilkinson, 1996a). The basis for this interpretation method is that invasion is a downhole flow experiment. The resistivity profile from this invasion process contains information that can be inverted to predict how the well would perform under production, provided the displacement process is similar to that of invasion. The inversion procedure yields fractional-flow curves, filtrate loss, and residual and movable saturations at every depth. To facilitate comparison between production data and the interpretation, we present a two-phase flow model, in which the fractional-flow curves obtained at every depth are converted to cumulative oil and water flow rates inside the wellbore. The model assumes a commingled system, with a constant pressure drive. The numerical algorithm to compute the phase flow rates in the wellbore can be run in two different modes: (1) when a single-phase permeability is available at every depth and (2) when the measured total (oil+water) rate is used as an input. The algorithm is flexible enough to accommodate the low reliability of logs across shoulder beds. An arbitrary combination of layers in which the log-based interpretation may be superseded is available to the user. We discuss the application of the flow-based interpretation method in a Middle-East well, in which the AIT Array Induction Imager Tool and the MicroSFL tools were run. With a known petrophysical relationship for the field, the log data were processed to deduce the two-phase flow characteristics. The processing yielded the hydrocarbon and water intervals, and the permeability barriers successfully. The water-producing interval was identified to be due to seawater encroachment. Because the well was produced with an openhole completion, we could directly compare the production log-based oil and water flow measurements with the model predictions. The production log algorithm was run with both of the above-mentioned modes. Except for some minor anomalies, the agreement between the predictions and the measurements is good. Finally, we show how the laboratory core fractional flow data and the log-processed results match one other. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER Y ACOUSTIC VELOCITY AND POROSITY SYSTEMATICS IN SILICICLASTICS Lev Vernik ARCO Exploration and Production Technology Company ABSTRACT Using several well characterized data sets, including logs and laboratory velocity measurements in fluid-saturated sandstone/shale sequences, it is shown that improved porosity and lithology prediction from sonic logs is possible if unique trends related to sediment deposition, compaction, and diagenesis are recognized. Realistic, petrographically observed evolution of the pore geometries in consolidated, grainsupported sandstones (less than 25-30% porosity and less than 15% clay volume) can be successfully modelled using effective medium theories resulting in a quasi-linear velocity-porosity relation deviating from linearity within the accuracy of velocity measurements. These relations are (1) well defined for clean arenites (essentially pure quartz sandstones) and arenites (slightly shaly sandstones with Vcl <= 12%), (2) largely stress-independent, and (3) yield an improved porosity prediction when compared to the Raymer-Hunt-Gardner equations. Poorly consolidated grain-supported sandstones and sands are characterized by a much steeper gradient of velocity increase with porosity reduction as compared to their consolidated counterparts. The effect can be explained by competing mechanisms of grain rearrangement and initial cementation during early diagenetic history of the sediments. The impact of clay in poorly consolidated sands is diminished, while the dependence on the grain sorting, loading history, and pore fluid chemical activity is increased. The compilation of experimental and log data on acoustic velocities in essentially clay matrix-supported siliciclastics (Vcl > 12%), including shales and wackes (substantially shaly sands), suggests that their bedding-normal elastic stiffnesses vary non-linearly in the whole range of porosity reduction from around 80% to almost zero. Theoretically motivated, empirical models are given for the day frame elastic moduli of the three major petrophysical groups of sandstones (clean arenites, arenites, and wackes) in the porosity range from zero to 40% which can be utilized in porosity prediction or fluid-substitution modeling. Gassmann equation-based fluid-substitution modeling for arenites and comparison with water saturated core measurements at ultrasonic frequency suggest that (1) a substantial pore fluid sensitivity is only typical of poorly consolidated sandstones and sands, and (2) frequency-related velocity dispersion in poorly consolidated sands is relatively minor. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER Z AN IMPROVED METHOD FOR QUANTIFYING IN-SITU INTERMEDIATE WETTABILITY USING WELL LOGS Ali A. Garrouch and Abbas A. Alikhan Petroleum Engineering Department, Kuwait University, Safat, Kuwait ABSTRACT This paper examines experimentally the effects of fractional wettability on the saturation exponent of Archie’s equation, also commonly known as the resistivity index exponent. A four-electrode experimental setup has been used to measure glass-bead-pack resistivity using both the dynamic displacement technique and the steady-state technique. Mixtures of oil-wet and water-wet beads were used covering the entire wettability spectrum. Wettability of the glass beads was reversed using both asphaltic crude oil and Quilon-C in isopropyl alcohol solution. Similar experiments were also duplicated on Berea sandstone core samples. A linear relationship between the resistivity index exponent and fractional oil wettability was observed. This exponent seemed to vary between values of approximately 2.0 for water-wet conditions and approximately 5.0 for oil-wet conditions for both glass-bead-packs and Berea sandstone rock samples. This dependence between degree of wettability and saturation exponent can be used to estimate in-situ reservoir rock wettability by combining data from EPT and resistivity logs for the same formation. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER AA AN EXPERIMENTAL INVESTIGATION OF METHANE IN ROCK MATERIALS Christian Straley Schlumberger-Doll Research Center, Ridgefield, Connecticut ABSTRACT High-pressure, nuclear magnetic resonance (NMR), laboratory measurements of methane have been made with the goal of improving the understanding of logging measurements made in gas fields. Preliminary measurements of bulk methane at 2k, 3k, and 4k psi agree with literature values for relaxation time, methane density, and diffusion constant. These measurements demonstrate the utility of the pressure cell, the high-pressure gas lines and, the gradient coils, which had to be built in-house. Measurements on dry sandstones and carbonates pressurized with methane show that surface relaxation dominates the T1 relaxation of the methane, and internal magnetic gradients affect the measured T2 relaxation times. Methane in dry carbonate rock shows weaker surface-induced relaxation and has insignificant effects due to internal gradients when compared to sandstones. In partially water-saturated rock systems, surface-relaxation also dominates the relaxation of methane in both sands and carbonates, contrary to the assumptions usually made for theoretical models of gas/rock systems. Additionally, in sandstones the effect of internal gradients on the methane T2 is important in the partially watersaturated rock. The comparison of measurements with and without applied field gradients illustrates differences between internal and applied gradients. The T2 relaxation due to diffusion of methane in the internal gradients is not described by free diffusion in a linear gradient field. In contrast, in applied linear gradient fields, the gas peak in rock is shown to shift due to diffusion as anticipated when the contributions of bulk relaxation, surface relaxation and diffusion in internal field gradients are removed. Laboratory examples of the Differential Spectrum Method and Shifted Spectrum Method for detecting natural gas are provided. However, because of the surface relaxation component of methane relaxation in rock materials, methane resonances will be found at shorter relaxation times than previous estimates based on applied gradient strengths and bulk methane parameters. Qualitatively, the unexpected shifts can lead to confusion when using the current interpretation models, and quantitatively, attempts to compensate for incomplete polarization using only bulk methane relaxation parameters will lead to overestimates of gas porosity. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER BB MULTIWELL LOG DATA COHERENCE CHARACTERIZATION USING THE SIMILARITY THRESHOLD METHOD Hugues Thevoux-Chabuel Laboratoire d’Informatique Appliqude aux Sciences de la Terre, Université de Paris, Paris, France Antoine Veillerette and Philippe Rabiller CSTJF, Elf Aquitaine Production, Pau, France ABSTRACT Field-wide interpretation methods rely on model calibration in one or more key wells. Whether the model can be successfully propagated to all other wells depends on the similarity of the key wells to the rest of the data set-s. The Similarity Threshold Method (STM) is a new technique used for checking the similarity of two logged intervals. Our experience acquired from reservoir studies shows that this method is more efficient and easier to implement than conventional empirical methods (crossplots) or statistical methods (principal component or regression analyses). Based on the k-nearest neighbors theory, this method characterizes the lithological response of the reference intervals in ‘n’ dimensions log space by integrating all the log data. The application data sets are then compared against the reference set for each depth level and a decision rule is applied to determine whether the depth level can be represented by the reference set. If a depth level is rejected it is possible to tell which logs are causing its rejection. The STM method is applied as a preliminary process before field-wide application of an interpretation model in order to check the coherence of the data. On a field, containing 37 wells, a permeability prediction model has been generated from core data. The STM method was used to see for what non cored wells and intervals the model could be applied. The method has the added benefit of finding new facies variations, not seen in the reference data, or being able to distinguish badly calibrated tools and poor log response due to bad hole conditions. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER CC IMPROVED ESTIMATION OF HIGH-RESOLUTION FORMATION RESISTIVITIES FOR THIN-BED ANALYSIS M.A. Frenkel, AG. Mezzatesta, and K.-M. Strack Western Atlas Logging Services, Houston, Texas ABSTRACT During the past decade, oil production from thin laminated formations has become increasingly important. An accurate, high-resolution description of the borehole surrounding is of fundamental importance in dealing with the interpretation and evaluation of thinly laminated reservoirs. Besides the imaging devices, not many logging tools are commercially available that provide the vertical resolution and depth of investigation of the Thin Bed Resistivity Tool (TBRTSM). An interpretation technique, based on inversion methods, has been developed to interpret the TBRT measurements, providing a resistivity distribution with resolution of 2 inches and higher that can be used for net pay count in laminated reservoirs. Parameters such as borehole rugosity and invasion must be considered in the interpretation scheme to accurately evaluate the spatial distribution of formation resistivities in the borehole surroundings. Two-dimensional numerical modeling of TBRT data, combined with inversion methods, allow for simultaneously considering the effects of borehole rugosity, shoulder beds, and invasion, leading to an accurate, higher resolution description of the thin-bed formation resistivities. The computer resources required to solve an inverse problem grow very rapidly with the dimensionality, formation model size (number of parameters), and the number of logging depths used in the interpretation. A fast-inversion technique has been developed that requires significantly less numerical calculations than standard inversion methods and is applied to the interpretation of TBRT data. To validate the interpretation technique, a number of typical cases involving environments with different rugous boreholes, invasion, and shoulder beds, have been studied and are presented here. The interpretation technique has also been applied to field cases, whose results show how measurements of different resolution can be combined into a single and efficient interpretation process, which provides an accurate thin-bed formation resistivity estimation for a further enhanced evaluation of residual and movable hydrocarbons. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER DD DRILL-BIT SEISMIC A SERVICE FOR DRILLING OPTIMIZATION Masahiro Kamata Schlumberger K.K., Wireline & Testing, Fuchinobe, Japan William Underhill Schlumberger Well Services, Sugar Land, Texas, Richard Meehan Schlumberger Cambridge Research, Cambridge, UK Lea Nutt Schlumberger Technical Services, Montrouge Cedex, France ABSTRACT The recently-developed Drill-Bit Seismic technique utilizes the acoustic energy radiated during the drilling process to provide vital information about the subsurface structure. This information, produced in real time at the wellsite, can be used to optimize the drilling process, leading to significant cost savings and enhanced safety. Acoustic energy radiating from the bit is recorded by sensors both at the top of the drill-string and placed on the ground in the vicinity of the rig. The signal recorded by sensors on the top of the drill-string is analyzed to give accurate travel time information within the drill-string and also the nature of the reflections generated within the drill-string. Signals from the subsurface recorded by the sensors on the ground are dominated by very large “ground roll” noise emanating from the rig. State-of-the-art signal processing techniques have been developed to improve signal-to-noise ratio from these recorded signals. Travel times recorded by the sensors on the ground are corrected for drill-string travel times and subsequent VSP processing of the data can lead to information about the nature of reflections both generated within and below the bottom of the well. Extensive field tests conducted in a wide range of both land and offshore environments confirm the ability of this new technique to provide reliable time-depth and formation velocity information at the wellsite, allowing the position of the bit to be plotted on the surface seismic section as the well progresses. Examples will be shown from recent land and offshore jobs where the Drill-bit Seismic technique has been used whereby generating look ahead VSP images, the approach to critical horizons, e.g., casing points, overpressure zones, or coring point, was monitored. The value and cost implications of being able to make these predictions accurately will also be discussed. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER EE APPLICATIONS AND INTERPRETATION OF AZIMUTHALLY SENSITIVE DENSITY MEASUREMENTS ACQUIRED WHILE DRILLING William W. Carpenter Schlumberger Wireline & Testing David Best and Mike Evans Anadrill ABSTRACT Azimuthal formation density measurements made while drilling are being used to add detail and accuracy to the characterization of formation properties and near-borehole environmental conditions. Crossing thin beds at a high relative angle to the wellbore trajectory causes a smearing of the response of circumferentially averaged density measurements. By examining the azimuthal densities, bed boundaries are accurately detected and discrete bed properties correctly measured. Fluid invasion affects density logs as a function of azimuthal fluid migration and time after drilling the formation. Interpretation of azimuthally acquired density data assists the user in properly identifying formation fluid types. The effects of time can be recognized through time lapse logging to develop a better understanding of a reservoir’s dynamic behavior. Actual field examples are shown to illustrate the use of azimuthal density measurement to enhance formation evaluation. Images of density, photoelectric factor (Pef), and delta-rho are used as log analysis tools, differentiating formation responses from borehole-induced responses. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER FF INTERPRETATION AND ANALYSIS OF SONIC-WHILE-DRILLING DATA IN OVERPRESSURED FORMATIONS K. Hsu Anadrill, Sugar Land, Texas M. Hashem Shell Offshore, New Orleans, Louisiana C. L. Bean Chevron USA Production, New Orleans, Louisiana R. Plumb Schlumberger Cambridge Research, Cambridge, UK G. N. Minerbo Anadrill, Sugar Land, Texas ABSTRACT With the availability of sonic and other logging-while-drilling (LWD) measurements, the detection and estimation of overpressured formations in realtime at the wellsite is now possible and significantly improves the efficiency of drilling, casing and completion operations. A suite of sonic and other LWD data acquired in an overpressured formation in the Gulf of Mexico is used to demonstrate this capability. Contrary to a normal compaction trend, the ∆t values measured while drilling increase with depth and approach abnormally high values - an indication of overpressured formation. The overpressure trend is confirmed by other LWD and drilling measurements. This divergent trend is further corroborated by the wireline sonic and neutron-density porosity logs acquired two weeks after drilling. It is observed that the LWD and wireline sonic measurements are in a close agreement in the interval above the overpressured formation. Detailed features in the two logs are well correlated but the two measurements show a small and consistent difference in the overpressured formation, with the LWD at measurements slower by a few microseconds per foot Given limited physical information, the most likely cause of this time dependent decrease of ∆t (or increase of velocity) is an increase of effective stress and a concomitant reduction of porosity near the wellbore. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER GG MULTI-PARAMETER PROPAGATION RESISTIVITY INTERPRETATION W. Hal Meyer Baker Hughes INTEQ ABSTRACT When propagation resistivity tools made just one measurement, only the apparent resistivity could be determined. When they made two measurements, another parameter could be determined by assuming that the separation between the two measurements was totally caused by a single effect. However, many current propagation resistivity tools make several measurements. The primary purpose of multiple measurements is to be able to determine several parameters simultaneously. The effects that cause separation in the propagation resistivity measurements include anisotropy, eccentricity, anomalous dielectric effects, vertical resolution differences, borehole conductivity, and invasion. Some of these effects, such as borehole conductivity and vertical resolution differences, can be removed accurately without calculating other parameters simultaneously. Other effects (such as anisotropy, anomalous dielectric constant, and invasion) must be interpreted simultaneously because an inaccurate value of one parameter will cause large errors in the others. After correcting for these effects, the data can be combined to produce apparent resistivities at several fixed depths of investigation. The result is a display much like the log of modem induction tools. The propagation resistivity data compare favorably to induction logs in the same well. In this study, borehole corrections are first performed on eight compensated measurements. This produces eight very accurate measurements of formation properties. Eight raw phase measurements are also borehole corrected and then “normalized” to one of the compensated curves. This “normalization” is necessary because the raw phase measurements are neither compensated nor accurately calibrated. All 16 of these curves are then inverted to remove the effects of limited vertical resolution. The resulting curves can then be evaluated to determine invasion, anisotropy, and dielectric constant. Shallow invasion will normally affect only the raw phase curves. Anisotropy affects the compensated phases most of all, the compensated attenuations less, and the raw phases least of all. In addition, anisotropy affects longspaced measurements more than short-spaced measurements (the opposite effect of invasion). Anomalous dielectric constants affect the attenuations more than the phases and the two effects depart from the true resistivity in opposite directions. However, dielectric effects are about the same at short or long spacing. Anomalous dielectric constants have very little effect on any of the raw phases. Effects of dielectric constant, invasion, and anisotropy are all less at lower frequencies. A nonlinear regression scheme finds the parameters that fit the data best. After this, the curves can be combined to produce the apparent resistivities at four fixed depths of investigation. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER HH REAL-TIME SONIC LOGGING WHILE DRILLING IN HARD AND SOFT ROCKS J. Aron, S. K. Chang, D. Codazzi, R. Dworak, K. Hsu, T. Lau, O. Minerbo, and E. Yogeswaren Anadrill, Sugar Land, Texas ABSTRACT A new sonic-while-drilling tool can obtain real-time formation slownesses over a wide measurement range. Real-time sonic logs provide valuable data for drilling applications and allow real-time decision-making. This measurement has previously been available only in hard rocks, where downhole semblance processing has been shown to give excellent agreement with wireline logs. The measurement of compressional slowness in slow formations was more challenging because the formation arrival is later in time and often has a lower amplitude. By significantly reducing the collar arrival, real-time sonic measurements in slow formations can now be made accurately and efficiently from the downhole processing. Field test results showing both recorded and real-time data are presented along with wireline and check shot comparisons. Although the logging-while-drilling and wireline logs agree in general, experience shows that there are depth intervals, especially in shales, where the measurements can be different. Several explanations are proposed for these observations. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER II TIME LAPSE OPENS NEW OPPORTUNITIES IN INTERPRETING 2-MHZ MULTISPACING RESISTIVITY LOGS UNDER DIFFICULT DRILLING CONDITIONS AND IN COMPLEX RESERVOIRS Jacques Tabanou, Sue Bruce, Steve Bonner, and Peter Wu Anadrill ABSTRACT Curve separation between 2-MHz resistivity multi-depth-of-investigation logs is often the first indicator at the well site that “something” is happening in the reservoir that does not fit a preconceived paradigm. Such curve separation can result from dynamic hole damage by heavy mud, anisotropy in highly deviated wells or permeability variations in carbonates. The thrust of this paper is to emphasize the importance of time-lapse logging. Time-lapse logging helps in differentiating curve separation caused by reservoir geometry, such as layered reservoirs drilled in high-angle wells, from separations induced by fluid displacement and formation damage. Several examples of ARC5 Array Resistivity Compensated tool logs are discussed that demonstrate the power of 2 MHz multidepth of investigation resistivity logs when combined with advanced modeling capabilities and multipass logging. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER JJ MEASURING RESIDUAL OIL SATURATION IN WEST TEXAS USING NMR M. B. Crowe Chevron Petroleum Technology Company D. E. Jasek, S.C. Van Dalen Chevron U.S.A. G. A. LaTorraca, K. J. Dunn, and M. T. Donovan Chevron Petroleum Technology Company ABSTRACT This paper presents a NMR log-inject-log method (NMRLIL) to estimate both permeability and residual oil in carbonate wells at multiple depths of investigation. Earlier methods have been published on how to determine residual oil using NMR. Our approach involves a two step process of coring and NMR logging followed by under-reaming with doped mud, and relogging using both low and high frequency MRIL tools. NMR core measurements have been used to improve the interpretation. The method includes NMR (Nuclear Magnetic Resonance) logging both before and after doping the mud with MnCl2. The paramagnetic ion (Mn++) causes the water signal to decay more rapidly allowing the oil signal to be differentiated and measured. After doping the mud, the hole is under-reamed to ensure sufficient invasion of the dopant into the formation before relogging. Comparison of the NMR logs from before and after doping yields the remaining and/or residual oil saturation profile for the well. In one well, both the NUMAR MRIL and Schlumberger CMR logs were run. The results from both tools were consistent except in rugose hole sections. Dual depth of investigation MRIL tools were run to determine the extent of filtrate invasion and whether “remaining” or “residual” oil saturations were being measured. In the upper section of the well the lower frequency (greater investigation depth) tool indicated higher remaining oil saturation than did the higher frequency tool. The difference decreased to zero with increasing depth. This new method has been applied on two wells in West Texas to assess the feasibility of a CO2 pilot project in a dolomite formation. A third well is being planned. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER KK CARBON/OXYGEN LOGGING IN COMPLEX BOREHOLE COMPLETIONS A. Badruzzaman Chevron Petroleum Technology Co. T. Badruzzaman Pacific Consultants and Engineers P. T. Nguyen Chevron Petroleum Technology Co. A. O. Adeyemo Chevron Nigeria Limited M. A. Limon Chevron Overseas Petroleum Inc. ABSTRACT The feasibility of using Carbon/Oxygen (C/O) measurements in complex completions that are beyond standard tool calibration is examined via field measurements and three-dimensional nuclear modeling techniques. The conditions studied include logging in large boreholes with the tool inside the casing only, inside casing and tubing, inside casing and blast joints in the tubing opposite open perforations, etc. C/O tools are not normally characterized in such borehole completions. Consequently, we studied the response characteristics of a dual-detector, slim CO tool using the Monte Carlo radiation transport technique, in conjunction with field measurements. Caveats and limitations of C/O measurements in complex completions identified by the study are described and the on-going research to use the measurements optimally are discussed. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER LL A NEW APPROACH TO FORMATION COMPACTION MONITORING IN A GIANT DEEPWATER GOM OIL AND GAS FIELD DEVELOPMENT Arne J. de Kook Shell Offshore Inc., New Orleans, LA T. Hagiwara, H. Zea, F. Santa Halliburton Energy Services, Houston, TX ABSTRACT Formation compaction in unconsolidated geopressured turbidite reservoirs in the Gulf of Mexico (GUM) is an issue of great uncertainty and concern. Typically, as the pressures decline in oil and gas producing reservoirs, the effective stress on formation grains increases, which causes the reservoir rock to compact. Compaction is most severe when pressure decline is high and the reservoir rock is loosely compacted and uncemented. As compaction occurs, it changes the porosity and permeability properties of the reservoir rock and can affect recovery efficiency and well productivity. Loads imparted by the compacting reservoir rock can deform well tubulars, creating operational problems and shortening well life. If compaction is significant, especially when multiple stacked reservoirs are involved (as in the case in this field example), then compaction can translate to the surface creating a subsidence bowl. In GOM offshore operations, this could cause platforms to subside deeper into the water and therefore potentially affecting its ability to sustain severe wave impacts in a hurricane situation. Therefore, failure to properly address compaction during the design and development phases of these capital intensive Deepwater projects could lead to severe financial setbacks. Because of the broad impact of compaction and seafloor subsidence, it is crucial to have a compaction monitoring program in place. As a result, we developed a new Formation Compaction Monitoring Tool (FCMT) and new methods of measurement and interpretation. The FCM’I’ is a wireline device that uses multiple gamma-ray detectors to determine locations of and precise distance between radioactive markers planted in the formation or on the casing. Compaction of the formation can be measured by changes in the distance between the markers. For precise estimation of the vertical distance between a pair of markers, the new method uses an array of three detectors. To achieve higher accuracy, we studied a method that used a four-detector array. By examining the tool response to a marker, we developed a new method to determine the exact vertical and lateral location of the marker using a Lorentzian response model. Consequently, not only the vertical compaction but also lateral displacement of markers can be monitored with the new method. The accuracy of the tool was established in a test facility where gamma-ray sources were placed at precisely known intervals. We ran the tools at three logging speeds (5, 10, and 15 ft/mm), and collected data at 0.1-in, interval. The vertical distances SPWLA 38th Annual Logging Symposium, June 15-18, 1997 between a pair of radioactive markers spaced 30 ft apart were measured accurately to within 0.1 in. This paper reports the results and conclusions of the successful first four actual logging jobs in Deepwater Gulf of Mexico reservoirs. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER MM FIELD EXAMPLES WITH A NEW COMPACTION MONITORING INSTRUMENT Richard Pemper, Lars Fjell, and Randy Gold Western Atlas Logging Services, Houston, Texas Frode Ramstad Phillips Petroleum Company Norway, Stavanger, Norway ABSTRACT Formations sometimes compress as a result of decreased pore pressure from hydrocarbon extraction. Possible consequences include casing collapse and surface subsidence. An important technique for monitoring this phenomenon involves periodic monitoring of the distance between radioactive markers placed into the formation or onto casing. The resulting data are useful for estimating the compressibility of producing formations and can be used to predict future compaction, subsidence, and hydrocarbon production. A new logging system, the Compaction Monitoring Instrument (CMI), has been developed for this purpose. The average precision of 507 formation intervals measured thus far has been 1.15 mm. The logging instrument incorporates four NaI(T1) gamma ray detectors, temperature and pressure sensors, a tri-axis accelerometer, and two casing collar locators. Special pressure resistant housing joints complete with index marks enable the instrument to be disassembled and reassembled without loss of the detector spacing calibration. As a result, the calibration is performed at the base and does not have to be repeated at the well site. The CMI calibration system utilizes an invar bar with source spacings traceable to the U. S. National Institute of Standards and Technology (NIST). Numerous independent calibrations have shown the measured variation of the detector spacings to be less than 0.2 mm. After a well is logged, gamma ray profiles of subsurface, radioactive markers are used to ascertain their depth locations. These locations are determined by finding the peak of each gamma ray distribution after it is filtered with a Gaussian smoothing function matched to the detector response. The distance between marker pairs is calculated at specified intervals of the well by averaging the measurements from all possible detector combinations. Outliers are rejected on a statistical basis. A special correction is used to translate the cable motion portion of each individual measurement into a more precise distance normalized to the detector spacing calibration. Numerous field examples demonstrate that measurement precision is influenced by irregularity in the velocities of both the tool and depth wheel, by marker signal strength, and by the number of log passes. Specific factors affecting the smoothness of tool velocity include logging cable thickness, winch system configuration, and general borehole conditions such as scale buildup, floating debris, and buckling of the casing. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER NN APPLICATIONS OF AN EPITHERMAL NEUTRON MEASUREMENT IN FORMATION EVALUATION H. Darling, H. Scott and A. Toufaily Schlumberger Wireline & Testing, Houston, Texas ABSTRACT The introduction of an openhole pulsed neutron tool that measures both formation hydrogen index and sigma capture cross section has improved quantitative formation evaluation. The measurement of formation hydrogen index with the new epithermal neutron device has provided an opportunity for the quantitative use of neutron logging not previously possible. To support this development, hydrogen index values for wet and dry clays have been determined to allow improved use of the neutron-density crossplots for determination of clay type, clay volume, formation porosity and flushed zone saturation. Interpretation with thermal neutron logs has been somewhat qualitative for many years because of the difficulty in correcting for the clay density and light hydrocarbon effects and the possible presence of thermal neutron absorbers. Use of the epithermal neutron-derived hydrogen index from the Accelerator Porosity Sonde (APS) eliminates the data shifts or compensation algorithms required previously. Many low-resistivity low-contrast zones have been abandoned as a result of inadequate formation characterization derived using the thermal neutron porosity input. Such zones now can be evaluated properly using the hydrogen index-density comparison in a “quicklook” fashion. The invaded zone gives the first indication of hydrocarbon mobility. Mobility determination is enhanced by neutron tools that have a depth of investigation similar to the density and shallow resistivity. If hydrocarbons are indicated by the saturation results from deeper reading tools, these values can now be compared with the flushed zone hydrocarbon saturation determined from the hydrogen index, density and shallow resistivity measurements. Sigma provides a new parameter for improving openhole interpretation. The use of a sigma-GR crossplot in conjunction with the hydrogen index-density crossplot provides another powerful quickbook interpretation strategy for well site decisions. Examples show this methodology has additional value for identifying the presence of unusual minerals, such as volcanic materials. Differences on the sigma-GR crossplot can indicate where the GR is elevated because of the presence of radioactive minerals or where sigma is reduced by the presence of gas in the flushed zone. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER OO MEASUREMENT OF TOTAL NMR POROSITY ADDS NEW VALUE TO NMR LOGGING R. Freedman, A. Boyd. G. Gubelin, D. McKeon and C. E. Morriss Schlumberger Wireline & Testing C. Flaum Schlumberger-Doll Research Center ABSTRACT The CMR* Combinable Magnetic Resonance logging tool has been upgraded to provide a robust measurement of clay bound water and microporosity. The upgraded measurement is now possible because of a 50% increase in the signal-to-noise (S/N) ratio, a decrease in the minimum echo spacing from 0.32 to 0.2 ins and improved signal processing software that has maximum sensitivity to fast decays. The new CMR hardware and signal processing software reduce the transverse relaxation time (T2) sensitivity limit (minimum detectable T2) of the measurement for depth logging from 3 to 0.3 ms. The total CMR porosity measurement is referred to here as TCMR to distinguish it from previous log outputs. The TCMR measurement adds new value to formation evaluation applications. • • • • Computed clay bound water volumes can be used to compute more accurate hydrocarbon saturations from electrical log measurements. Deficits between density log porosity and TCMR logs in shaly sands can be used to detect the presence of gas and/or hydrocarbon liquids that have a reduced hydrogen index. TCMR provides a lithology-independent porosity measurement that in complex environments can be more accurate than traditional methods based on neutron and density porosity logs. The reduced T2 sensitivity limit of the TCMR measurement extends the range of heavy oil viscosities that can be detected from approximately 1000 to 10,000 cp. T2 distributions computed from laboratory nuclear magnetic resonance (NMR) measurements on a suite of shaly sand reservoir samples with clay bound water saturations ranging from 26% to 49% are shown. The T2 distributions demonstrate the fast relaxation times that can be associated with clay bound water in shaly reservoir rocks. The results show the importance of the reduced T2 sensitivity limit for logging SPWLA 38th Annual Logging Symposium, June 15-18, 1997 total NMR porosity in shaly formations. Monte Carlo simulations were performed using synthetic spin-echo data generated from a suite of 30 model T2 distributions. Some of the T2 distributions have significant signal amplitude associated with relaxation times below 0.3 ms. The spin-echo data were processed using the new signal processing software and a 0.2-ms echo spacing. The simulations show that the T2 sensitivity limit of the TCMR measurement is 0.3 ms for depth logging conditions and 0.1 in for station logging. Field examples from wells in shaly sand formations are used to compare TCMR and effective porosity with an approximate 3-ms T2 sensitivity limit (CMRP) to density and neutron porosity logs. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER PP NMR APPLICATIONS IN THE GULF OF MEXICO Steve Crary Schlumberger Wireline & Testing Freddie Pellegrin Texaco Exploration Production Inc. Bob Simon Chevron USA Production Company ABSTRACT Density, neutron and induction logs have traditionally been used to evaluate sandstone sediments of the Gulf of Mexico. The combination of these measurements has proved cost effective for determining formation lithology, porosity, and saturation. These reservoir properties are then used for predicting the producibility of the interval. In many zones, accurate producibility prediction is difficult. Zones with high irreducible water caused by very fine grained rock texture or thin beds can make producibility predictions pessimistic. In other cases the wellbore environment creates difficulties. Oil-base mud, in particular, can make interpretation difficult by invading deeply into the formation and masking traditional log response. Nuclear magnetic resonance (NMR) techniques have long been available to supplement traditional measurements. The NMR measurement responds to the formation’s porosity and pore size and thus provides additional information on the producibility of the reservoir. However, the main limitation of NMR technology has been the time and cost of acquiring the data. Although the need for mud doping has been eliminated, the relatively long cycle of the NMR measurement still results in very slow logging speeds. This paper demonstrates cost-effective applications of NMR logging to Gulf Coast land and offshore petrophysical analysis. Case studies are presented that utilize NMR in combination with traditional logging measurements. Situations include cases where NMR data were acquired simultaneously with density-neutron-induction data to provide a complete analysis in real time. Additional cases demonstrate how wireline NMR measurements can be combined with measurements acquired while drilling to provide a cost-effective evaluation in highly deviated wellbores. These examples, complete with production results, illustrate the application of NMR interpretation in wells drilled with fresh water mud systems and oil-base drilling fluids, and in reservoirs ranging from high porosity clean sandstones to formations with lowresistivity pay. In conclusion, NMR measurements combined with traditional measurements can be used in a cost-effective manner to provide a more complete petrophysical analysis of Gulf Coast reservoirs. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER QQ A NEW CHARACTERIZATION OF BULK-VOLUME IRREDUCIBLE USING MAGNETIC RESONANCE G. R. Coates, D. Marschall, D. Mardon NUMAR, Houston, Texas J. Galford NUMAR Australia, Perth, Western Australia ABSTRACT Irreducible water volume from the new Magnetic Resonance (MR) logging tools provides the log analyst with insight into a formation’s permeability and its water-cut potential. However, the traditional T2 cutoff method to determine the bulk volume of irreducible water (BVI), currently in wide use, has been found to be inadequate for some formations and fluid conditions. A new method to characterize bulk volume irreducible that addresses these issues is presented in this paper. The method is based on the premise that each pore size has its own inherent irreducible water saturation. Given that relaxation time is related to pore size, this method utilizes core MR measurements to relate each relaxation time to a specific fraction of capillary bound water. Thus, the bulk volume irreducible becomes a direct output of the inversion of the echo data, and it utilizes the entire T2 distribution. Core data are presented that demonstrate the Spectral Bulk Volume Irreducible (SBVI) petrophysical model and the method used for its characterization. Log examples of the SBVI implementation are presented to demonstrate the improvements brought by this development. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER RR EFFECTS OF TOOL DESIGN AND LOGGING SPEED ON T2 NMR LOG DATA Carl M. Edwards Western Atlas Logging Services ABSTRACT Advanced NMR logging applications depend critically on capturing and resolving the entire T2 spectrum. Much of the critical T2 information is contained in the long T2 components beyond 256 ms when hydrocarbons or large pores are present. Unfortunately, an NMR logging tool’s ability to faithfully capture long T2 data depends on tool design and logging speed, as well as rock and fluid properties. Logging speed is necessarily a compromise between often conflicting requirements for signal-to-noise, depth resolution, NMR experiment time and operational constraints, the relationship of logging speed, vertical aperture, vertical averaging and sensed volume is well understood. Generally, a low logging speed is required for enhanced vertical resolution and improved signal-to-noise. However, low logging speeds can significantly increase logging time (and costs) and can increase the risk of tool sticking. What is often not realized is that the CPMG echo data collected with a moving tool may be distorted, depending on tool design and logging speed. Previous modeling work by Akkurt focused on one particular tool geometry. We generalize Akkurt’s work to include other tool geometries and the effects of tool geometry and logging speed on T2 data. Apart from the depth resolution of the measurement, tool motion and tool design affect data quality and potentially limit interpretation possibilities. For a moving tool, the formation excited by the initial pulse of the CPMG sequence is not the same volume from which the last echo originates. This skew in data collection can be modeled as an additional, undesirable, relaxation term. Incomplete polarization can be important when light oils and gases are present in the formation, because the tool length is finite. Finally, the vertical profile of the static magnetic field when combined with the tool motion can introduce a speed-dependent phase angle into the echo data. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER SS IMPACT OF NMR IN THE DEVELOPMENT OF A DEEPWATER TUIRBIDITE FIELD Ridvan Akkurt NUMAR M. Andrei Moore Shell Offshore Inc. J. Justin Freeman Shell E &P Tech. Co. ABSTRACT The Tahoe field is a deepwater turbidite field in the Viosca Knoll area of the Gulf of Mexico. The main pay horizon is a highly laminated formation of middle Miocene age. Following a year of initial production, a second phase of development was undertaken to fully develop the field by drilling several horizontal wells. One of the challenges in the second phase was the typing of hydrocarbons and identification of fluid contacts for proper horizontal well placement. Hydrocarbon typing has been problematic at Tahoe due to an abundance of shale laminae. Conventional logs such as neutron-density fail to detect gas due to their inability to resolve sand beds. Pressure gradient analysis can be used for gas detection in this field, however, many operational problems exist with this approach. Given the poor history of the above techniques, NMR was employed for fluid typing and contact definition. Using Time Domain analysis of data acquired with dual-wait times, NMR was used to successfully type the hydrocarbons in all three wells of the program. Relying on NMR as the primary openhole logging tool resulted in simplified operations and decreased costs. Resistivity based water saturation models fail at Tahoe due to the highly laminated nature of the formations. Capillary pressure models have been used for reserve estimates given that certain assumptions are satisfied. NMR was considered as an alternative to this technique. The free and bound fluid information from NMR combined with structural information to determine whether the reservoirs were at irreducible, was used to estimate movable hydrocarbons. The most likely uncertainty (one sigma) in total net feet of hydrocarbon from the NMR approach was 14%, compared to 20% for the capillary pressure model. The results obtained at Tahoe show that NMR is a powerful petrophysical tool in turbidite formations since it provides unique formation evaluation and reservoir engineering information. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER TT LOG WATER SATURATION MODEL VALIDATION USING NMR LOG AND CORE DATA J. D. Klein and P. R. Martin ARCO Exploration and Production Technology, Piano, Texas Wisnu Widjanarko Atlantic Richfield Indonesia, Inc., Jakarta, Indonesia ABSTRACT A low invasion core was obtained in the BM-2 well, B-Field, offshore Northwest Java, Indonesia, for the purpose of validating wireline log porosity and water saturation models. The reservoir sands are lower Miocene in age and are generally clay rich and poorly consolidated. The first step in validating the log models was to compare log and low invasion core porosity data. The average core porosity in the reservoir sands was nearly 4 PU higher than the average porosity obtained with the log model. When examined on the basis of bulk density the core data were low compared to the measured log by nearly 0.1 gm/cc, which is 5 to 10 times larger than the accepted accuracy of the density tool. We concluded that the core, which had high water content, had likely been irreversibly changed when preserved by freezing, so we took an alternate approach for model validation using NMR logs in the adjacent BM-3 welL NMR measurements were carried out on core plugs from the BM-2 well. The NMR data were obtained with the core fully saturated with brine and after desaturating on centrifuge with air at approximately 20 psi. The two data sets were then compared to obtain a T2-cutoff of less than 10 msec. This relatively fast T2 value may be due to the presence of iron-bearing clays. The BM-3 well was drilled nearby and logged with the CMR tool as well as a conventional suite of logs. No core was obtained in this well. The NMR logs from the BM-3 well were initially processed using a 33 msec T2 cutoff, Based on the BM-2 NMR core data, the logs were reprocessed using T2 cutoffs of 10 and 20 msec to obtain revised values of irreducible water saturation. The conventional log model porosity and total NMR porosity agreed on average to within 1 porosity unit. Irreducible water saturation from NMR logs and a WaxmanSmits water saturation agreed qualitatively, but important differences were noted when compared on a more detailed basis. These differences could be understood if explained in terms of variation in T2-cutoff with capillary pressure, and when accounting for separate hydrocarbon columns present in the B-Field. Overall the NMR logs supported the accuracy of the conventional log models. Our results suggest the possibility of using NMR data, under favorable conditions, as another means of calibrating conventional water saturation models. This might be a viable alternative when undisturbed low invasion cores are difficult to recover, for example in unconsolidated sands or sands with high water saturation. A complication in the use of NMR data is the appropriate pressure to use in the core desaturation step SPWLA 38th Annual Logging Symposium, June 15-18, 1997 during determination of the T2 cutoff. Since capillary pressure varies with height in a reservoir, the T2 cutoff should also be determined as a function of height, which is not normally done. For the reservoir rocks in the B-Field, with a wide range in pore types and sizes, variation in fl-cutoff with capillary pressure must be accounted for when quantitatively interpreting NMR logs for hydrocarbon pore volume. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER UU INTEGRATION OF NUCLEAR MAGNETIC RESONANCE CORE ANALYSIS AND NUCLEAR MAGNETIC RESONANCE LOGS: AN EXAMPLE FROM THE NORTH SEA, UK. Tom Walsgrove Amoco (UK) Exploration Company, London, UK Simon G. Stromberg, Ben D. Lowden and Paul B. Basan Applied Reservoir Technology, Suffolk, UK ABSTRACT We demonstrate that log interrogation and facies analysis can be used to integrate core NMR response models into the log. T2 shift and T2 rate-of-change analysis interrogates the fundamental parameters that control the NMR log response. Because of differences in borehole environment, logging speed, fluid saturation and other tool dependent factors, NMR response models developed from core analysis are not directly comparable to the log. Consequently, we have developed NMR facies techniques that compensate for the inability to make direct comparisons. NMR facies are defined as ‘a set of similar NMR distributions that summarise the characteristics of the rock’. NMR facies analysis techniques are used to group similar T2 distributions and relate NMR response models obtained from core analysis to the log. These methods are tested on an NMR log from the North Sea. Core-to-log integration subdivides the log into units that have petrophysical significance. These subdivisions assist in finding variations in petrophysical properties, and thus, define flow units in the reservoir. NMR facies analysis allows NMR petrophysical characteristics to be linked to other geological criteria that are predictable away from the well bore. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER VV CONFIRMING RW BY RECONSTRUCTING THE S.P. USING FORMATION EVALUATION RESULTS EMPLOYING GRADIENT FIELD MAGNETIC RESONANCE LOGS G.R. Coates and D. Miller NUMAR ABSTRACT The new NMR measurements; mineralogy free porosity, direct hydrocarbon volume and type and its tie to CEO, have brought a new level of quality to log derived water saturation both in the flushed and virgin conditions. However, the inversion of resistivity logs into water filled porosity remains heavily dependent upon the resistivity ( Rw ) of the formation water, and having no direct measure of Rw the analyst, on many occasions is left unsure of the Rw value to use. Such an event led to a search for methods which could be employed to quality control the Rw selection, a method that was found through use of NMR and the SP. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER WW CONCILIATION OF CORE DATA AND LOG DERIVED PERMEABILITY FROM NUCLEAR MAGNETIC RESONANCE AND STONELEY WAVE IN NORTHERN MONAGAS AREA, EASTERN VENEZUELA M. J. Prieto and R. Barbato Corpoven S.A. A. Sinha and G. Gomez Western Atlas Logging Services, Venezuela ABSTRACT In the Northern Monagas Field of Eastern Venezuela, the main producing formation is called Naricual which has a thickness up to 1500 ft., associated with varying grain sizes and different clay types. The productivity of this sand is controlled by variation of permeability in the sand body. Therefore a reliable estimation of the permeability, to identify the zone for completion, is a big challenge for the petrophysicist, specially in absence of a continuous permeability measurement tool in the wireline industry, In the past, several attempts have been made to derive a reliable permeability of the reservoir based on wireline logs. Recently permeability calculations based on irreducible water saturation from Nuclear Magnetic Resonance Logs is gaining more and more acceptance in the industry. Similarly several investigations indicate that the permeability in rocks are correlatable with Stoneley wave attenuation and a permeability index curve can be derived from analyzing the attenuation of the Stoneley wave. Permeabilities derived from Nuclear Magnetic Resonance, attenuation of the Stoneley Wave and core analysis are studied in depth and are compared with the actual production data. An attempt is made to understand the differences in permeability obtained from different sources. Based on these studies, a model is built to predict a reliable permeability to reflect the productivity in future wells. This model already tested on several wells will be used in the field to decide the completion zone. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER XX PERMEABILITY FROM NUCLEAR MAGNETIC RESONANCE LOGGING IN A GASCONDENSATE FIELD J.J. Howard and J.S. Williams Phillips Petroleum Co., Bartlesville, OK D.C. Thorpe Phillips Petroleum Co (ZOC), West Perth, WA, Australia ABSTRACT Nuclear Magnetic Resonance (NMR) technology was applied to assist in the reservoir characterization of the gas-bearing sandstones of the Bayu-Undan Field in the offshore of N. Australia Zone of Cooperation. NMR log permeabilities based on the standard algorithms underestimate core permeabilities by as much as two orders of magnitude in the high porosity, low clay content sandstone intervals. These initial log permeabilities also overestimate permeability in the lower porosity, interlaminated shaly sandstone intervals. Improved NMR log permeability values were obtained by adjusting inputs to the porosity and pore size terms in the permeability algorithms. The presence of gas with its low hydrogen index in the logging tools’ sensitive region reduces the NMR porosity and alters the relaxation time distribution. The porosity loss can be corrected by replacing the NMR porosity with a Density-Neutron crossplot porosity value. This correction improves the permeability estimates in the clean sandstone intervals to almost a 1:1 correlation with core results, but still overestimates in the low permeability intervals, Adjustments to the pore size term allow for a better match in the low permeability range. Mean relaxation times from the logs are not useful here because they show considerable scatter when compared to the laboratory measured relaxation times measured a fully- and partially-saturated conditions. The free fluid/bound water volume value does work when several modifications are made. The very clean nature of the sandstones requires that the threshold time be shifted to 80 milliseconds from 33 milliseconds. Contributions from the larger pores also are added to the estimate of bound water volumes based on a statistical analysis of relaxation times as a function of partial saturations for these Bayu sandstones. The comparison between core permeability values and the log-estimated permeabilities over a four-order magnitude range produces a correlation coefficient of greater than 0.75 in a single well. Vertical resolution limits of the logging tools affect the comparison of individual layers, but the agreement in permeability trends in the various sandstone layers is very good. Net pay calculations using the NMR permeability curve as cutoffs agree favorable with estimates based on permeability thresholds derived from porosity and shale volume correlations. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER YY FAST NMR LOGGING FOR BOUND FLUID AND PERMEABILITY J.M. Singer and L. Johnston Schlumberger Wireline & Testing R.L. Kleinberg and C. Flaum Schlumberger-Doll Research Center ABSTRACT Nuclear magnetic resonance (NMR) logging provides valuable and unique information for formation evaluation. One drawback has been the long time it takes to acquire the data, as much as 10 times longer than traditional neutron-density logs. To overcome this, a specific CMR Combinable Magnetic Resonance Tool acquisition sequence has been designed to provide the unique NMR measurement of bound fluid, and to predict permeability in combination with other measurements. This sequence can be run many times faster than conventional NMR logs, at logging speeds that are typical for neutron-density logging. It can be run most conveniently in combination with these logs. Reliable bound fluid and permeability data can therefore be obtained with fast NMR logging. A full NMR measurement requires a long wait time to polarize all components of the formation, and a long acquisition time to measure the longest relaxation times. However, experience has shown that the T2 relaxation time of the bound fluid is in most cases less than 33 msec in a sandstone and 100 msec in a carbonate. In fast NMR logging, it is possible to use short wait times by accepting less accuracy in measuring the longer components. In addition, a short echo spacing and an appropriate number of echoes reduce the acquisition time and ensure that the measurement volume does not change significantly because of the faster tool movement. Full NMR and fast NMR logs have been recorded together in several wells, with good-quality bound fluid logs obtained from the fast pass. Statistical comparisons show that the accuracy, precision and vertical resolution of the bound fluid volume from the fast run is close to that of the full run. As in full NMR logging, the bound water volume can be further analyzed for the content of clay bound water and irreducible water. With full NMR logging the permeability is calculated from the NMR porosity and either the mean T2 or the free fluid volume. With fast NMR logging, porosity is taken from another log, typically the density log in shaly sands, or the density-neutron crossplot in gas and carbonates. Free fluid is taken from the difference of porosity and bound water. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER ZZ ELASTIC AND ELECTRICAL PROPERTIES OF COARSE SAND UNDER THREE-PHASE FLUID SATURATION G. Yu Western Atlas Logging Services, Houston, Texas L. Leunt BHP Research, Newcastle Laboratories, Newcastle, N.S.W., Australia K. Vozoff HarbourDom Pty. Limited, Sydney, N.S.W., Australia ABSTRACT In recent years, geophysical techniques have increasingly been used to characterize oil reservoirs, oil contaminated sites, and to evaluate remediation progress or success. However, little information has been available on the elastic and electrical property values needed to define the relationships between subsurface bulk-rock geophysical signals and the extent and nature of oil and various oil-related contaminants. An experimental study aimed at determining these properties under three-phase fluid saturation in a laboratory condition has been conducted at Macquarie University, Sydney, Australia. The experimental results showed the elastic and electrical property changes with different fluid saturation that can be used to interpret the geophysical survey data. Laboratory studies are described for elastic and electrical property measurements of coarse sands (density sands) saturated with mixtures of air, deaerated water, diesel oil, marine bunker oil, and tar. Elastic properties include P-wave velocity, peak transmission frequency, and attenuation; electrical properties include resistivity, phase response, and relative dielectric constant, in the frequency range of 1 Hz to 32 MHz. All measurements were conducted at room temperature under a 1.0 MPa confining pressure. Ultrasonic compressional wave velocities in coarse sand specimens were strongly affected by water and heavy hydrocarbon saturation. Significant increases occurred in amplitude spectra and peak frequency of P-waves with water saturation. Amplitude and peak frequency decreased with increasing diesel oil and decreasing water saturation. The peak frequency in fully and partially heavy hydrocarbonsaturated specimens was lower than that in fully water-saturated specimens. Electrical property changes in coarse sand were found to depend strongly on water saturation, As the water saturation increased and the diesel oil or heavy hydrocarbon saturation decreased, the resistivity values decreased and the dielectric constant increased in the sand specimens. The experimental results of this project demonstrated that both the elastic and electrical properties are different between air-dried, partially water- and oil-saturated, and fully water-saturated, unconsolidated materials. The experimental results should be useful indicators of the states of water and oil saturation in unconsolidated SPWLA 38th Annual Logging Symposium, June 15-18, 1997 materials, and provide valuable information for interpreting geophysical surveys conducted to characterize oil reservoirs as well as oil-contaminated sites. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER AAA SPATIAL DISTRIBUTION OF PETROPHYSICAL PARAMETERS ON A CORE SCALE USING MAGNETIC RESONANCE IMAGING A. Brancolini, A. Cominelli AGIP S.p.A., Milan, Italy R. Kulkarni and A.T. Watson Texas A&M University, Department of Chemical Engineering, College Station, TX ABSTRACT Spatial heterogeneity at the core sample scale is studied using nuclear magnetic resonance imaging (MRI) techniques. MRI is used to obtain two dimensional porosity maps on several core samples. Variogram analysis is used to better describe and compare the heterogeneity information from porosity maps. Permeability, maps are estimated using known correlations that relate absolute permeability to porosity and NMR relaxation parameters. The absolute permeability maps are compared against maps provided by minipermeameter. A key step in quantification of MRI data is the estimation of the intrinsic magnetisation intensity on a pixel-by-pixel basis. We use an inversion process based on continuous distribution of relaxation rates. In this technique the parameter estimation problem at each pixel is linear, which offers considerable computational advantage and convenience over other numerical implementations that frequently require use of non-linear parameter estimation methods. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER BBB NEW MWD RESISTIVITY TEST FORMATION FOR 2D AND ECCENTRIC TOOL EXPERIMENTS Mac M. Wisler. Jian-Qun Wu, John A. Signorelli, Larry W. Thompson, and Thomas Bell Baker Hughes INTEQ ABSTRACT Response characterization of formation evaluation tools should be solidly based both on theoretical analysis and on laboratory measurements. Recent years have seen extensive modeling work on the theoretical response of wave propagation resistivity tools. but due to difficulty in constructing laboratory formations, experimental work has lagged behind. In this paper we describe a new test formation which for the first time simulates a set of actual downhole conditions for resistivity tools. The formation consists of a tank containing a sequence of five beds of varying porosities and a permeable borehole, The construction of the tank is described and propagation resistivity measurements are made in the test formation and compared with model calculations at both 2 MHz and 400 KHz at two spacings using the Baker Hughes INTEQ MPR tool. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER CCC ANALOG MODELING FOR ACOUSTIC IMAGE LOGGING TOOL G. Yu, M. Civarolo, S. Painchaud, K-M. Strack, and D. Tetzlaff Western Atlas Logging Services, Houston, TX ABSTRACT Borehole imaging technology is one of the most rapidly advancing fields in the geotechnical industry. Some of the most important applications in downhole imaging are advanced reservoir engineering, enhanced hydrocarbon recovery, stimulation of tight reservoirs, and geomechanical studies of deep exploration boreholes. Although the two most common imaging measurements, micro-electrical and acoustic, appear to be independent, Western Atlas Logging Services has combined them into one subsurface device. In order to quantitatively interpret the image logging data as well as validate numerical modeling results, a rock borehole model has been built to study the responses of the image logging tools to known geological features. This paper presents the results of analog modeling for the acoustic image logging tool. An analog modeling facility has been built to conduct analog modeling for both the resistivity imaging and the acoustic imaging tools. The equipment is designed to measure the sensor response to small and/or complicated features such as vugs, fractures, tilted beds, thin beds and mudcake. These complicated features are often difficult to model with current numerical modeling capabilities. A rotating platform from the acoustic section of the imaging tool was used for the acoustic image modeling. The analog modeling results characterize tool responses to geologic features, and are being used to develop quantitative interpretation software for the image logging data. They also serve to independently validate numerical modeling. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER DDD A COMPARISON OF WIRELINE AND LWD RESISTIVITY IMAGES IN THE GULF OF MEXICO Jeff Prilliman Schlumberger Wireline & Testing Clarke L. Bean Chevron USA Moharned Hashem Shell Offshore, Inc. Tom Bratton, Mark A. Fredette, and John R. Lovell Anadrill ABSTRACT Borehole resistivity images provide a wealth of information for petroleum geologists and petrophysicists. Such images are generated by measuring formation resistivity azimuthally around the borehole and assigning a graduated color scale to the range cf resistivities encountered. The borehole images are oriented by magnetometers and can be plotted with reference to the top of the hole or north. Traditionally, such imaging measurements could be made only using wireline-conveyed tools. Recent technological advances allow quantitative resistivity images to be generated from measurements taken while drilling. Although these images do not have the vertical resolution available from some wireline imaging tools, the while-drilling images have many uses, including picking structural dips, identification of some fractures, lithology indication, and thin-bed analysis. Inherent benefits of the while-drilling measurement include 100% borehole coverage and better hole conditions than at wireline time. Data are recorded versus time in the while-drilling mode which facilitates time-lapse logging to show features such as invasion and borehole stability. Some limitations of the current while-drilling image acquisition technology include artifacts resulting from time/depth data recording discrepancies and a requirement that the drillstring be rotating above 30 rpm. A well recently drilled in the Gulf of Mexico was logged with both wireline and logging-while-drilling (LWD) tools. Raw data from these measurements was used to generate resistivity images. The advantages and disadvantages of each type of measurement are investigated. In addition, the limiting factors and complications in the processing of each type of data are discussed. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER EEE HYDROCARBON IDENTIFICATION FROM SHEAR SONIC LOGS IN THE NEUQUEN BASIN, ARGENTINA Mario Schiuma Chauvco Resources, Buenos Aires, Argentina Michel Claverie Schlumberger W&T, Buenos Aires, Argentina Alain Brie Schlumberger K.K., Wireline & Testing, Fuchinobe, Japan ABSTRACT Sonic slownesses are sensitive to the type of formation pore fluid. The sensitivity is large in the case of unconsolidated sands and when gas is present in the formation. The influence of fluid type is reduced in more compact formations. Up to now the use of sonic slownesses for fluid type identification was recommended only in sands with a porosity of at least 20% and to differentiate between liquid and gas. The present study reports how sonic slownesses were successfully used to identify hydrocarbons and to distinguish between oil- and gas-bearing intervals in wellcompacted sands in the Neuquen basin in Argentina. The reservoir environment is fluvio-deltaic with the fluid distribution controlled not only by the structure, but also by stratigraphy, and it is common to find anomalous fluid distributions, with oil zones overlaying gas zones. The shallow, low porosity sands produce under various degrees of depletion. In these conditions the traditional neutron-density technique often has difficulties in distinguishing between oil and gas, probably because of intense flushing of the hydrocarbons in the invaded zone. A crossplot technique of Vp/Vs versus compressional ∆t is successfully used to identify the hydrocarbon type. This is done at the wellsite and is available during logging to decide on further testing and completion programs. A more detailed evaluation is made on a workstation system at the office. The technique has been used with good success both in open and cased hole conditions. The predictions have been confirmed by production tests and by correlation on more than 20 wells, Examples of evaluations in different wells are presented and compared with other openhole logs. These results show that the use of sonic slownesses for hydrocarbon identification can be extended in some instances to the case of compact, depleted sandstones. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER FFF PETROPHYSICAL PROPERTIES AND GEOLOGY OF SELECTED INTERVALS IN THE FRIO FORMATION, STRATTON FIELD FOR MODELING INTERWELL SEISMIC LOGGING RESPONSES Jorge O. Parra Southwest Research Institute, San Antonio, Texas Hughbert A. Collier Tarleton State University, Stephenville, Texas ABSTRACT Surface seismic (2D and 3D) may not detect and resolve deep thin beds at Gulf Coast gas fields in the Frio Formation. Interwell seismic measurements between wells may enhance resolution in on shore gas fields to map in detail the deep subsurface geology. Since gas well separations in the Texas Gulf Coast exceed 1/2-mile, multiple offset measurement techniques such as traveltime tomography and reflection imaging are currently not economically justifiable for mapping the subsurface between wells in Gulf Coast gas reservoir formations. An alternative approach that can be complementary to surface seismic is continuity logging using guided waves. This method can be used to measure and analyze inter-well seismic data to assess the continuity and discontinuity of target zones of interest and delineate the interwell geology. A model study was conducted to demonstrate the feasibility of transmitting and detecting guided seismic waves in Gulf Coast gas reservoirs. Geological cross-sections between selected wells in the upper and middle Frio Formation were compiled using well logs and core information. The upper Frio Formation consists of laterally continuous sandstones interbedded with mudstones. Both the sandstones and mudstones are continuous for distances exceeding one mile, making an ideal environment to conduct experiments for proof of concept of transmission and detection of guided waves between gas wells. The middle Frio contains laterally discontinuous sandstones and is a challenge for continuity logging. A petrophysical analysis of selected wells was undertaken to provide the rock physical parameters (i.e., velocities, density and thicknesses) to simulate guided waves, The P-wave and S-wave velocities of each bed in the zone of interest was classified as either low velocity (consistent, uniform velocity mudstone), intermediate velocity (heterogeneous, impure mudstone), high velocity (generally the sands), and very high velocity (generally lower porosity sands). The cross-section containing velocity zones of interest were modeled using P-wave velocity and S-wave velocity parameters. The computer models predict the waveform characteristics associated with some of the target zones in the upper Frio Formation for planning continuity logging experiments. The model includes the source strength level and noise levels for given velocity contrasts, Q~ and Q quality factors, and well separation. The computer model results determine that three-component geophone arrays may be appropriate to detect guided seismic waves at a well separation of 800 m using a source having a volume expansion of 10 cm3 in a noisy environment of 100 mPa. In addition, modeling results demonstrate that both leaky and normal modes are the most probable SPWLA 38th Annual Logging Symposium, June 15-18, 1997 events to be excited in shale waveguides if continuity logging experiments are conducted in the upper Frio Formation. SPWLA 38th Annual Logging Symposium, June 15-18, 1997 PAPER GGG COMPLEX RESISTIVITY MEASUREMENT IN FREQUENCY / TIME DOMAIN AND ESTIMATION OF FORMATION WATER RESISTIVITY Peng Shilin, Hu Xinmiao, Wang Jingnong, Sun Baodian, and Zeng Huaxiu Jianghan Well Logging Institute, CNPC ABSTRACT: Rock complex resistivity measurement and analysis is a new method applied to the study of rock electrical property. The research may be conducted in frequency and time domain. A set of equipment “Rock Complex Resistivity Measurement and Analysis System (CRMAS)” has been developed, which measures the resistance and impedance of rock samples at simulated reservoir temperature and pressure (max. temperature 150 °C and max. pressure 70 MPa). The scope of frequency measurement is from l0-3 Hz to 100 MHz. A four-electrode coreholder is used. The samples of different lithology taken from several oil-fields are measured and analyzed. Our experiments show that there is a frequency dispersion in all different lithology samples. When the measuring frequency is less than 100Hz or greater than 1MHz, the frequency dispersion phenomenon is more obvious. Temperature and pressure also affect the amplitude of the resistance measured. These results can be applied not only to develop new well logging tools for formation rock complex resistivity measurement, but also to aid to the research of Array Logging Tools working in multi-frequencies. Rock complex resistivity measurement is also made in time domain, that is to measure decay characteristics of rock polarization potential and polarizability. The relationships between rock polarizability and cation exchange capacity (Qv) and the salinity of formation water are established. Based on laboratory work, we have developed the induced-polarization well logging tool which shows a good geological results in the determination of reservoir formation CEC, formation water salinity and the waterflooded performance of the pay zones. The paper presents methods and instruments of rock complex resistivity measurement in frequency /time domain. A mathematical equation is drawn to represent relations of rock complex resistivity with formation water resistivity, shale content or cation exchange capacity (Qv). The field examples are given to verify its geological results. TRANSACTIONS \ OFTHE SPWZA THIRTY-NINTH ANNUALLOGGINGSYMPOSIUM Sponsoredby THE SOCIETY OF PROFESSIONALWELL LOG ANALYSTS, INC. 8866 Gulf Freeway,Suite 320 Houston, Texas 77017 Presentedat THE KEYSTONE CONFERENCECENTER Keystone,Colorado May 26-29, 1998 NOTICE TO EDITORS: Permission is hereby granted to publish elsewhereany of these transactions tier May 29, 1998,provided that conspicuousacknowledgementis given to the original presentationof the paper and the authors of the paper have agreedto the republication. (The statementsand opinions expressedin thesetransactionsare those of the authors and should not be construedas an official action or opinion of the Societyof ProfessionalWell Log Analysts, Inc.) SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER A INTERPRETATION OF MULTIARRAY INDUCTION LOGS IN INVADED FORMATIONS AT HIGH RELATIVE DIP ANGLES Thomas D. Barber, Tracy Broussard, Gerald N. Minerbo, Zlatko Sijercic Schlumberger Oilfield Services David Murgatroyd Rice University ABSTRACT A new processing algorithm for multiarray induction tools has been tailored for highly deviated wells. This algorithm provides the same interpretation for invasion that has been available in the past only in vertical wells. Induction arrays are designed to be interpreted in vertical wells. All focusing techniques, whether traditional multicoil focusing or modern software focusing, have as their goal the isolation of the response of the tool into a reasonably thin layer when the borehole is perpendicular to the bedding planes. The new processing algorithm isolates the tool response into thin layers at any relative dip angle. Previous dip correction algorithms for induction logs were limited in practice to angles less than 50° because of the increasingly nonlinear response to dip. As a result, the interpretation of resistivity from induction logs at apparent dip angles over 500 has been limited to iterative forward modeling using ID forward models. Even with fast codes, this procedure is time consuming in long logs or where thin beds are of concern. The presence of invasion has until now made resistivity interpretation at high dip angle possible only with 3D codes. The new algorithm is based on maximum-entropy inversion of the raw borehole-corrected array data through a fast 1 D forward model. It allows interpretation of multiarray induction logs even in the presence of invasion. Tests on a wide range of invasion profiles computed with a 3D induction forward model code at relative dip angles as high as 85° show that the new inversion allows determination of R~0, R., and the invasion profile with similar accuracy to that determined in vertical wells. In theory the process works to 90°; however, the current parameterization requires that the wellbore cut all beds of interest. Studies of sensitivity to incorrect dip angle show that in most cases the relative dip must be known to ±5° to satisfy reasonable petrophysical requirements. Sensitivity studies to other sources of error, including coherent borehole noise, show that the new process has a sensitivity to these effects similar to the current field processing. Because the nonlinear response of induction arrays to conductivity is handled explicitly, and without approximation, the new process also handles large shoulder-bed dip angles. Application to field logs at a variety of relative dip angles confirms that the results predicted from modeled data transfer to the real world, producing R, estimates that are fully corrected for dip effect. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER B THE EFFECT OF CROSSBEDDING ANISOTROPY ON INDUCTION TOOL RESPONSE B. I. Anderson, T. D. Barber Schlumberger S. C. Gianzero Sensor Modeling, Inc. ABSTRACT Many hydrocarbon reservoirs consist of crossbedded deposits. Crossbedding can result from water deposition or aeolian deposition. Grain size sorting and differential cementation can give rise to resistivity anisotropy that is not directed along the bedding planes. Previous studies of resistivity anisotropy have assessed the sensitivity of induction tools to anisotropy in transversely isotropic media aligned parallel to the bedding planes. In this paper, we address full three-dimensional anisotropy having arbitrary dip and strike relative to the beds. Crossbedding effects can now be studied for the first time. A computer code was developed to model the response of an induction logging tool traversing many beds, each of which possesses different crossbedding characteristics. The crossbedding in each medium is described by means of a uniaxial conductivity tensor whose principal axes have strike and dip angles oriented arbitrarily with respect to the bedding planes. The code is numerically efficient; response for a tool logging through several beds can be generated in less than 15 minutes on a modem workstation. The overall results show that, for anisotropy coefficients less than 5, computed responses for both two-coil and multicoil devices vary in a continuous manner as the sondes cross a single bed boundary separating two infinitely thick beds. Furthermore, after correction for skin effect, the limiting log values far from the bed boundary are entirely predictable from a previously published formula. However, in vertical wells, when the crossbedding dip angle is 75° or greater and the anisotropy coefficient greater than or equal to 5, anomalously large readings appear in the vicinity of the bed boundaries. These large readings are similar to the polarization horns that occur in dipping beds at high-contrast isotropic interfaces. In the case of a thin bed (e.g., 5 ft) located between two massive shoulder beds, the large anomalies from the bed boundaries merge into a single anomaly at the center of the bed. This behavior can be quantified only by modeling. Modeled results are also used to analyze AIT Array Induction Imager tool response in a crossbedded reservoir in the Nugget formation where we expect different values of Rv and Rh. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER C FIELD MEASUREMENTS AND INVERSION RESULTS OF THE HIGH-DEFINITION LATERAL LOG R.G. Hakvoort Shell International Exploration and Production, The Netherlands A. Fabris, M.A. Frenkel Western Atlas Logging Services, Houston, Texas, U.S.A J.M.V.A. Koelman Shell Petroleum Development Oman AM. Loermans Nederlandse Aardolie Maatschappij, The Netherlands. ABSTRACT The High-Definition Lateral Log (HDLL) tool is a type of a Multi-Electrode Resistivity Tool (MERT). It is a new tool that has been developed in a joint project between WALS (Western Atlas Logging Services) and Shell. The tool was developed to address shortcomings with conventional dual laterolog (DLL) technology (restricted vertical resolution, artefacts in deviated boreholes), The new tool has a single current-injection electrode and 18 potential measurement electrodes at various distances from the injection electrode. The tool measures absolute potentials and first differences (potential difference between two neighbouring electrodes). The raw responses are combined in such a way as to produce a shallow-, a medium-, and a deep-reading synthetically focussed curve at the well-site. In addition, inversion is given as a postprocessing HDLL service. The inversion method is based on a direct inversion of the raw measured data, both in 2D (layered formation, including effects from borehole and invaded zone) and in 3D (layered, dipping formation, including effects from deviating borehole and invaded zone). In this paper, two case studies are discussed. The first one is a vertical well in a formation that includes many thin layers. The second is a highly deviated well. For both cases, the wellsite deliverables and the inversion results are shown. A comparison with DLL measurements shows clear advantages of the HDLL compared to the conventional DLL technology (better vertical resolution, fewer artefacts in deviated boreholes). SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER D INTERPRETATION OF PROPAGATION RESISTIVITY LOGS IN HIGH ANGLE WELLS W. Hal Meyer Baker Hughes INTEQ ABSTRACT The interpretation of propagation resistivity logs at very high relative dip angles is much more challenging than the interpretation of vertical logs. Several effects degrade the usefulness of propagation resistivity data when the borehole is horizontal or the relative dip angle is very high (over 70°). These effects include anisotropy, non-circular invasion, and eccentricity in oddly shaped boreholes. Some of these effects cannot be properly quantified so their effect on the data cannot be fully removed. Other effects can be removed, but the remaining presence of these unquantified effects makes this more difficult. An entirely new strategy is necessary to interpret horizontal and high angle wells, and even then the results will not be as accurate as they are in low angle wells. The first problem with interpretation of high angle wells is the inability to remove the effects of differing vertical resolution (more properly called axial resolution in this case). Lateral changes in the formation itself may have a larger impact on the data than the progress of the tool vertically through the formation (which is very slow at high relative dip angles). As a result, a normal inversion will increase the size of the changes within the formation more than it will reduce the effects of nearby bed boundaries. In addition, high angle wells often use the resistivity data to geosteer the well. Geosteering recognizes the nearby bed boundaries as variables and any attempt to eliminate these variables by inversion techniques is not desirable. Therefore, a strategy which treats the distance to the bed as a variable is required. This strategy must also reduce the effects of anisotropy, anomalous dielectric permittivity, and invasion. All of these effects have to be interpreted simultaneously because they all result in various types of separation of the apparent resistivity curves. If one of the parameters is analyzed individually, the resulting attempt to explain all of the separations with a single effect will cause an error in that parameter. In addition, it will then be impossible to determine the other parameters. In this paper several of these formation effects have been simultaneously inverted to produce “true” formation parameters. This method has been used to interpret some field logs. While the results demonstrate the advantage of this strategy over previous methods, the interpretation is still not as effective as interpretation in low dip angle formations. However, it is still possible to produce accurate resistivity curves at fixed depths of investigation. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER E DETERMINATION AND APPLICATION OF FORMATION ANISOTROPY USING MULTIPLE FREQUENCY, MULTIPLE SPACING PROPAGATION RESISTIVITY TOOL FROM A HORIZONTAL WELL, ONSHORE CALIFORNIA. D. MacCallum, M. Dautel Baker Hughes INTEQ Christopher Phillips Tidelands Oil Production Company ABSTRACT As part of a Department of Energy cost share program, a horizontal well was drilled in thin heterogeneous Miocene age turbidite sands. The challenge was to economically drill and exploit remaining reserves in the 60 year old Wilmington Field (Long Beach, California). The solution was to use new technology and sidetrack an existing wellbore with a horizontal lateral to capture hydrocarbon reserves uneconomically recoverable with historically used conventional methods. The new technologies included detailed reservoir characterization, 3D geologic modeling, geosteering in thin beds and modeling the Logging While Drilling (LWD) responses. Propagation resistivity measurements can be affected by eccentricity, invasion, variations in dielectric permittivity and thin beds. In situations of high relative dip, adjacent beds and formation anisotropy become significant factors in the log response. The use of a multiple spacing, multiple frequency propagation resistivity tool enables the calculation of multiple independent sets of vertical and horizontal resistivities. In addition to identifying and quantifying anisotropy, this also helps to determine additional borehole and formation effects. This case history demonstrates the application of forward modeling and inversion processing to enhance understanding of the horizontal log response and the reservoir structure of a complex horizontal well drilled onshore California. Current geosteering techniques frequently use offset wireline or LWD data from vertical or low angle wells. These logs predominantly measure the horizontal resistivity of the formation. Vertical resistivity cannot be accurately determined, if at all, in these situations. At high relative dip angles (e.g. in horizontal wells), a model generated from horizontal resistivity alone will not be representative of the actual log response. Horizontal and vertical resistivities derived from the inversion processing and subsequent modeling were in excellent agreement with both the offset wireline data and the actual LWD log. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER F ULTRASONIC VELOCITY AND ATTENUATION MEASUREMENTS IN HIGH DENSITY DRILLING MUDS Eric Motz, Duane Canny, and Eddie Evans Baker Hughes INTEQ, Houston, Texas ABSTRACT A parameter critical to the depth of investigation of an MWD acoustic (ultrasonic) standoff measurement is the ultrasound absorption in the circulating drilling mud. Previous work demonstrated that ultrasound absorption increases with mud weight to the heaviest (14.0 lb/gal [1.68 g/cc]) mud tested. Continued attenuation increase in the heavier muds would greatly limit the performance of acoustic standoff gauges. The attenuation mechanism is generally believed to involve the viscous relaxation within the mud. However, using typical values of drilling mud plastic viscosity in the Navier-Stokes equations yields attenuation coefficients orders of magnitude smaller than those observed. Some explanation is needed. Attenuation was measured in 9, 12, 15, 18, and 20 lb/gal water- and oil-based muds using pulse-echo amplitude detection. The front face of a highly-damped piezoelectric transducer is submerged in the test mud, facing a flat aluminum or sandstone reflector. After pulsing the transducer, the reflection amplitudes and arrival times are measured as the reflector distance is varied. Thus, the mud ultrasonic velocity and attenuation per unit distance can be calculated. The samples were tested at room temperature, without additives or flow. Two broadband transducers were tested, one that produced 280 KHz peak ultrasound and another that produced 180 KHz peak ultrasound. Rather than increasing with mud weight or plastic viscosity above 14 lb/gal as expected, the 280 KHz attenuation peaked in 15 lb/gal mud and decreased monotonically in the higherdensity muds. This was observed in both water and oil-based muds and is similar to the absorption peaks found in highly-viscous fluids where these peaks represent the cutoff between liquid-like and solid-like behavior. When 180 KHz ultrasound was tested, the attenuation peak position and shape showed frequency-dependence similar to those associated with ultrahigh viscous relaxation. The liquid-solid interaction in suspension mixtures like drilling mud may produce such a result and could explain why using the plastic viscosity to calculate the ultrasonic attenuation in drilling muds produces values much lower than measured attenuation. Comparative sound velocity data is presented. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER G APPLYING ANISOTROPY CORRECTIONS TO MWD ELECTROMAGNETIC WAVE RESISTIVITY DATA IN A SHALY-SAND RESERVOIR Sigit Sutiyono Unocal Indonesia Company Steve Mack Sperry-Sun ABSTRACT Large differences in MWD resistivity values were observed between horizontal wells and low-angle wells in a shaly-sand reservoir located offshore Indonesia. These differences made calculated water saturation (Sw) values too low in the horizontal wells by up to 30 saturation units. These Sw discrepancies, if not understood, could have led to erroneous net pay and reserve estimates. Tool modeling software was used to invert four phase-shift resistivity measurements to produce a horizontal resistivity (Rh) and a vertical resistivity (Rv), which helped in the evaluation of this reservoir. Rh was used for Rt in the calculation of Sw with satisfactory results, matching the values obtained in adjacent low-angle wells. Constructing a detailed geologic model of discreet anisotropic units, that could be verified using forward modeling code, was impossible given the complex nature of this reservoir. Modeling results from adjacent shales and analysis of offset core data suggest that the observed macroscopic anisotropy is due to fluid stratification and not shale lamination. The shale section situated stratigraphically higher than the reservoir does not exhibit any obvious anisotropy effects, regardless of the relative dip angle (RDA). The core data from adjacent wells show the permeability to vary significantly on a sample-by-sample basis, consistent with a reservoir that is made up of thin layers of differing Sw. The study area is the Yakin Field, which is located offshore East Kalimantan, Indonesia. The sandstone reservoir is an unconsolidated, tidally influenced, distributary channel with lithic fragments, feldspars, clays, and detrital coal. A horizontal-drilling project, implemented to drain oil from this reservoir, was planned and executed using MWD tools for formation evaluation. The MWD tool string provided real-time and recorded gamma ray, neutron porosity, and formation density measurements. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER H A NEW PRODUCTION LOGGING SERVICE FOR MULTIPHASE HORIZONTAL WELLS David Chace, Darryl Trcka, and Dan Georgi Western Atlas Logging Services Olaf Bousché and Alex van der Spek Shell International Exploration and Production Hassan al Nasser Petroleum Development Oman ABSTRACT New hardware and interpretation methods are introduced that are designed specifically for production logging and inflow profiling in multiphase horizontal or deviated wells. The new method is intended to allow the measurement of three phase flow rates in any flow regime, watercut, and gas volume fraction. The new production logging instrumentation utilises arrays of capacitance sensors to make across-the-borehole measurements of liquid level, hold up, and phase velocities while simultaneously making average borehole measurements of pressure, temperature, diameter and noise. Newly developed downhole flow modelling software helps pre-job planning and post-job evaluation of data consistency. flow loop data shows the resulting accuracy of the phase flow rate measurements to be approximately 20%. Log examples illustrate the measurement of multiphase inflow profiles. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER I GAS HOLDUP TOOL APPLICATIONS IN PRODUCTION LOGGING Gary Frisch, Margaret Waid, Calvin Kessler, and William Madigan Halliburton Energy Services ABSTRACT Horizontal and multilateral wells are being drilled as a cost-effective approach to increasing production and adding reservoir value. Conventional center-sample production logging tools have proven to be inadequate in horizontal and highly deviated wells. A new quick well analysis has been developed, along with a new generation of tools used for production logging and reservoir monitoring. This suite of tools is used to improve understanding of downhole flowing conditions and to expedite the evaluation of the reservoir The Gas Holdup Tool (GHT) has proved to be a major component of this new approach. Less than 3 feet in length, the GHT is a 1-11/16-inch-OD through-tubing production logging tool, which directly and accurately determines the volumetric fraction of gas over a crosssectional volume element of the wellbore. This sensitive downhole tool operates in horizontal, highly deviated, and vertical cased wells. It uses real-time downhole pressure and temperature, casing inside diameter, and gas gravity measurements to generate a 0% to 100% gas holdup log in stratified or homogenized flows. The GHT uses a low-energy cobalt-57 source and a sodium iodide detector located a short distance from the source and separated by a tungsten shield. The detector counts gamma rays that are scattered back from the production fluid to the detector; the magnitude of the count rate is directly related to the gas holdup. The measurement is not affected by the composition and density of materials outside the casing. Log examples illustrate a variety of applications of the fullbore gas holdup tool measurements, Guidelines are provided for running production logs for evaluating horizontal completions and improving reservoir production and value. The examples demonstrate the new wellsite production analysis program, which incorporates new interpretation algorithms, and features fast automatic processing capability to permit the user to quickly determine the type and rate of produced fluids while the logging unit is still on location. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER J MONITORING ANNULAR WATER FLOW IN DUAL CASING STRING COMPLETIONS USING CONTINUOUS OXYGEN ACTIVATION LOGGING David M. Chace, Hans-Christian Freftag, Darryl E. Trcka, and René W. Mayer Western Atlas Logging Services Ahmed W. Sufi Zakum Development Co., UAE ABSTRACT New continuous oxygen activation logging technology is being used to monitor injection profiles in several carbonate reservoirs in the Middle East, where dual string completions are employed to control relative injection into intervals of markedly different permeability. In dual string completions, where short string water injection is conveyed via the tubing-casing annulus, oxygen activation logs often provide the only method for monitoring injection performance and for evaluating add and polymer treatments used to modify injection profiles. Accurate measurement of water flow in the annulus leads to a better understanding of hydraulic properties of the reservoir and improved design of water injection programs. Oxygen activation logging is also being applied to monitor water entry and rates in producing wells. Recent literature has described the latest improvements in logging instrumentation and interpretation methods allowing continuous oxygen activation logging for diagnosis of complex water flow conditionst5’6~. This paper focuses on recent experience in the Middle East with applications to a number of challenging monitoring problems uniquely associated with monitoring the short string injection or production in dual string completions, including: • • • • • • • Determination of short string injection profiles for multiple sets of perforations Delineation of injection into sublayers of a reservoir within a single set of perforations Monitoring injection profiles during multi-rate injection tests Identification of fractured intervals and effects on injection profiles Monitoring the effectiveness of polymer treatments for water shut-off Mechanical integrity testing and location of channeling or tubing leaks Determination of water entry points and measurement of water rates in producing wells Recent examples with logs and histories are presented highlighting many of these applications. In one particular example, a comparison of continuous oxygen activation logs run in the dual completion and conventional spinner flowmeters run in casing prior to the dual completion illustrates the viability and accuracy of the method. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER K A NEW SMALL-DIAMETER, HIGH-PERFORMANCE RESERVOIR MONITORING TOOL L. A. Jacobson, R. Ethridge, and G. Simpson Halliburton Energy Services ABSTRACT This paper discusses a new small-diameter, high performance pulsed-neutron spectrometry tool that has been introduced recently for reservoir monitoring applications. Basic measurement principles are explained, and important operating parameters are presented. Field examples illustrate multiple applications of the new device and compare the performance of the new instrument with a larger pulsed-neutron spectrometry tool. A summary of main points presented in the paper follows. The new 2-1/8-inch tool is designed for logging through 2-7/8-inch or larger tubing. The tool is slightly larger than traditional 1-1 1/16-inch through-tubing tools. This design accommodates larger detectors, thereby improving the performance of the carbon-oxygen (C/O) measurement. Two bismuth germanate detectors yield high gamma ray count rates with good spectral resolution. To optimize C/O measurements, one mode of tool operation interlaces 5ms of background measurements with 2Oms of 1O-kHz neutron pulses. The tool records 256-channel spectra from each detector: • • • during the neutron bursts (for analyzing inelastic events), during the interpulse period (for analyzing capture events), and during the background period (for analyzing activation and background events). The tool also provides a simultaneous neutron capture cross section (sigma) measurement. Carbon-oxygen and calcium-silicon (Ca/Si) response characteristics for saturation and lithology analysis are derived from laboratory measurements and are illustrated in the customary fan-chart format. In a second mode of operation, the tool optimizes the sigma measurement and provides high-quality neutron capture spectra for quantitative lithology determination. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER L THE APPLICATION AND ACCURACY OF GEOLOGICAL INFORMATION FROM A LOGGINGWHILE-DRILLING DENSITY TOOL Erhard Bornemann Schlumberger Oilfield Services Tim Bourgeois and Ken Bramlett Shell Deepwater Production Inc. Kyel Hodenfield and David Maggs Schlumberger Oilfield Services ABSTRACT Since their introduction in 1986 electrical borehole images have become the major source of geological information from log data. Structural dip, sediment transport direction, reservoir geometry, thin-bed analysis, dual porosity systems, even permeability can be recognized or estimated with greater confidence by using high-resolution images. Recently a clear industry trend toward geosteered, highly deviated wells has emerged. This is particularly apparent in the high stakes offshore development arena where Logging While Drilling (LWD) tools are often the preferred, if not the only logging choice, and where the range of imaging tools is limited. A new 4¾ in LWD density sonde measures Bulk Density (RHOB) and Photoelectric effect (Pe), in sixteen sectors around the borehole. In a 6 inch hole, it can provide formation images with a pixel size of about 1.2 in. The tool functions in conductive, non conductive and oil-based muds. Initial trials have shown that lithology boundaries are well defined, and allow the computation of dips for structural analysis. For net pay estimation, the density image provides an azimuthally selected density trace that is relatively free from hole effects. These applications are demonstrated in two horizontal wells from the Gulf of Mexico, deepwater Ram Powell project. The examples illustrate the successful application of density images for computing structural dip, estimating net pay and identifying large scale stratigraphic features. Recommendations on the acquisition and use of density images are made. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER M AUTOMATIC HIGH RESOLUTION TEXTURE ANALYSIS ON BOREHOLE IMAGERY Shin-Ju Ye Image Laboratory, Institute of Geodynamics University of Bordeaux, Talence, France Philippe Rabiller and Noomane Keskes Elf Exploration Production, Pau, France ABSTRACT This paper describes a new automatic processing methodology for extracting texture information from electrical borehole images. A field case is used to illustrate the successive stages of the methodology and its integration in sedimentological interpretation. The texture of rock is an important item of geological information since it affects fluid flow, and the recovery factor. Texture is also the signature of the depositional mechanisms which control reservoir architecture and its extension. Only by coring or borehole imaging, can the texture of the rock be described. The method proposed here obviates restrictive a priori assumptions about the texture types. A statistical texture model proposed by S. D. Ma and A. Gagalowicz is first used to extract texture parameters (histogram and autocovariance function). The choice of this model is based on the possibility to create a synthetic image which is comparable to the original one, in order to control the representativity of texture parameters used. These statistical parameters can characterize almost all possible textures, even those which were not observed before. We then use the self-organizing map of Kohonen for classifying texture parameter and ordering the texture classes by their visual similarities. The advantage of this classifier is that no a priori knowledge about observed textures is necessary. This rules out the possible bias introduced by interpreters. This procedure generates textural logs which can be integrated with conventional wireline logs. A 2-D synthetic texture map is created in order to facilitate the calibration of the textural logs to cores. Thanks to the ordered texture classes, the vertical variation of sedimentary and petrophysical relative changes in rock textures can be easily assessed from textural logs. The method has been tested on several field cases. Combined with the results of facies analysis of conventional logs and high resolution dip trends from borehole, images, this methodology has proved its capacity to improve sedimentological interpretation. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER N NEAR-WELLBORE 3D RECONSTRUCTION OF SEDIMENTARY BODIES FROM BOREHOLE ELECTRICAL IMAGES H. Anxionnaz and J.P. Delhomme Schlumberger Wireline & Testing ABSTRACT Borehole electrical image analysis unravels the internal organization of the sedimentary bodies crossed by wells. It is the key to determining the direction of sediment transport at the time of deposition, which in turn conditions the overall geometry of the sand bodies. From a production standpoint, cross-beds as well as parallel beds can have contrasting permeability in alternating laminae, resulting in a significant anisotropic effect on fluid flow. However, the full use of borehole images for reservoir-oriented applications in cross-bedded formations is still hampered by a cylindrical view restricted to the borehole size, coupled with the limited ability of interpreters to visualize in 3D. A new 3D presentation of the information conveyed by sedimentary dips is proposed: a series of block diagrams is drawn along the borehole trajectory. The diagrams correspond to individual cross-bed units (cross-beds with similar morphology and orientation), are oriented along the principal directions of deposition of these units and extend to a few meters around the wellbore. By keeping a 1:1 aspect ratio, this display presents crossbed laminae in their true 3D setting. It epitomizes the sedimentary information contained in the borehole images in an accessible way, even at a 1/40 or a 1/100 scale. The reconstruction of bedform geometry away from the wellbore, along the walls of each cubic block diagram, is constrained by the information available on the cylindrical borehole wall. This task is achieved by numerically solving an inverse problem. A possible solution to the corresponding forward model consists of a superposition of sine surfaces that represent the interfaces between laminae. A similar parametrization of the problem is used in the current implementation for the inverse problem. In addition to this geometrical reconstruction of foresets and their bounding surfaces, a 3D reconstruction of the spatial distribution of electrical conductivities is performed. A case study from cross-bedded deposits is shown to illustrate the technique. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER O ANALYSIS OF FULL-WAVEFORM SONIC DATA ACQUIRED IN UNCONSOLIDATED GAS SANDS Brian E. Hornby and Eric S. Pasternack ARCO Exploration and Production Technology ABSTRACT Experimental full-waveform sonic data were collected in an exploration well penetrating unconsolidated sand-shale sequences. Using a low frequency (2,000 Hz) compressional wave source drive, formation compressional (P) wave arrivals were recorded in acoustically-slow, gas bearing sands. In some intervals formation velocities were slower than the borehole fluid speed. Standard (10,000 Hz) P-wave acquisition failed to detect formation signals across the same intervals. Shear wave acquisition was accomplished using dipole transducers. Interpretation of the measured Vp/V3 ratio highlighted gas-bearing intervals where the Vp/V5 ratio dropped below the background compaction trend. One of these gas-bearing intervals had commercial saturations of gas while the other zones were both sands and shales with low, non-commercial gas saturations. The zones with low gas saturation appeared to be caused by a variety of mechanisms. Very slow P-wave responses (borehole fluid speed or slower), positively indicating the presence of gas, were noted in several shale formations. Using a model of thin, gas charged silts or sand embedded in the shale matrix, effective medium modeling indicated that the recorded slowness could be caused by a concentration of less than 20 percent gas charged silt layers. This is significant because it indicates that gas charged shales can give amplitude anomalies, indicating hydrocarbons, on the seismic section. Slow P-wave responses indicating the presence of gas were also noted in high-porosity sands that appeared wet, with the water saturation approximately 90 percent, using conventional log interpretation. Pressure profiling gave a water gradient, confirming that water was the dominant phase in those intervals and indicating only a low gas saturation. However, this low gas saturation was sufficient to cause a significant slowing of the P-wave velocity and a corresponding amplitude anomaly on the seismic section. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER P A NEW METHOD FOR DERIVING PERMEABILITY FROM BOREHOLE STONELEY WAVES AND ITS APPLICATION IN NORTH MONAGAS FIELD OF EASTERN VENEZUELA A. Sinha, M. Rangel Geoscience, Western Atlas Logging Services, Venezuela R. Barbato PDVSA, Puerto La Cruz, Venezuela X. Tang Western Atlas Logging Services, Houston, USA ABSTRACT With the proposed new advance technique for estimating permeability from borehole acoustic logs, a continuous formation permeability profile can now be routinely obtained from full wave acoustic log. The success of this method lies in the separation of permeability related wave attributes from other effects unrelated to permeability. The new method uses an effective processing procedure to suppress noise and scattering effects in the data. It then performs wave modeling to account for wave attributes related to borehole and elastic property changes. Finally, it applies an inversion procedure to derive permeability from fluidflow related wave attributes. The permeability obtained from this method has been compared with MRJL (Magnetic Resonance Imaging Log) derived permeability with remarkable correlation. This technique was applied in the North Monagas field of Eastern Venezuela to derive Stoneley permeability from MAC (Multiple Array Acoustic) data The formation consists of sand / shale sequences with porosity around 20% and permeability up to a few hundred millidarcies. In wells where, in addition to Full Wave Acoustic, Magnetic Resonance Imaging Log (MRIL) and Borehole Image logs were also recorded, the similarities and differences between two permeability profiles provided valuable information for formation evaluation. MRIL derived permeability reflects porosity and pore size distribution and is primarily due to rock matrix. Stoneley-derived permeability, however, is related to fluid movement and it is more sensitive to formation fractures. Thus, by comparing MRIL and Stoneley permeability, it was possible to separate permeability contributions arising from the matrix and fractures. The fractures were identified with the help of Borehole Image log. Since Stoneley waves are related to fluid movement, the wave attributes are also affected by formation fluid viscosity and compressibility which can help to detect formation fluid property variations by comparing MRIL and Stoneley permeability profiles. Such comparisons demonstrated the value of combined Stoneley wave and MRJL permeability measurements. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER Q FLUID IDENTIFICATION THROUGH DYNAMIC MODULUS DECOMPOSITION IN CARBONATE RESERVOIRS Raghu Ramamoorthy and William F. Murphy III, Schlumberger-Doll Research, Ridgefield, Connecticut, USA ABSTRACT It has been proposed that the appropriate method for interpreting petrophysical properties from acoustic logs is through interpretation in the modulus domain. The frame shear and the frame bulk moduli have been shown to be a function of porosity in quartz reservoirs. We can use the total porosity from logs to estimate the dry frame bulk modulus and, hence, the bulk modulus of the pore fluid, thereby identifying the saturating fluid. Here we extend the method to limestone reservoirs. We demonstrate that we can identify the presence of hydrocarbons based on the shear, compressional, density and porosity logs. The procedure must account for the pore geometry that is observed in these rocks. We propose the use of more accurate relationships among the acoustic moduli, porosity and pore type. The method is applied to a carbonate formation in the Middle East that has hydrocarbon-bearing zones. The presence of hydrocarbons as indicated by acoustic logs is confirmed by resistivity logs and from local knowledge. Many investigators have questioned the validity of time-lapse four dimension (4D) seismic studies and direct hydrocarbon detection in carbonates. They observed an apparent lack of sensitivity to fluid content in laboratory experiments on compressional velocity. We expect, based on theory and experiment, a measurable effect with both oil or gas depending on the gas/oil ratio of the oil. We discuss the applicability of the procedure to study the feasibility of using 4D seismic studies to monitor carbonate reservoirs. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER R ADVANCED INTERPRETATION OF FRACTURED CARBONATE RESERVOIRS USING FOURCOMPONENT CROSS- DIPOLE ANALYSIS Bob Joyce, Doug Patterson, and Jane Thomas Western Atlas Logging Services, Houston, TX ABSTRACT Industry demand for interpretive techniques utilizing acoustic full-wave compressional and shear data sets has been growing in the last decade. The increase in industry demand is due to the need for more efficient use of the petrophysical and geophysical data sets gathered on exploration and development logging programs. This is evident as more questions arise in regard to how borehole acoustic data can be used in increasingly complex solutions above and beyond the conventional seismic tie and rock properties. Recent advances in acoustic dipole logging techniques and processing have resulted in enhanced accuracy for interpreting location and direction of natural fractures in carbonate reservoirs. Quality four-component cross-dipole data are combined with new, advanced geoscience solutions to identify fracture regimes and their azimuthal distribution in the reservoir. The solution utilizes a new, robust package exhibiting unique inversion and wavematching processing techniques ideally suited, but not limited to fracture determination in carbonates. The interpretation features are presented and a comparison of the new processing technique to conventional cross-dipole processes is discussed. To substantiate the analysis, two very different evaluation methods using image data and Stoneley permeability are compared. Quality indicators and wellbore constraints on azimuthal dipole measurements will also be discussed. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER S IMPORTANT IMPLICATIONS FROM A COMPARISON OF LWD AND WIRELINE ACOUSTIC DATA FROM A GULF OF MEXICO WELL. Paul Boonen Halliburton Energy Services Clarke Bean Chevron USA Production Company Rob Tepper and Ron Deady Halliburton Energy Services ABSTRACT This paper demonstrates that high-quality LWD acoustic data can be obtained over intervals where wireline data may be unreliable. Early-time data acquisition and special tool technology contributed to this LWD capability. In some instances, the high quality of the LWD data has resulted in improved time-to-depth correlations, The paper analyzes data from LWD and wireline acoustic tool runs in a Gulf of Mexico well drilled with oil-based mud. Consistent differences between data from the two tools suggest pronounced shale alteration may have occurred between LWD and wireline runs. In zones where the borehole was washed-out above 14 inches, sonic data quality decreased for both wireline and LWD tools. In good hole conditions, the systematic differences observed between LWD and wireline in sandstones may be explained by different pore fluids (resulting from invasion) at the various times of logging. The LWD acoustic tool used on the well employed dense sampling and downhole semblance processing. These features allowed lower-quality waveform data to be discarded and only higher-quality waveform data to be stored in tool memory for supplementary uphole processing. Because only higher-quality data were used downhole to determine At values, higher confidence could be placed in the At values transmitted uphole while logging. As an added benefit, the efficient use of memory via selective storage of data permitted longer intervals to be logged before tripping was needed to retrieve memory data. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER T NEW 3 1/8 INCH DIAMETER PROPAGATION RESISTIVITY TOOL FOR MWD; SMALL SIZE BUT NO ACCURACY COMPROMISE Mac M. Wisler, Larry W. Thompson, Andreas Larscheid, John A. Signorelli, Manna Pope, and Steve M. Wright Baker Hughes INTEQ ABSTRACT A new 3 1/8 inch diameter propagation resistivity tool has been built for coiled tubing and rotary drilling applications. This tool is the first to provide resistivity measurements with the accuracy required for formation evaluation and geosteering in the very slim hole sizes. The development was made possible by achieving size reduction of electronics and antennas without sacrificing measurement accuracy or tool strength. The tool can be run in holes as small as 3 ¾ inches rates with build rates of 30 degrees per 100 ft. Applications are in geosteering and formation evaluation in the reentry and very slim hole drilling markets. The tool works at two frequencies (2 MHz and 400 KHz) with a single spaced compensated antenna design. The ultra-small state of the art electronics and robust antenna design yield measurement performance as good or better than the present state of the art in resistivity tools, while preserving strength for drill string integrity. Four fully compensated resistivity measurements are made at four different depths of investigation. Phase resistivity accuracy at 2 MHz is better than 1/2 mmho at high resistivities and 1 percent at low resistivities. The measurement range is 0.1 ohm meters to 3000 ohm meters in all mud types. Details of the design are presented along with well logging data. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER U IMPLICATIONS OF RESISTIVITY DISPERSION ON LWD LOG INTERPRETATIONS Ali A. Garrouch Department of Petroleum Engineering, Kuwait University ABSTRACT For matching LWD and wireline resistivity logs, it has been customary to include the effects of several factors such as invasion, dip, anisotropy and dielectric constant. Experimental results were obtained by measuring the complex electrical resistivity of Berea sandstone, tight-gas-sand rocks and Ottawa sand-bentonite mixtures saturated with NaCl brine solutions at frequencies between 10 Hz to 10 MHz. These results indicate that rock resistivity becomes dispersive at frequencies below 1 MHz for the various conditions of salinity, wettability, clay content, and degree of rock consolidation investigated. Therefore, an additional correction for frequency dependence has to be added for matching LWD and wireline resistivity logs. Similar conclusions have been reached by simulating the complex impedance of shaly sands using a generalized Hanai-Bruggeman model, and carbonate rocks using the Complex Refractive Index model. Supporting well log examples were obtained in non-permeable nondipping shales of two Conoco test wells. A systematic trend of steady resistivity decrease with increasing frequency has been observed with all the electromagnetic tools operating at frequencies ranging from 20 KHz to 1 GHz run in these test wells. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER V HIGH-DEFINITION LATERAL LOG - RESISTIVITY DEVICE: BASIC PHYSICS AND RESOLUTION G.B. Itskovich, A.G. Mezzatesta, K-M. Strack, and L.A. Tabarovsky Western Atlas Logging Services, Houston, Texas ABSTRACT The High-Definition Lateral Log (HDLL) system is a new logging resistivity array developed by Western Atlas Logging Services (WALS) in cooperation with Shell International Exploration and Production, Research and Technical Services (SIEP-RTS). The logging instrument has been designed to provide high-resolution resistivity distributions in both the radial and vertical direction. The HDLL tool has a single-current electrode and acquires a multitude of measurements, including electrode potentials and electric fields (represented by first-potential differences) at several distances from an injection electrode. The high accuracy under which the electric fields are measured allows for an accurate calculation of the potential second differences that are used to detect formation boundaries. High-resolution estimates of formation resistivities are obtained by inverting the HDLL acquired data. Based on a subset of the acquired data, a number of synthetic-focused measurements can be generated that provide a first indication of the formation resistivity distribution in the formation surrounding the borehole. A separate paper at this conference presents techniques for generating synthetic curves, HDLL inversion, and results of a field study. Here, we mainly focus on the basic physical principles and HDLL array resolution. We evaluate three zones of a current flow, which have different links to formation parameters and also consider an asymptotic theory leading to the concept of current leakage. The resolution of the HDLL array is also discussed and we analyze how noise affects the interpretation. The resolution analysis comprises, in a single scheme, the simulated data, their associated sensitivity to formation parameters, and a statistical noise model. Synthetic and field examples are presented that illustrate the high-information content and resolution power of the HDLL instrument. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER W A NEW SIMULTANEOUS ANISOTROPY AND DIELECTRIC CORRECTION ALGORITHM FOR LWD RESISTIVITY MEASUREMENTS C. Jackson and T. Hagiwara Halliburton Energy Services ABSTRACT A correct resistivity measurement is crucial to an accurate formation evaluation. Wireline resistivity measurements are often affected by invasion, shoulder bed properties, and environmental effects, such as borehole size and tool position. Though Logging While Drilling (LWD) resistivity measurements may be affected less by these factors, the use of a high frequency (.4 to 2 MHz) induces other unwanted effects in the LWD resistivity measurements. The most significant effects are due to the dielectric constant and the resistivity anisotropy of the formation. They may affect log response even in homogeneous formations when other effects are small. These two effects can be present simultaneously. This paper demonstrates a new algorithm that simultaneously corrects for both dielectric and anisotropy effects when using multiple-spacing LWD resistivity measurements. This algorithm determines the true formation (horizontal) resistivity, the resistivity anisotropy, and the formation dielectric constant simultaneously. Several field examples are presented that demonstrate when the effect of the dielectric constant and the resistivity anisotropy, or both, had to be corrected to obtain the true formation resistivity from the LWD measurements. The results are compared with those without simultaneous corrections and where only enough resistivity data are available to correct for variations in a single formation property, either the dielectric constant or the formation resistivity anisotropy. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER X DUAL LATEROLOG RESPONSE IN 3-D ENVIRONMENTS H. M. Wang and L. C. Shen University of Houston, Houston, Texas, U. S. A. G. J. Zhang University of Petroleum, Dongying, Shandong, China ABSTRACT Computer codes based on the finite element method (FEM) have been developed to model dual laterolog responses in 3-D environments. Validation of the 3-D FEM codes are carried out by showing that results of the codes agree with those obtained by an analytic solution, 2D FEM codes, and 3-I) hybrid method codes. Using the 3-D FEM codes, dual laterolog responses in highly deviated wells, horizontal wells, dipping anisotropic formations, and fractured carbonate formations are studied. For a laterolog tool in a formation consisting of a resistive bed between two conductive shoulder beds, one tends to believe that the greater the dip angle is, the lower the reading of the apparent resistivity at the center of the bed will be. However, such a behavior is not always true. The resistivity reading depends on the bed thickness and on the resistivity contrast between the resistive bed and the shoulder beds, In some cases, it is seen that when the dip angle varies from 0º to 75º, readings at the center of the resistive bed actually increase with the dip angle. Therefore in thin dipping beds, the response of either the deep or the shallow laterolog is difficult to predict without carrying out computer modeling. The study also shows the dual laterolog response in a horizontal well in a formation containing a resistive bed. The readings of both the shallow and the deep tools are strongly affected by the shoulder beds when the tools are in the resistive bed. Even when the bed thickness is increased to 3 m, shoulder beds still have a strong influence on the deep laterolog. The effect of anisotropy on dual laterolog is small even at large dip angles. The apparent resistivity readings of dual laterolog are determined mainly by the horizontal resistivity of the anisotropic formation, As expected., in a fractured formation the decrease in the apparent resistivity reading is directly related to the increase of the opening width of the fracture and to the decrease in the resistivity of the fluid in the fracture. It is shown that in some cases the fracture with at least a 5-rn extent has the same effect as the one with an infinite extent; and a 100 micrometer fracture opening has an apparent thickness of one meter. When there are at least five fractures within a meter along the borehole axis, then the fractured formation can be modeled as an anisotropic formation. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER Y GETTING THE HYDROCARBON VOLUMES RIGHT - A QUANTITATIVE USE OF NMR DATA IN WATER SATURATION MODELLING Lars Helge Flølo Statoil, Norway ABSTRACT A saturation model which utilises the NMR bound fluid measurement is proposed for the purpose of calculating the volume of hydrocarbons in a reservoir. In this model the water saturation is split into an irreducible water saturation and a capillary bound water saturation. The irreducible water saturation is derived from the volume of bound fluid as determined by the NMR tools, and the capillary bound water saturation which is a height dependant quantity, is derived from capillary pressure data or from log saturations in wells penetrating the transition zone. The saturation model expresses the water saturation as a function of porosity, permeability and height above the free water level. These are the variables that are typically being mapped in a reservoir mapping process. The irreducible water saturation and the capillary bound water saturation is therefore expressed as functions of these variables. In addition it is shown how the Timur-Coates NMR permeability equation may be used either to generate permeability maps, or as direct input to the model equation. The saturation model is presented with an example from the Heidrun Field in the Norwegian Sea. The saturation from the model matches very well the saturation calculated from resistivity logs in the wells. The saturation modelling concept was also found particularly useful in the stochastic reservoir model which was developed for the field. This saturation model only requires that an NMR tool is run in one key well in a field to establish the necessary correlations. The only required data from the NMR tool is the bound fluid log which in principle can be acquired at normal logging speeds. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER Z MINIMUM AND MAXIMUM PAY ESTIMATION USING RESISTIVITY LOG DATA INVERSION M.A. Frenkel and A.G. Mezzatesta Western Atlas Logging Services, Houston, Texas ABSTRACT The estimation of oil in place heavily depends on the accuracy of resistivity data and the reliability of their interpretation. Conventional interpretation of resistivity logging data is normally accomplished by using computer based correction charts that are generally applied on a level-by-level basis. This type of interpretation is approximate and does not allow for extracting all information contained in the data. This may result in overlooking oil- and gasbearing formations. Modeling and inversion methods allow one to produce reliable estimates of formation properties. In addition, inversion provides means for incorporating data uncertainties in the process and estimate the uncertainties of formation properties. Uncertainties (error bounds) in estimated formation properties can then be used to provide minimum and maximum pay estimates. A case study is presented that illustrates the ability of inversion technology to provide lower and upper limits for the estimated formation properties and shows how these results can be extended to compute lower and upper limits of hydrocarbon estimates. The case study is from a well in Oman for which resistivity data from the High-Definition Lateral Log (HDLL) service is available. A two-dimensional (2-D) resistivity inversion was performed, followed by a petrophysical interpretation of all log data available on this well. The interpretation results show the capability of the proposed methodology to obtain both an accurate reservoir delineation as well as minimum and maximum pay estimates. These results clearly show the benefits of an integrated evaluation of oil/gas-bearing reservoirs. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER AA SUCCESSFUL PREDICTION OF WELL PRODUCTIVITY FROM OPENHOLE LOGS IMPROVES PROFITABILITY IN SEVERAL ROCKY MOUNTAIN FORMATIONS: CASE HISTORIES S. Schubarth, M. Mullen and C. Kessler Halliburton Energy Services ABSTRACT This paper describes methods used to quantify reservoir properties through the analysis of openhole logs and integrate the results into completion and stimulation designs. The results show that accurate prediction of well productivity can be made such that stimulation treatments can be designed to truly optimize production from a well. Optimization of stimulation treatment design is greatly dependent on understanding the permeability of the formation. Direct measurement of permeability is not economically available in many “tight” formations. We will demonstrate through integrating log analysis with production evaluation that correlations can be constructed which allow the analyst to interpret in-situ formation permeability from conventional openhole logs. The three formations being presented are: the Mesaverde group of the Piceance Basin, the Frontier formation of the Moxa Arch and the Green River formation in the Uinta Basin. The first two of these formations are “tight” gas sands, and the third is a low-permeability oil-producing formation. In each case the authors have been successful in increasing profitability of the well completions. The Piceance Basin case involves development of a reservoir containing multiple sand lenses. The methods will demonstrate how to accurately predict the gas to be produced from each sand, and therefore improve the efficiency of a completion by selecting only the sands that will contribute significant quantities of gas to pay for the expense of completion. The Moxa Arch case involves development of a single sand. The variation in reservoir quality between wells suggests that variations in stimulation treatment design could improve economics of the development. This paper presents the method used to quantify the individual well quality and design the optimum treatment for that unique well. The Uinta Basin case extends the concepts developed in the two “tight” gas sand plays to oilproducing sands. The results from this development project indicate that production has more than doubled through improved treatment design, which comes from increased reservoir understanding through log analysis. We feel that the presentation of the various methods used in this paper will demonstrate the value of inter-discipline communication concerning the economics of oil & gas exploration and development. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER BB APPLYING INTEGRATED FORMATION EVALUATION TO ADVANCED RESERVOIR CHARACTERIZATION IN CALIFORNIA’S MONTEREY FORMATION SILICEOUS SHALES Thomas A. Zalan, Mike F. Morea, and Dale R. Julander Chevron North America E&P Stan A. Denoo Schlumberger Wireline & Testing ABSTRACT A joint project between Chevron and the United States Department of Energy is underway to determine the technical feasibility of implementing a CO2 enhanced oil recovery project in the Antelope Shale (Monterey Formation siliceous shale equivalent) in the Buena Vista Hills Field, Kern County, California. The project consists of four main components: Reservoir Matrix and Fluid Characterization, Fracture Characterization, Reservoir Modeling and Simulation, and CO2 Pilot Flood and Evaluation. This paper focuses on the first phase of the project to apply a variety of advanced reservoir characterization techniques to determine the production characteristics of the Antelope Shale reservoir. The Antelope Shale in the San Joaquin Valley contains an estimated 7 billion barrels of oil in place. The Monterey Shale that underlies much of California’s coastal area and offshore holds nearly 3 billion barrels of oil in place. Thus, the focus of this project is the 10 billion barrel (oil in place) fractured siliceous shale resource. Extensive core and log data were acquired in a 1996 project well. Mineral end points for each of the important minerals in the Brown Shale and Upper Antelope Shale members were determined. Accurate determinations of reservoir mineralogy over selected intervals were then made for testing the accuracy of mineral and lithology determinations from mineralbased wireline log analysis. Three different mineral models from three different combinations of logging suites were developed and compared to core mineralogy. NMR porosity was found to closely agree with core porosity. The NMR log saw most of the reservoir fluid as irreducible. This was expected, as most of the matrix porosity is constituted by micropores in the diagenetically altered siliceous shale. A satisfactory NMR based permeability model has not yet been developed for siliceous shale. For assisting in choosing a completion interval for the 1996 project well, carbon/oxygen log results were analyzed to high-grade the choice of completion interval. Correlations of porosity and drill stem test permeability to shale volume from spontaneous potential were developed to extend the reservoir characterization fieldwide, where most wells have only vintage electric logs and spontaneous potential. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER CC ACCURATE OIL SATURATION DETERMINATION USING CARBON/OXYGEN LOGS IN THREEPHASE RESERVOIRS Ahmed Badruzzaman Chevron Petroleum Technology Robert H. Skillin and Thomas A. Zalan Chevron North America Exploration & Production Tahmina Badruzzaman Pacific Consultants & Engineers Phuong T. Nguyen, Chevron Petroleum Technology ABSTRACT The ability of carbon/oxygen logs to predict oil saturation in three-phase reservoirs (steam, oil and water) is examined using Monte Carlo modeling and field data. First, the oil saturation from carbon/oxygen (GO) logs is compared with that from core data. Then conventional two-phase algorithms and ad-hoc modifications to such algorithms to account for the lower density of the formation fluid in such reservoirs are tested. Finally, a three-phase saturation equation relating the two-phase saturation to an independently determined gas saturation is derived from theory. Saturation prediction in three-phase reservoirs using both the spectral-fitted GO ratio and the windows ratio are examined using the new algorithm. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER DD UTILIZING MWD GEOMETRY AND INNOVATIVE NEUTRON DETECTION TECHNIQUES TO MINIMIZE THE EFFECTS OF THERMAL NEUTRON ABSORBERS ON LOG RESPONSE Daniel C. Minette Baker Hughes INTEQ ABSTRACT The advantages of utilizing epithermal neutrons for the measurement of formation porosity have long been known. Epithermal neutron measurements are less influenced by environmental effects such as borehole salinity. They are not dependant on trace amounts of elements with high thermal neutron absorption cross sections. Thus, they can give a more accurate determination of the porosity of formations. Even with these known advantages for epithermal neutrons, thermal neutron measurements still predominate. This is the result of the reduction in count rate typically associated with epithermal detectors. Most epithermal detectors are ordinary detectors surrounded by a thin sheet of a strong thermal neutron absorber, such as cadmium, which results in count rates < 20% of their thermal counterparts. A new method for obtaining epithermal neutron measurements without shielded detectors has been developed. This method utilizes the additional steel that exists in a MWD tool to thermalize neutrons as well as detectors that have a higher effective epithermal neutron cross section to obtain a predominantly epithermal neutron measurement. This technique eliminates the drastic reduction in count rate associated with shielded detectors. The epithermal nature of the measurement will be demonstrated through the use of both Monte Carlo modeling and test pit data. As a first step, data are taken with cadmium shielded detectors in the test pits. This afforded a measurement of the epithermal/thermal mix of neutrons counted at the detectors. Subsequently, Monte Carlo runs are made both with and without a thin cadmium sheet surrounding the tool. Use of this shield, which virtually eliminates the thermal neutron flux entering the tool, will allow for a calculation of the fraction of detected neutrons that were epithermal when they entered the tool. By combining these two methods, one can determine the ratio between formation epithermal neutrons and formation thermal neutrons measured by the detectors This understanding will be expanded through the analysis of the tool response in known formations. These test pit measurements focus on the responses most dependent on the thermal absorption cross section. Included in these are borehole salinity, formation salinity, and water tank with varying salinities. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER EE PRESSURE TEST ANALYSIS OF GAS BEARING FORMATIONS Jaedong Lee, Ekrem Kasap, and Amit Sarkar Western Atlas Logging Services ABSTRACT The Reservoir Characterization Instrument (RCI), a new generation wireline formation tester developed at Western Atlas, is capable of controlling fluid sampling operations from the surface, and acquiring and analyzing corresponding rate-pressure data in real-time to estimate formation pressures and core-scale permeabilities. These data, along with fluid properties derived from the collected samples, are essential in evaluating reservoir performances, and designing well completions and surface facilities. For interpretation of pressure test data in liquid saturated formations, a Formation Rate Analysis (FRA) technique has been developed recently. The technique is not directly applicable to gas formations because compressibility, density and viscosity of gas are strong functions of pressure. Moreover, appreciable temperature changes and inertial pressure drop or non-Darcy flow effects may occur during gas formation testing. This paper modifies the FRA technique for gas formation testing and is termed OFRA. The method calculates gas pseudo-potentials and analyzes the variation of pseudo-potential versus formation rate during a pressure test by utilizing the geometric factor concept. Forchheimer’s equation, instead of Darcy’s equation, was used to study the non-Darcy flow effects. The technique is verified using simulated well test data generated using a 3-fl near wellbore simulator. Finally, the technique is applied to a field test to estimate the formation permeability. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER FF SOME EXCEPTIONS TO DEFAULT NMR ROCK AND FLUID PROPERTIES Q. Zhang, S.-W. Lo, CC. Huang, G.J. Hirasaki, R. Kobayashi, and W.V. House Rice University ABSTRACT Two correlations currently exists for the estimation of the hydrogen index (HI) of crude oils. One is based on API gravity of stock tank oil and the other on the density of alkanes. The former will not describe the reduced HI of gas condensate systems and the latter will overestimate the HI of heavy, aromatic crude oils. Both systems can be correlated with density and the hydrogen:carbon ratio. The hydrogen:carbon ratio of the stock tank oil can be determined by measuring the NMR HI and density. The HI of the live oil can then be predicted from the formation volume factor and solution gas/oil ratio. The relaxation time of crude oils is currently estimated by a viscosity/temperature correlation. These correlations are usually based on measurements at ambient conditions with the oil saturated with air and in the absence of methane. Hydrocarbons have a significant air solubiity and molecular oxygen will significantly shorten the relaxation time of light hydrocarbons. Pure methane relaxes by a different mechanism than heavier hydrocarbons and the relaxation time has a dependence on viscosity/temperature that is the inverse of that for liquid hydrocarbons. Live crude oils (modeled as mixtures of methane and n-decane) have a relaxation time dependence on viscosity/temperature that deviates from the correlation based on hydrocarbon liquids at ambient conditions. The estimation of BVI assumes that the relaxation time distribution below the T2 cutoff is due to capillary-bound water in the smaller pores. However, crude oil may have a relaxation time distribution with a significant overlap in the region of the By!. This becomes worse when there is wettability alternation. This results in overestimation of By! and underestimation of permeability. The surface relaxivity is the parameter that correlates the brine relaxation time distribution to the rock pore size distribution. The T2 cutoff for estimating BVI/FFI and the prefactor in the T2,LM permeability correlation are dependent on the value of surface relaxivity. It was observed that a 200 md sandstone with pore lining chlorite has T1,LM/T2,LM >> 1.6 with the ratio increasing with echo spacing. Also, the BVI/FFI T2 cutoff for this sandstone is 18 ma and 5 ma for echo spacing of 0.2 ma and 2 ma respectively; much less than the default value of 33 ms. These results depart from typical values for sandstones and are due to the internal gradients resulting from the high magnetic susceptibility of the iron-rich chlorite lining the pore walls. It is usually assumed that if a system is water-wet, the oil will relax at the same rate as bulk oil. T2 measurements with this chlorite coated sandstone are an exception to this assumption because the oil is relaxed by echo-spacing dependent diffusion resulting from the large internal gradients. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER GG ENHANCED DIFFUSION: EXPANDING THE RANGE OF NMR DIRECT HYDROCARBON-TYPING APPLICATIONS Ridvan Akkurt NMRPlus, USA Duncan Mardon Exxon Exploration Co., USA John S. Gardner, Dave M. Marschall NUMAR, USA Fernando Solanet YPF, Argentina ABSTRACT This paper introduces a new gradient NMR logging technique, called the Enhanced Diffusion Method (EDM), which exploits the diffusion contrast between oil and water to separate the NMR signals from each fluid. Unlike standard NMR logs acquired with short inter-echo time spacing TE, EDM data are acquired using relatively long TE’s to deliberately accentuate diffusion effects. The fundamental concept in EDM is that diffusion establishes an absolute upper bound for the apparent T2 for water that is accurately predictable from formation temperature and usercontrolled tool characteristics (TE and magnetic field gradient). Signal with apparent T2’s greater than this limiting value for water therefore provides a direct and unambiguous indication that oil is present in the sensitive volume probed by the logging tool. EDM is sensitive only to intermediate viscosity oils in the range of approximately 1 to 50 cP. Thus, EDM is complementary to the established direct NMR hydrocarbon-typing applications that are designed for light hydrocarbons, such as the Differential Spectrum Method. Besides expanding the viscosity range of NMR based hydrocarbon-typing applications, EDM can be potentially useful in the evaluation of carbonate reservoirs, and in the determination of residual oil saturation (ROS). Processing and interpretation of EDM data in the T2 domain are straightforward and can be performed at the well site. Although interpretation in the T2 domain is suitable for qualitative applications, quantitative applications require special data acquisition and processing methods that are currently under development. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER HH METHODS FOR COMPUTING SWI AND BVI FROM NMR LOGS S. Chen Western Atlas Logging Services, Houston, Texas, USA R. Arro and C. Minetto Comodoro Rivadavia, Argentina D. Georgi and C. Liu Western Atlas Logging Services, Houston, Texas, USA ABSTRACT Irreducible water saturation (Swi) and bulk volume irreducible (BVI) water from nuclear magnetic resonance logging are not directly measured quantities but are derived from the T2 distribution and the effective porosity. Thus, they are dependent on models and the associated parameter used in the interpretation of the T2 distribution data. It has been standard practice to use a T2 cutoff value to partition the T2 spectrum into irreducible and moveable fluids. This assumes that small pores are filled with irreducible water and that large pores contain moveable fluids (either hydrocarbons or water). Such an approach brings forth arguments both from scientific considerations and from the practical applications point of view. Scientifically, it is also possible that pores are incompletely drained; a film model may be more suitable for describing BIKE. In practice, a sharp T2 cutoff may result in very small or “zero” BVI, if either the T2 cutoff value or the estimated T2 distribution is inaccurate. Such a phenomenon has been observed on logs from the Gulf of San Jorge Basin, Argentina and is known also to occur in formations elsewhere. We investigated the relationship between T2 cutoff and the film model and, for simple pore geometric models, derived transcendental equations for predicting film model BVI weighting functions based on T2 cutoff values. We found that the BVI weight functions are not very pore geometry sensitive and based on that, a procedure to compute a generic BYE weighting function is derived. The method is illustrated with core samples from the Gulf of San Jorge Basin and has been applied routinely since 1995 to several hundred NMR well logs. In addition, we used a second approach to estimate BVI weighting functions by forming the ratio of individual incremental porosity bins. of the 100% saturated and desaturated core NMR T2 distributions. This approach appears more reasonable for cases when the short T2 bin porosities in the desaturated T2 distribution exceed the corresponding bin porosities in T2 distribution of the fully saturated data. Both approaches work well with San Jorge Basin data and are easy to use. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER II COMBINING NMR AND DENSITY LOGS FOR PETROPHYSICAL ANALYSIS IN GAS-BEARING FORMATIONS R. Freedman, Chanh Cao Minh and Greg Gubelin Schlumberger Oilfield Services J.J. Freeman Shell E&P Technology Company Thai McGinness and Bob Terry Amoco Production Company David Rawlence Woodside Offshore Petroleum ABSTRACT A new well logging method for evaluating gas-bearing reservoirs has been developed. The method combines total porosity from the CMR Combinable Magnetic Resonance tool (TCMR) and density log-derived porosity (DPHI). It is based on new gas equations derived recently by Freedman (1997) and will be referred to as the Density—Magnetic Resonance (DMR) method. The equations and the method are also applicable to reservoirs with gas condensate or light oil near the wellbore. The method provides new petrophysical equations for (1) gas-corrected total formation porosity and (2) flushed-zone gas saturation. This paper describes the method in detail and applies it to the evaluation of field data. The DMR gas-corrected total porosity (DMRP) is a new formation evaluation parameter. DMRP from the new method can be used in volumetric calculations to provide more accurate reservoir volume estimates than previously possible. Also, more accurate formation gas saturations can be computed when using gas-corrected total porosity in conjunction with deepreading resistivity tools. The improved gas saturations and reservoir volumes provide better estimates of gas reserves. Gas-corrected total porosity can also be used in conjunction with the Coates-Timur equation to provide better permeability estimates in gas-bearing zones. Attractive features of the DMR method include (1) faster logging in many environments because the gas polarization can be minimized, (2) robust gas evaluation because the separation in porosity is accentuated by the opposite effect of gas on the DPHI and NMR logs, (3) total porosity corrected for gas effect and (4) simple interpretation analogous to the familiar neutron-density gas detection. The equations for gas-corrected total porosity and flushed-zone gas saturation are derived from the petrophysical response equations for total NMR porosity and formation bulk density. In gas-bearing reservoirs, gas-corrected total porosity is shown to obey a simple approximate equation that can be used to make a semi-quantitative estimate of DMRP by visual inspection of DPHI and TCMR logs. The effects that uncertainties in input parameters have on the outputs of the gas equations are studied using equations derived in Appendix A. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 Numerical examples using synthetic data and field data are used to demonstrate the relative insensitivity of the gas equation outputs to uncertainties in the inputs. The method is applied to field logs from three commercial gas and oil wells. In the first field example the gas effect on the neutron log is suppressed by thermal neutron absorbers and the neutron-density logs fail to show gas in a gas-bearing zone. The large separation between DPHI and TCMR identifies the zone as gas bearing. In the second field example, gascorrected total porosity logs are compared to neutron-density logs and to porosity measurements on conventional core. Logs of gas-corrected total porosity including the uncertainties SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER JJ A METHOD FOR INVERTING NMR DATA SETS WITH DIFFERENT SIGNAL TO NOISE RATIOS K-J Dunn Chevron Petroleum Technology Co., La Habra, CA, USA D. J. Bergman Tel Aviv University, Tel Aviv, Israel G. A. LaTorraca, S M. Stouard, and M. B. Crowe Chevron Petroleum Technology Co., La Habra, CA, USA ABSTRACT The inversion of NMR data set of a T2 echo train to obtain a T2 amplitude distribution often assumes a constant measurement error for each echo for the whole echo train. However, when the NMR logging tool provides two sets of echo train data, one with a high signal to noise ratio, e.g., short echo trains with a very short recovery time; and another one with a low signal to noise ratio, e.g., long echo trains with an ample recovery time; the measurement error for the two data sets are very different and the inversion of the NMR data with due consideration to both data sets presents a problem. The current approach for inverting the two NMR data sets with drastically different signal to noise ratios uses a splicing technique. The data sets with high and low signal to noise ratios are processed separately with different sets of T2 relaxation times. Then, the long T2 components for the short echo trains are discarded due to incomplete recovery. The remaining short T2 components for the clay-bound water are concatenated with the T2 distribution for long echo trains to form the total porosity T2 distribution. This splicing technique is computationally fast, simple to implement, and may be appropriate for small numbers of T2 relaxation times. However, it can result in discontinuous T2 distributions when a large number of T2 relaxation times (which allow for a smooth T2 display) are used. The discontinuity can occur where the two distributions are spliced. To resolve the discontinuity problem and ensure that no information is lost, we present a new method which accounts for both high and low signal to noise ratio data and processes them simultaneously. The resulting T2 distribution is not only smooth throughout, it also reveals subtle features in the short relaxation times which are not observed in the splicing technique. Synthetic, core and log examples are presented to delineate the differences. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER KK JOINT INTERPRETATION OF FORMATION PERMEABILITY FROM WIRELINE ACOUSTIC, NMR, AND IMAGE LOG DATA X.M. Tang, M. Altunbay, and D. Shorey Western Atlas Logging Services, Houston, TX, USA ABSTRACT Formation permeability is the key to reservoir characterization and management. Wireline acoustic logging and nuclear magnetic resonance (NMR) logging can now provide continuous permeability profiles. This paper compares the permeability profiles obtained from the two fundamentally different measurements. The two profiles exhibit a remarkable correspondence for various data sets around the world. They also show significant differences in some situations. We analyzed the differences based on the measurement principles of the two methods. It was found that the differences are often associated with the presence of gas, fractures, vugs, and hard mudcake. We demonstrate these examples using data from various formations with gas saturation, fractures, vugs, stiff mudcake, and carbonate scenarios, etc. Image data are also used to aid the analyses and interpretation. These examples show that jointly interpreting the acoustic, NMR and image data provides not only valid and reliable formation permeability profiles, but also an effective means for formation characterization. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER LL CHARACTERISATION OF THE ATHEL SILICILYTE SOURCE ROCK/RESERVOIR: PETROPHYSICS MEETS GEOCHEMISTRY Jean-Louis Alixant, Neil Frewin, Peter Nederlof, and Nashwa Al Ruwehy Petroleum Development Oman ABSTRACT The log evaluation of an expelling source rock is notoriously challenging because the kerogen and the generated oil coexist in the source rock. The Athel silicilyte is a unique geological setting found in the South Oman Salt Basin where the issue of source rock evaluation is directly related to that of reservoir characterisation. With no secondary migration path available, this sealed intra-salt source rock matured into a large reservoir, with micro-crystalline porosities up to 30 pu and oil saturations in excess of 80 su. In the absence of known analogues, well Al Noor-2 was fully cored, and extensive core analyses provided the basis for reservoir description. A simple density-porosity model was calibrated against the core data. The foundation of this model was recently challenged, when the analysis of cores taken in shallower Athel prospects revealed that the organic matter comprised in the matrix is highly soluble in organic solvents. There was significant concern that the Al Noor-2 cores may have been affected in a similar manner, albeit to a lesser degree. Petrophysics and geochemistry teamed up to resolve the challenge. A detailed analytical programme confirmed the initial observation, and established that the shallower, less mature Athel prospects contain viscous bitumens rather than kerogen. The organic matter comprised in the deep Athel silicilyte is mature, and is Soluble in organic solvents: it is a kerogen. TOC (Total Organic Carbon) measurements on cuttings showed that the organic matter content of Athel silicilyte varies with depth in a given well. The resulting grain density changes should be accounted for in the interpretation of the density log. Traditional wirelinebased TOC evaluation methods developed for the evaluation of source rocks do not readily apply, and have limited accuracy. Furthermore, knowledge of TOC alone is insufficient to determine the grain density reliably, due to the occurrence of additional minerals, including pyrite. A more general petrophysical model, honouring all available data simultaneously, is being calibrated to provide the detail required for reservoir characterisation. Measurement procedures, processing methods, and evaluation techniques have had to be revised. The integration of disciplines and the introduction of new measurements provided the key to bracket the organic matter fraction and to reduce evaluation uncertainties to an acceptable level. Preliminary results obtained from NMR (Nuclear Magnetic Resonance) in the laboratory and in the field are promising and indicate the potential of the technique to unravel the unusual porosity of the Athel silicilyte, with possible further application to the characterisation of organic-rich reservoirs and source-rocks. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER MM WELL LOG EVALUATION OF GAS HYDRATE SATURATIONS Timothy S. Collett US Geological Survey, Denver, CO ABSTRACT The amount of gas sequestered in gas hydrates is probably enormous, but estimates are highly speculative due to the lack of previous quantitative studies. Gas volumes that may be attributed to a gas hydrate accumulation within a given geologic setting are dependent on a number of reservoir parameters; one of which, gas-hydrate saturation, can be assessed with data obtained from downhole well logging devices. The primary objective of this study was to develop quantitative well-log evaluation techniques which will permit the calculation of gas-hydrate saturations in gas-hydrate-bearing sedimentary units. The “standard” and “quick look” Archie relations (resistivity log data) yielded accurate gashydrate and free-gas saturations within all of the gas hydrate accumulations assessed in the field verification phase of the study. Compressional wave acoustic log data have been used along with the Timur, modified Wood, and the Lee weighted average acoustic equations to calculate accurate gas-hydrate saturations in all of the gas hydrate accumulations assessed in this study. The well log derived gas-hydrate saturations calculated in the field verification phase of this study, which range from as low as 2% to as high as 97%, confirm that gas hydrates represent a potentially important source of natural gas. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER NN USING CONSONANT-MEASUREMENT SENSORS FOR A MORE ACCURATE LOG INTERPRETATION P.A. Casu AGIP S.p.A M. Andreani & W. Klopf Schlumberger Wireline & Testing ABSTRACT The highly layered nature and complex lithology of many fields make formation interpretation difficult, especially for evaluating permeability and fluid mobility with logs. The problem is compounded when log interpretation is based on logging measurements that investigate different volumes of rock. The introduction of new tools that measure colocational rock volumes such as nuclear magnetic resonance, dielectric propagation time and attenuation, microresistivity, litho-density and epithermal pulsed neutron improves log interpretation significantly. This paper reviews an interpretation methodology that associates these measurements with R~ logs and shows how they can be used in siliclastic formations of complex lithology to provide more accurate determinations of permeability index, silt volume, water-cut prediction, and fluid mobility and volumes. The accuracy of the results, however, depends on a close relationship between the measurements. For this reason the concept of consonance between volumes investigated by logs is defined and analyzed. Logging tools take readings in different ways: some measure a cylindrical rock section centered on the bore-hole and others measure a ring of rock isolated from the borehole but still influenced by bore-hole effects. A third group of sensors investigates a segment into the rock of up to 45 degrees. Because most rocks are heterogeneous, to obtain permeability and characterize rock reservoir potential it is essential to associate logs that measure in the same way. In this paper, the measurements selected are from consonant sensors that investigate co-locational rock volumes of the segment form. This paper addresses the following interpretation issues: determination of a reliable silt index in complex lithology, evaluation of thin gas-bearing beds, log interpretation in oilbase mud, values of the in and n parameters in the Archie formula, radioactive sands and control of lithology effects on magnetic resonance measurements. The difficult case of finding the silt index of formations of complex lithology is reviewed in detail. The coherent measurements studied are highly complementary and provide accurate estimates of the potential reserves in place. The petrophysical model, which includes all fluids in invaded zone, is reliable for the characterization of the reservoir rock in any mud environment. Furthermore, the assessment of the total hydrocarbon reserves depends not only on accurate R~ values but also on the way they were entered in the model. The examples are supported by open-hole test results and core comparisons. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER OO DETERMINATION OF ROCK TYPES FROM PORE THROAT RADIUS AND BULK VOLUME WATER, AND THEIR RELATIONS TO LITHOFACIES, CARITO NORTE FIELD, EASTERN VENEZUELA BASIN Juan Carlos Porras Petróleos de Venezuela S.A., Exploración y Produccion Puerto La Cruz, Venezuela ABSTRACT The Canto Norte Field is located in the North of Monagas trend in the Eastern Venezuela Basin (Figure 1). Reservoirs in this trend are characterized by high initial temperature and pressure, and high initial production rates. A tar mat is present at the base of the oil column, acting as a barrier between the aquifers below and the oil-containing formations above. Water saturation, all of which is immobile, is dependent on pore geometry, with pore throat being the dominant control on the flow characteristics of the reservoirs. The pore throat radius on a capillary pressure curve where the non-wetting fluid saturates 35% of the porosity, R35, is a good measure of the largest interconnected pore throats in a rock with intergranular porosity. Pore throat radii, estimated from mercury injection capillary pressure data, were related to several reservoir responses, such as permeability, porosity, and irreducible water saturation, and were used to determine rock types, which range from mega to nanno porous, and have different flow capacities. Lines of equal bulk volume water, which is the product of porosity and water saturation, were used to divide rock types into sub-types, according to theft storage capacity. Rock types were then compared to lithofacies determined from sedimentological analysis, resulting in an excellent correlation. A capillary pressure profile or pore throat type curve was determined for each rock type. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER PP FACIES PREDICTION FROM CORE AND LOG DATA USING ARTIFICIAL NEURAL NETWORK TECHNOLOGY Loveena Kapur, Larry W. Lake, Kamy Sepehrnoori The University of Texas at Austin Dave C. Herrick, Cindy T. Kalkomey Mobil Exploration and Producing Technical Center ABSTRACT Facies classification is widely used to divide well data to obtain, for example, meaningful porosity-permeability relationships. This classification is usually done visually on cores and then extended to wireline logs from the cored wells. The challenge of this problem is to apply this classification to uncored wells based on the relationships observed at the cored wells. This paper presents a novel method to predict facies based on artificial neural network (ANN) techniques. We use a back-propagation ANN algorithm for recognizing the patterns of different facies. The ANN is trained on each facies of cored wells based on gamma ray, density, neutron, and resistivity logs. The facies selected for training the ANN are turbidites, debris flow, shoreface, and lower shoreface. The accuracy of facies predicted from logs alone using the ANN ranges from 75% to 93%. Gamma ray and density logs are the most crucial for some types of facies while neutron porosity log are more important for others. The approach of this work can be applied to fields where quantitative classification of a large number of logs by visual observation can be time-consuming and tedious. This approach can also be used to determine which logs are the most crucial for determining different types of facies. This can provide insights into future data collection. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER QQ POROSITY ESTIMATION FROM P AND S SONIC LOG DATA USING A SEMI-EMPIRICAL VELOCITY-POROSITY-CLAY MODEL I. Goldberg and B. Gurevich The Geophysical Institute of Israel, Holon, Israel ABSTRACT Porosity estimation from compressional and shear acoustic logs is an important component of formation evaluation, especially for wells without other sources of porosity (e.g., water wells). A similar problem arises in the petrophysical interpretation of P and S velocities obtained from seismic data. We propose an algorithm for porosity estimation from P and S velocities using a send-empirical model that relates elastic wave velocities to porosity and clay content. The proposed model extends Krief’s velocity-porosity model to account for the effect of clay content on elastic wave velocities. The Krief’s model assumes that P and S velocities in a porous fluid-saturated rock obey Gassmann formulas with the Biot compliance coefficient. To use this model for clay-rich rocks we assume that bulk and shear moduli of the grain material, and Krief’s exponent A are functions of the clay content C. The moduli of the matrix grain material are defined by the low Hashin - Shtrikman bounds. The constants of the model are obtained by a calibration procedure, consisting of a multivariate nonlinear regression fit for P and S velocities as functions of porosity and clay content using data with known petrophysical properties. The proposed model (1) is consistent with the Biot - Gassmann theory of poroelasticity thereby providing an explicit continuous and smooth dependency of the P and S velocities on the pore fluid properties in the wide range of porosity and clay content (making it readily suited for inversion purposes); (2) provides realistic limits for predicted velocities in all limiting cases (Wood’s formula for high porosities, known properties of shales for rocks with high clay content); (3) allows for simple and natural tailoring to a particular environment through multivariate nonlinear regression. The forward model outlined above may be inverted for reliable porosity estimation from P and S velocities if the clay content data (e.g., from Gamma-ray log) are available. The porosity estimation algorithm was tested on a log data set from West Dixon #1 well, Western Australia (courtesy of Wiltshire Geological Services). The calibration was performed using Neutron-Density (ND) porosity and clay content from Gamma-ray log over the interval 41554420 m, where dipole sonic log was run. The constants obtained from the calibration were fitted into the inversion algorithm, which was then run in order to estimate porosity from the measured velocities and clay content. The results agree reasonably well with ND porosity. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER RR CAN NMR POROSITY REPLACE CONVENTIONAL POROSITY IN FORMATION EVALUATION? Stefan Menger and Manfred Prammer NUMAR ABSTRACT Since 1991, commercial NMR logging has significantly changed the way formation evaluation is done. The capability to reliably measure total porosity in combination with fluid typing has let NMR become an important cornerstone in formation evaluation. A brief review discussing the potential and limits of current NMR technology is followed by a proposal to standardize the nomenclature. NMR well log data is then compared with conventional logs and core results. The full description of these comparisons show that NMR does have the potential to replace conventional porosity. In the near future, further research and development will increase the impact of NMR technology on formation evaluation. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER SS FORWARD MODELS FOR NUCLEAR MAGNETIC RESONANCE IN CARBONATE ROCKS T. S. Ramakrishnan, L. M. Schwartz, E. J. Fordham, W. E. Kenyon, and D. J. Wilkinson Schlumberger ABSTRACT In the conventional approach to the interpretation of nuclear magnetic resonance (NMR) measurements on water-saturated reservoir rocks, it is assumed that the T2 distribution and the pore size distribution are directly related. However, both laboratory and log data show that this relationship breaks down in many pore systems, especially carbonates, which consist of micro (intragranular) and macro (intergranular) porosity. This breakdown limits our ability to predict permeability and movable fluid fractions. Physically, it is due to the diffusion of magnetization between the intra- and intergranular pores. We present here three geometrical models that help clarify the relationship between NMR measurements and the underlying pore geometry. All of the models characterize this geometry in terms of four parameters: (1) (2) (3) (4) the volume fraction of total porosity, φ the volume fraction of intergranular porosity, fm the pore volume to surface area ratio for the micropores, Vsµ the pore volume to surface area ratio for the macropores, Vsm In the first model, we apply random walk numerical simulations to an ordered cubic packing of consolidated microporous grains. For given values of the above parameters, the T2 distribution is evaluated as a function of surface relaxation parameter, ρ. In the second and third models, the microporous grains are treated as a continuum. For the ρ values of greatest interest, roughly, 1.50 → 7.50 µm/s, essentially identical results can be derived from a threedimensional (3D) analytical model. In addition, for all values of ρ, many features of the T2 distribution can be represented in terms of a one-dimensional (1D) model pore space. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER TT NMR FORMATION EVALUATION APPLICATIONS IN A COMPLEX LOW-RESISTIVITY HYDROCARBON RESERVOIR Carsten Slot-Petersen Statoil Exploration and Production Terje Eidesmo Statoil Research Centre Jim White Schlumberger Wireline and Testing Hákou G. Rueslâtten Statoil Research Centre ABSTRACT: NMR log data was acquired on a well in a complex lithology reservoir in the North Sea. The entire gross reservoir interval had been cored, using low invasion core bit and tracer doped synthetic oil based mud for the purpose of providing core NMR and core water saturation data for log calibration. The reservoir is characterized by low-resistivity pay zones, high matrix densities and standard log and core porosities of about 35 PU. The reservoir sands are fine to very fine grained, well sorted with abundant iron-rich glauconite and chlorite. The microporous glauconite reduces the reservoir effective porosity. The main objectives of the NIMR log were to help identify mobile oil, quantify irreducible water saturation, effective porosity and bound fluid volume. Most of these objectives were achieved but during the course of the interpretation a number of other useful applications of NMR technology came to light. The demonstrated field specific applications related to the subject well include; 1) detection of tar within the hydrocarbon zone reflected by a NMR “signature’ of missing porosity and shortened 12 decay; 2) determination of mobile fluid type through comparison of core calibrated NMR bound fluid volumes with water volumes computed from resistivity tools; 3) identification of glauconite and chlorite rich sands; 4) estimation of permeability through core calibrated NMR data; and 5) identification of gas by the density NMR indicator. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER UU BOUND WATER VOLUME, PERMEABILITY, AND RESIDUAL OIL SATURATION FROM INCOMPLETE MAGNETIC RESONANCE LOGGING DATA C. Flaum and R.L. Kleinberg Schlumberger-Doll Research J. Bedford Schlumberger Wireline & Testing ABSTRACT Magnetic resonance (MR) is a cyclic measurement consisting of a wait time followed by an echo acquisition period. Conventional MR data are presently acquired with wait times sufficiently long to substantially polarize all fluid protons in the formation. Protons in gas, light oil, oil base mud filtrates, and vug water polarize very slowly. Therefore very long wait times have been used when logging formations in which those fluids are present. The measurement can be made faster if the wait time is reduced. When the wait time is a few tenths of a second, the signals from fluids that are slow to polarize are suppressed. Then MR. logging is no longer useful for the determination of porosity, but the incomplete MR data still deliver capillary and clay bound water volumes. When these MR-unique outputs are combined with widely accepted and efficient measurements of resistivity and porosity, a complete suite of log outputs can be reliably obtained. MR data has been obtained in the North Sea with significantly improved logging speed, vertical resolution and/or data precision. We show that each of these specifications can be improved without degradation of the others. Heretofore it has been considered difficult or impossible to estimate residual oil saturation in wells drilled with oil base mud. This is particularly true when using a magnetic resonance T2 measurement to distinguish mud filtrate from low viscosity native oil. These fluids have little or no T2 contrast, but can have large T1 contrast. Using reduced wait time logging we have found significant amplitude contrast between fluids with different T1 relaxation times. We have used this principle in the North Sea to estimate the proportions of oil base mud filtrate and native formation oil in the flushed zone, from data obtained in one fast logging pass. Where deep resistivity measurements show a long transition zone, shallow NMR measurements find a sharp residual oil contact because there is no capillary pressure difference between OEM filtrate and native oil. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER VV A CASE STUDY DEMONSTRATING HOW NMR LOGGING REDUCES COMPLETION UNCERTAINTIES IN LOW POROSITY, TIGHT GAS SAND RESERVOIRS W. Scott Dodge Exxon Exploration Company Angel G. Guzman-Garcia Exxon Production Research Company Dave A. Noble Exxon Company U.S.A. Jack LaVigne Schlumberger Well Services Ridvan Akkurt NUMAR ABSTRACT Nuclear Magnetic Resonance (NMR) logging in low permeability gas reservoirs has been used to assist standard formation evaluation techniques in identifying productive reservoirs from those that lead to tight tests or produce formation water. By incorporating NMR logging into the standard logging suite, improved completion decisions are made regarding perforation intervals, hydraulic fracture program design and accurate estimates of producible gas. The deep gas reservoirs of the Vicksburg trend in this study contain complex clastic mineralogy derived from igneous rocks. Transport, deposition, and diagenesis play an important role in the producing characteristics of these reservoirs, Burial and diagenesis lead to low-porosity reservoirs with permeability in the range of 0.01 to 1 mD. Diagenesis of lithic rock fragments and feldspars creates significant quantities of micro-porosity, which degrades reservoir quality. The micro-porous rock holds large amounts of non-producible formation water, yet shows up as high water saturation in standard log estimates, Therefore, when conventional logging estimates of porosity and water saturation are used, it is not clear which reservoirs will produce gas free of formation water or not produce at all because of low permeability. NMR technology provides additional information on irreducible water-filled porosity and quantitative reservoir permeability not available from standard logging tools. In cases where the wells are drilled with oil-based mud and formation water resistivity is not known accurately, NMR reduces the risk of completing zones, which produce water while identifying tight gas zones, by the absence of oil-based mud filtrate in the flushed zone. When NMR measurements are combined with log-derived measurements of porosity and water saturation, both producible porosity and permeability thickness for these reservoir sands can be quantified. This paper is a case study showing the benefits of NMR logging and core analysis in low porosity, gas-bearing sandstones. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER WW PRACTICAL INVERSION OF HIGH-DEFINITION INDUCTION LOGS USING A PRIORI INFORMATION T. Fishburn Unocal Corporation, Sugar Land, Texas, USA I.M. Geldmacher, M. Rabinovich, and L. Tabarovsky Western Atlas Logging Services, Houston, Texas, USA ABSTRACT Model-based, fast 2-D inversion is becoming one of the key tools in petrophysical interpretation of array-type logging data. Several practical examples that show the fast 2-D inversion advantages of the High-Definition Induction Log (HDIL) service are presented. A log analyst can link the HDIL measurements with other geological and petrophysical information by trying various sets of a priori information at the postinversion stage. For example, the shallow-resistivity device measurements can be very helpful in the case of shallow-resistive invasion, Some indicators of permeable formations (for example, the ratio of deep-resistivity measurement to shallow) may reduce inversion nonuniqueness when the contrast between formation resistivity and the resistivity of the invaded zone is low or the radius of the flushed zone is small. The layering information taken from other logging devices can help when the shallow HDIL measurements are not sensitive to formation parameters (conductive mud systems, rugose borehole). Data uncertainties and resolution limits of the HDIL measurements allow us to drive the inversion solution towards a priori known petrophysical features of the formation. The standard technique of introducing a priori constraints complicates the objective function and significantly slows down the inversion. In addition, it is necessary to repeat the entire inversion run if a new set of a priori information must be tested. We have developed a method for incorporating a priori information at a postinversion stage, when new a priori information can be taken into account without actually performing a new run of the inversion algorithm. More accurate reserve estimates may be calculated by incorporating the results of the inversion into volumetric analysis and reservoir thickness computations. Induction measurements of deep resistivity can be suppressed by nearby conductive beds or even erroneously elevated in laminated sections. Examples are presented of volumetric analyses and pay/sand counts using both conventional curve data and “squared” inversion results. Oil in place calculations and the gross value of those reserves are also compared to demonstrate potential improvements in accuracy using these types of measurements. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER XX A PETROPHYSICS-.BASED RESOLUTION-ENHANCEMENT TECHNIQUE FOR ARRAY-TYPE INDUCTION LOGS J. Xiao, D. Beard, B. Corley, and I. M.Geldmacher Western Atlas Logging Services, Houston, Texas, USA ABSTRACT The physics of basic well-logging measurements control the fundamental vertical resolution of a well log. Logs with a greater depth of investigation have coarser vertical resolution than those of smaller depths of investigation when measured with the same physical principles. For log interpretation purposes, however, it is often convenient to have all logs at the same vertical resolution. Array-type induction devices, such as the High-Definition Induction Logging (HDIL) tool, are well suited for providing “resolution-matched” logs. The data measured by short- and long-spaced arrays are combined to generate logs of varying depths of investigation, but with matched vertical resolution. Current resolution-matching techniques used on array-type induction logs superimpose the conductivity variations of shallow-investigating curves onto the deep-investigating curves in an additive manner. The additive method gives correct resolution-matched curves when the conductivity variations far from the borehole are the same as those near the borehole. This condition only holds when there is no invasion. In practice, this technique also gives good results when there is resistive invasion; when there is conductive invasion, however, anomalous results are possible. The HDIL system uses a two-step focusing approach. In the first step, curves at each depth of investigation are generated with their natural resolution. These “true-resolution” curves can therefore be made free of near-borehole sensitivity and related artifacts. The second step is the vertical-resolution-matching process. Because the resolution-matching process is separate, it can be based on assumptions other than the conventional “no invasion” assumption described above. The “no invasion” assumption, which could also be called the “radially constant fluid” assumption, can be replaced by a “slow vertically varying fluid” assumption. The “slow vertically varying fluid” assumption is more reasonable from a petrophysical point-of-view and is more likely to be true, even when significant invasion is present. Synthetic data testing and practical applications illustrate the effectiveness and reliability of the new induction logging data processing technique. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER YY CASE STUDIES OF NUCLEAR MAGNETIC RESONANCE TEXAS AND LOUISIANA GULF COAST AREA Roy Guest, Marcel Di Giovanni, Stacy Smith, Otis Walter Texaco Exploration and Production Inc. Jon Musselman, Tom Pickens, Steve Crary Schlumberger Oilfield Services ABSTRACT Three case studies are presented illustrating how nuclear magnetic resonance (NMR) techniques can supplement traditional logging measurements to evaluate diverse reservoir types. These three cases studies from the Louisiana offshore and East Texas areas demonstrate how NMR data can be integrated with other information to provide costeffective formation analysis, resulting in successful identification and enhanced production of hydrocarbons in a variety of reservoirs. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER ZZ CASED-HOLE LITHOLOGY AND DENSITY MEASUREMENTS, PART I: APPLICATIONS IN PERMIAN BASIN RESERVOIR ANALYSES R. C. Odom Computalog Research Inc. G. P. Hogan II, C. B. Rogers Computalog Wireline Services R. L. Smith Applied Petrophysics M. A. Sirgo III MGB Energy Ltd. ABSTRACT Often, traditional cased-reservoir analyses are not comprehensive enough to be applicable to the mixed lithologies common to Permian Basin reservoirs. The accuracy of the two casedhole saturation measurements (SIGMA and Carbon/Oxygen logging) requires a porosity that is corrected for lithology effects. The saturation calculations also require inputs for the effects of the rock matrix. Along with correction and correlation of logging measurements, the lithology determination can be important in identifying reservoir characteristics, such as sand or anhydrite content, that are often linked to producibility or permeability. Two new measurements are being developed for the Computalog Pulsed Neutron System PND-S: a density-based porosity from inelastic scattering and an inferred photoelectric (FE) factor based on neutron-induced spectroscopy data (described in Part II of this paper). The combination of these new measurements with existing pulsed-neutron technologies is used to develop a more comprehensive reservoir analysis model. Using the cased-hole density with the neutron porosity can resolve the ambiguities posed by gas-filled porosity and changing rock matrix. Through consideration of the porosities, lithology, and saturation measurements, the more complete measurement set resolves many of the inherently under-determined situations common to cased-reservoir analyses. Applications of these new measurements are demonstrated in several reservoir analysis examples from the Permian Basin. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER AAA RECOGNITION OF HIGH PERMEABLE FACIES USING STATISTICAL LITHOFACIES CHARACTERIZATION OF BOREHOLE ACOUSTIC IMAGES IN NORTH MONAGAS FIELD, VENEZUELA A. Sinha , B. Gomez and G. Gomez Geoscience, Western Atlas Logging Services, Venezuela M. Giusto and E. Solorzano PDVSA , Maturin, Venezuela. ABSTRACT The main producing formations, Naricual and Jabillos, of Northern Monagas Field, contain thick sands of varying grain sizes. The production capabilities of these sands are mainly controlled by permeability. In the absence of a reliable continuous permeability measurement, identification of zones for completion or for injection has been a big challenge for the petrophysicist in this area. Borehole images are commonly used for detailed structural and stratigraphic analysis of the reservoir. But the image also contains valuable information in regard to lithofacies. Using statistical methods, several numerical image descriptor curves each describing composition, texture and the fabric of the penetrated formation are calculated which forms the backbone of this technique. These numerical descriptor curves are integrated with core information and other open hole logs to recognize facies. The method has been successfully applied to identify the permeable zones in uncored intervals of the well. Based on the core derived permeability values, the sands are divided into several classes (facies), each facie having a range of permeability values. A data base is built integrating these permeability classes, selected image descriptors and available other open hole logs. The data base is consulted to recognize different permeability facies for the uncored section of wells. This technology has helped in identifying zones with high permeability facie for completion or injection and thus increase production. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER BBB POROSITY DETERMINATION IN HORIZONTAL WELLS S. Ca1vert, M. Lovell, P. Harvey Department of Geology, Leicester University J. R. Samworth Wireline Technologies Ltd J. Hook Enterprise Oil Plc ABSTRACT Experience demonstrates that there is a need to assess whether or not logging tools are providing acceptable porosity values in horizontal wells. The economic viability of a field crucially depends on accurate porosity determination. Porosity can be derived from a number of different wireline and LWD tools; neutron, density, sonic, NMR and resistivity. Porosity estimates together with core measurements that provide important constraints for the range of porosity estimations. There are a number of unusual effects that have been seen in horizontal well porosity log data when compared with vertical wells through the same formation, typically resulting in increased porosity values from some of the porosity logs (notably density). A number of reasons can be proposed, such as: permeability anisotropy leading to irregular invasion and variable water saturation above and below the borehole, differential stress, micro-fractures and disturbed tool placement. Often the porosity values derived from the well data are applied field wide, this is regardless of whether the well location represents a ‘reasonable’ sample of the reservoir parameters or not. The well data are the best information available that can shed light on these quantities. Hence, there is a need to assess which logs provide acceptable porosity estimates. Horizontal wells allow a greater volume of the field for investigation than vertical wells and therefore represent a statistically ‘better’ sampled population of the field parameters. In the case study examined, examples of vertical well and horizontal side track data from the North Sea demonstrated inconsistencies in porosity values. The vertical well wireline derived porosity values and LWD neutron are less than the core porosity. But, in the horizontal well, the LWD maximum density derived porosity and neutron porosity values are lower than expected from the vertical well values. This suggests that through the horizontal section either the formation porosity is lower; increased gas effects are observed or borehole conditions are degrading the quality of the log values. For the dataset examined, the LWD maximum density derived porosity provided the best porosity estimator in the horizontal well and the borehole conditions are the most likely cause of the reduced porosity estimate. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 The values have been validated, calibrated and compared to assert the fundamental causes of the inconsistencies. The integrated data sets have then been used to formulate algorithms and approaches to interpretation by using a range of numerical methods to constrain the derived porosity values. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER CCC AN APPLICATION OF WELL DATA IN OIL AND GAS ASSESSMENT - ARCTIC NATIONAL WILDLIFE REFUGE Philip H. Nelson, Christopher J. Schenk U. S. Geological Survey, Denver, Colorado Kenneth J. Bird U.S. Geological Survey, Menlo Park, California ABSTRACT A current assessment of oil and gas resources in the Arctic National Wildlife Refuge (ANWR) 1002 Area by the U.S. Geological Survey relies upon seismic data, geological mapping of exposures south and west of the assessment area, and exploratory wells. No wells have been drilled within ANWR information is obtained from wells up to 50 1cm west and north of ANWR. Well data are used in this project in two modes: synthesis of all available data and presentation on a well-by-well basis, and extraction of volumetric parameters that were used to assess undiscovered accumulations. Well data assembled for 41 wells include: well logs, core descriptions and measurements, formation tops, biostratigraphic boundaries, drill-stem tests, casing points, fission-track age dates, vitrinite reflectance, and organic-carbon content. Quantitative and descriptive information are plotted at a scale of 500 feet per inch with expanded (100 feet per inch) sections over prospective intervals. Formation tops are inspected and adjusted where necessary. This assemblage of log and core-based data is useful to investigators of different disciplines working on assessment tasks. Distributions of five volumetric parameters are required to estimate hydrocarbon in place for any randomly selected prospect: net reservoir thickness, area of closure, porosity, water saturation, and trap fill. The proportion of reservoir quality rock in the section is computed from the well logs, using criteria suitable for each formation. Area of closure is estimated from seismic data and from geological setting. Porosity is estimated from core and well log data; again using criteria suitable for each formation. Water saturation is estimated from porosity using a constant bulk-volume-water criterion. Trap fill is assigned based upon structural and stratigraphic settings. For each play, a probability distribution is constructed for each of these five parameters; these distributions are then combined using Monte Carlo simulation to obtain statistical estimates of hydrocarbon volume in place, Ultimately, the spreads and median values of the distribution functions, coupled with the risks and number of prospects assigned to each play, determine the uncertainties (spreads) and average values of estimated resources. The synthesis of well data with other geological and geophysical data provides a quantitative foundation for resource estimates of ANWR. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER DDD CROSS-SECTION CONSTRUCTION FROM AUTOMATED WELL LOG CORRELATION: A DYNAMIC PROGRAMMING APPROACH USING MULTIPLE WELL LOGS I. Le Nir Schlumberger Wireine and Testing - RPC – Clamart, France N. Van Gysel Institut de Physique du Globe – Paris, France D. Rossi Schlumberger Wireline and Testing - RPC – Clamart, France ABSTRACT Cross-sections, typically constructed from correlation of a small number of log curves, are a basic tool for developing geologic maps and models of subsurface formations. An efficient dynamic programming algorithm has been developed that uses multi-channel log data to reliably correlate geologic formations between wells, especially in geologic intervals with thinning beds and missing or discontinuous units. Unlike many other methods currently in use, the log data need not be pre-filtered or normalized, reducing data preparation, and retaining familiar measurement values and character that are easily understood when evaluating the results. The only pie-conditioning required is to define a common top and bottom marker in each well. The algorithm is able to use up to 20 log curves per well, weighted according to data quality and geological environment By using an iterative, sequential, method, there is no predefined limit to the number of wells can that be correlated. Automated correlations were carried out on data sets from several oil and gas fields. In the Stratton field (US Gulf Coast), the gamma ray, neutron porosity, density, spontaneous potential, spherically focused laterolog and induction log medium log curves were used from a group of 10 wells in the fluvio-deltaic Frio Sandstone, for which an extensive public data set exists. Subsurface correlations are often complex in this mixed fluvio-deltaic environment because of paleo-shoreline movements that produced rapid lateral and vertical changes in depositional environments, further complicated by erosional unconformities and faulting. Correlation cross-sections produced by the automatic correlation methods are in good agreement with reliable hand-drawn cross sections. Typical depth tie agreement of automatically interpreted correlation surfaces is to within one meter of the published handdrawn sections. The method needs further testing, but these results indicate that by using multiple rather than single logs in an innovative dynamic programming approach, the geologist may more quickly develop a reliable and automatic alternative to hand-drawn correlations between wells from log data. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER EEE GEOSTATISTICAL ANALYSIS OF PETROPHYSICAL VARIABLES INFLUENCING RESERVOIR QUALITY IN THE FRACTURED “O LIMESTONE”, MAPORAL FIELD, WESTERN VENEZUELA Decio Flores PDVSA-E&P Reinaldo Gonzalez Universidad Central de Venezuela ABSTRACT This paper describes the geostatistical analysis applied to both core and log data of the “0 Limestone” to determine the petrophysical variables affecting productivity from this complex, fractured calcareous reservoir in the Maporal Field, Western Venezuela. The production history of the wells was also statistically analyzed in order to perform correlations with the interpreted petrophysical variables. Oil production has been found to be fades and diagenesis controlled, therefore, due to the fact that core data was scarce, a two-step approach was developed. First; core description and their corresponding analysis in three wells were used to characterize the four main flow units of the “0 Limestone”. Then, correlations with log parameters were established in order to extend the classification to wells for which only log data were available. Primary intercrystalline porosity is below 7% in the interval. Total porosity may reach up to 14% through secondary porosity due to dolomitization, mineral dissolution and fractures. Geostatistical correlations and stochastic models were made to ascertain both vertical and areal distributions of petrophysical parameters of the four geological subdivisions. The uppermost interval showed such low porosity and permeability as to be considered nonreservoir rock throughout most of the field. The following two central intervals were found to be the better flow units, while the lowermost one only locally showed producing capability. The applied method should be applicable to other nearby fields in which the “0 Limestone” has been found to be oil producing An optimized log program for new wells was derived from the study. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER FFF NEURAL NETWORK APPLICATIONS TO UPSCALE CORE DATA AND ANCIENT LOGS TO PETROPHYSICAL PARAMETERS OF FLOW UNITS, CARO FIELD, EASTERN VENEZUELA. Alfonso Quaglia F. and Roberto Barbato Q. Petróleos de Venezuela S.A., Exploración y Producción, Puerto La Cruz, Venezuela ABSTRACT A common challenge for optimizing limited log data and other well bore data in mature basins is the need to upscale limited core data to log and reservoir scale. A major handicap to scaling up from older data is the wide range in log type, quality and vintage. In old fields, most of the logs fall in the ancient category, usually with S.PJGR curves, several resistivity curves, cased hole neutron and, if fortunate, some open hole porosity curves. A few new wells may have a relatively complete suite of logs, but limited core data. If cores are avaible, they tend to be from the older wells. Neutral Net (NN) applications are being widely used as pattern recognition tools in the industry. Their application is being successfully used to upgrade and upscale well bore data. The Caro field, one of several along the Carisito structural trend in the Eastern basin of Venezuela (Fig. I and 2) is used as a pilot area to test the NN applications. The objective of the project was two fold; 1.- Teach the NN “brain” to upgrade older, limited log data to values of porosity, water saturation (Sw) and air permeability (Ka), within acceptable error bar limits. 2.- Upscale this new information to net pay of flow units by rock type for each sandstone in the field, in order to relate too and predict performance behavior (decline curves). A sequential process was developed and successfully applied to accomplish better than expected results. The field, one of four along the trend, produces from 18 wells, which have recovered 21 mm barrels of oil from 22 sandstone containers. The potential to identify and recover significant additional oil is based on performance data, validated with the volumetric and recovery models developed using the NN output. Six wells were selected to test the full capacity of the NN application; 1.- Key training well (1). Modem logs with selected core data and productions tests. 2.- Test well (2). The next well chosen was used to test the brain developed from the key well, and to further train the brain with a more limited data set 3.- Evaluation wells (A.B.C.D). The ultimate test is using the ‘brain” to process the older data sets. The results developed with the brain occurred within very acceptable error bar ranges for each parameter tested, from each data set processed. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER GGG TRUE PAY THICKNESS DETERMINATION OF LAMINATED SAND AND SHALE SEQUENCES USING BOREHOLE RESISTIVITY IMAGE LOGS Ray R. Reid and Milton B. Enderlin Phillips Petroleum Company ABSTRACT Once an exploration well has been logged, economic evaluation begins. True pay thickness from logs is an important input to that economic evaluation. Knowledge of the true pay thickness is of particular importance in sands and shales laminated at a scale below the resolution of the standard logging suit. Resistivity image logs provide information about the spatial distribution of shallow resistivity. Since resistivity is a function of both the rocks and included fluids, resistivity images can provide information about the spatial nature of the rocks and fluids. With proper processing (which includes data transformation from depth to the time domain, correction for tool acceleration, transformation back to the depth domain, and static normalization) the resistivity images can provide a quantitative measure of the shallow resistivity at a resolution of a few centimeters. A resistivity-to-pay sand cutoff operator is selected by the optical application of petrophysical reasoning. The resistivity-topay sand operator determines which sand layers are pay and their apparent thickness. Further processing can resolve the n-tie dip of individual pay sand layers. By combining the local structural dip interpreted from true dip of the pay sand layers with borehole orientation data, the apparent thickness of each pay sand layer can be converted into a true pay sand layer thickness. Summing over all the true pay sand layer thickness yields the true pay thickness. Resistivity images from a Gulf of Mexico exploration well are used to illustrate a processing technique to achieve an understanding of the true pay thickness. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER HHH GET THE “RHYTHM OF THE ROCKS” AND FIND THE DIP Milton B. Enderlin and Tony Kratochvil Phillips Petroleum Company ABSTRACT Once an exploration well is drilled, pre-drill/postdrill reconciliation begins. Often the structural dip interpreted from seismic conflicts with the structural dip interpreted from dipmeter or image logs. An independent measure of dip, “Rhythm of the Rocks”, can help to reconcile this conflict. The “Rhythm of the Rocks” is seen in the oscillatory response of well logging tools. One attribute of the oscillating log response is measured depth wavelength. If the wavelength over a particular succession of rocks is known from a wellbore oriented normal to the rock layers, then an increase in wavelength for the same or correlative succession of rocks penetrated by a non-normal wellbore can be explained by the angular relationship between the wellbore and rock layers. Measurement of normal and non-normal wavelengths can be accomplished graphically or digitally. The trigonometric relationship between the normal and non-normal wavelengths determines the relative angle between the wellbore and a line normal to the rock layers. A plot of the relative angle centered at wellbore deviation and direction on a Stereonet yields true dip magnitude of the rock layers as a function of azimuth. If the wellbore is vertical, only the true dip magnitude is known; dip direction remains a mystery. In deviated wells, the true dip magnitude is known, but now an understanding of the dip direction is possible. The greater the deviation, the greater is the understanding. “Rhythm of the Rocks” is used here to determine true dip magnitude and help substantiate dipmeter results in an example from Garden Banks, Gulf of Mexico. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER III MWD ULTRASONIC CALIPER ADVANCED DETECTION TECHNIQUES Gary Aithoff Sperry-Sun Drilling Services, Houston Abbas Arian, A. Balaji Kavaipatti, Georgios L. Varsamis, Laurence T. Wisniewski SensorWise, Inc., Houston ABSTRACT Several Measurement-While-Drilling (MWD) caliper tools, which utilize acoustic technology, are available to the industry (Moake 93, Orban 91), but they typically have a limited operating range and are based on analog detection techniques. Technical issues which could preclude the use of acoustics for an MWD caliper tool include the noise associated with the downhole drilling environment, tool accuracy, tool reliability, errors associated with mud attenuation and signal detection, tool eccentering, formation cuttings, gas, etc. A new digital standoff sensor has recently been introduced (Maranuk 97), that utilizes full waveform digitization and processing to enhance the accuracy and operating range of the tool. The sensor represents an evolution similar to that of other digital sensors when compared to their previous analog counterparts. The sensor technology includes fill waveform digitization and periodic storage, as well as the capability for signal processing downhole. Stored data can be recovered and further processed at the surface after the bit run, Additionally, the sensor includes a magnetometer and an accelerometer that provide information for the derivation of key characteristics of elliptical boreholes. Significant emphasis was placed in the performance and repeatability of the ultrasonic transducer itself; during the design and testing phase. This paper describes the basic design of a digital ultrasonic caliper for MWD applications with an emphasis on the utilization of advanced detection techniques. The detection techniques employed by the tool are presented, discussed and analyzed through several log examples and some corresponding stored waveforms. The options for “on-the-fly” processing downhole, as well as for surface processing are described and several interesting cases, such as tool performance in soft formations, are illustrated. Through all of the examples the stability and performance of the ultrasonic transducer in a variety of downhole environments is demonstrated. Finally, some conclusions are drawn. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER JJJ GAS SATURATION DETECTION MODEL APPLIED TO HETEROGENEOUS RESERVOIR USING TDT G.M. Hamada College of Engineering, King Saud University, Riyadh, Saudi Arabia S.A. Heikal Suez Oil Company, Cairo, Egypt. ABSTRACT The Zeit Bay field reservoir units consist of sandstone and carbonates, partially overlaying a tilted block of fractured basement reservoir with a complex drive mechanism. A secondary recovery scheme of gas re-injection into the original gas cap was initiated to maintain reservoir energy and minimise pressure decline. Hence accurate detection of gas movement is very critical. Several difficulties to monitor gas-oil contacts were encountered in a considerable number of wells. Some of these difficulties were, gas channelling behind the casing, gas coning, wellbore fluid changes, porosity and lithology changes, wellbore fluid invasion into the reservoir and the presence of formation stimulation fluid (HCl). The application of conventional methods using. the response of gas indicator curves could result in a false indication of formation gas-oil contacts. This paper discusses the approach adopted in order to determine the gas-oil contact in wells where such problems occur. A database was established including more than 70 TDT runs, open hole log and pressure data of 12 infill wells, and production performance records of all Zeit Bay wells. The approach follows the Polyachenko model of functional relationship between count rates and gas saturation. Several cross plots for the same range of porosity and connate water saturation, e.g. formation capture cross section (SIGM), total selected near detector counts (TSCN), total selected far detector counts (TSCF), the capture cross section of the bore hole (SIBH), and inelastic far detector counts (INFD). Each crossplot gives a definite diagnostic shape around the depth of the formation gas-oil contact. By using these crossplots it would be possible to calculate gas saturation from a stand alone run. The model was validated by RFT and open hole log data from infill wells. Also it was successfully applied in wells which showed an ambiguity in the detected formation gas-oil contact. The field gas-oil contact in Zeit Bay was revised using the results of the model. This revision lead to an accurate definition of the oil leg and to the drilling of three additional wells in the field. The open hole log results of these wells verified the gas-oil contact determined by the model. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER KKK ANALOG AND NUMERICAL MODELING FOR BOREHOLE RESISTIVITY- AND ACOUSTICIMAGING TOOLS G. Yu, S Painchaud, and K.-M. Strack Western Atlas Logging Services, Houston, Texas ABSTRACT Borehole imaging technology plays a key role in formation characterization and hydrocarbon reservoir evaluation. Analog (or physical) modeling is a method for studying the effects of subsurface bodies or structures by comparison with the response of earth models built in the laboratory. Historically, analog modeling has been used to support tool development and data interpretation. The modeling results have been too complex, however, to explain the physics of the imaging tool response. Subsequently, numerical simulation has replaced analog modeling to support tool development and data interpretation. With advanced computer technology, numerical simulation can model and explain the physics of complicated models. Due to the spatial geometry of tool and rock formations, however, even three-dimensional numerical modeling has its limitations. Both analog and numerical modeling methods must be used to understand the relationship between real geology features and image logs. The combination of analog and numerical modeling assists the development of both acoustic and it also resistivity-imaging tools as well as interpretation software and provides an effective way to extract better geologic information. One resistivity pad was used to perform the resistivity image analog modeling. The acoustic section of the imaging tool was used for the acoustic image modeling. Both imaging tools were tested over the rock borehole model using bentonite to simulate mudcake. The mudcake did not influence the resistivity image but smeared small and fine features on the acoustic image. Fractures parallel and perpendicular to the borehole axis were created in the rock model. The acoustic image tool showed higher sensitivity on fractures perpendicular to the borehole axis than on those parallel to the borehole axis. Overall, the resistivity image showed fracture resolution at least twice that-of the acoustic image. The numerical modeling for the electrical sensor focused on the support of tool design and characterization. Numerical modeling has been used to develop environmental corrections in order to obtain a calibrated resistivity (K-factors). SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER LLL CASE STUDIES OF TIME-LAPSE SONIC LOGGING K. Hsu , G. Minerbo, J. Aron, D. Codazzi, V. Ernst, T Lau Schlumberger Oil Field Service, Sugar Land Product Center, Sugar Land, Texas ABSTRACT With the recent availability of sonic logging while drilling (LWD), it is possible to measure compressional & values shortly after the formation is drilled. In this paper, we present two case studies of LWD and wireline measurement comparison with emphasis on the analysis and interpretation of time-lapse sonic logging. For the cases shown, the LWD and wireline sonic logs show good agreement in interbedded sand-shale sequences. However, there are zones, especially in shale, where the two measurements can be different. This is attributed to formation alteration caused by drilling damage, plastic deformation and water take-up by shale. Formation alteration can form a low-velocity annular zone around the borehole that causes wireline sonic tools to read the slower at values whereas the LWD sonic can still read the virgin formation. Data also suggests that the thickness and properties of the altered zone progressively change over time. A reverse trend that the LWD sonic is slower than the wireline sonic is observed in gas sands, which can be explained by the replacement of gas by mud filtrate in the pore space of the rock. Finally, because data shows that formation alteration can happen in a short period of time (i.e., less than a half hour) after the rock is penetrated, a sonic-while-drilling tool with a long transmitter-to-receiver spacing provides a better opportunity to log the virgin formation & than a tool with a shorter spacing. SPWLA 39th Annual Logging Symposium, May 26-29, 1998 PAPER MMM OPTIMIZED SPEED CORRECTION Mark G. Kerzner SHMSoft ABSTRACT A number of papers investigate the irregularities of logging tool movement in the borehole, and point out the needs of correcting for it. The obvious advantage is better depth determination. For logs that involve image processing, speed correction also results in better quality of the resulting picture. The two published approaches are straightforward integration of the accelerometer measurements, and Kalman filtering. These approaches have problems accounting for all limitations. Therefore, they have to use additional cutoffs on the result, raise special flags, adjust filter parameters, etc. The optimized speed correction (OSC, patent pending) takes the opposite approach. It starts with the desired results and finds the mathematical method of arriving to it.. The problem of finding correct depth is formulated as a problem of mathematical optimization. The givens in this problem are: the measured depths, the accelerometer measurements, and the known limitations, such as maximum tool speed and maximum cable stretch. Other data, such as cable tension, can also be included. The desired outcome is the tool’s trajectory as a function of time. The solution has to start at the starting depth and come to the ending depth, and correspond most closely to the measured accelerometer data, while fulfilling all the limitations. Dynamic programming algorithm provides a solution to this optimization problem. The design of the algorithm contains a few key points that reduce the problem to a computationally manageable task. The formulation of the speed correction problem as a problem of mathematical optimization allows the following: 1. Formulate the problem in terms of the desired result known from physics and geological research. 2. Guarantee the quality of the solution. 3. Allow to take into account multiple limiting conditions. 4. Formulate a feasible computation algorithm. The speed button correction is also treated and an approach that uses dynamic programming curve matching is suggested. TRANSACTIONS OFTHE SPWLA FORTIETH ANNUALLOGGINGSYMPOSllLlM §ponsod by THE SOCIETYOF PROFESSIONALWELL LOG ANALYSTS, INC. 8866Gulf Freeway,Suite320 Houston,Texas77017 Presentedat THE HOLMENKOLLENPARRHOTEL RICA Oslo,Norway May 30-June3,1999 NOTICE TO EDITORS: Permissionis hereby granted to publish elsewhereany of these transactionsafter June 3, 1999, provided that conspicuousacknowledgementis given to the original presentationof the paperandthe authorsof the paperhaveagreedto the republication. (The statementsandopinionsexpressedin thesetransactionsare thoseof the authorsand should not be construedasan official actionor opinionof the Societyof ProfessionalWell Log Analysts, Inc.) th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER A INVASION IN SPACE AND TIME M. Peeters Colorado School of Mines Golden Colorado D. Allen Schlumberger Doll Research, Ridgefield Connecticut, it Comes Colorado School of Mines, Golden Colorado / Petrobras Brazil J. I. Kristiansen Norsk Hydro, Bergen Norway ABSTRACT All wireline and logging while drilling tools are to some extend adversely affected by the borehole fluids and invasion. The severity of these effects depends on the depth of investigation of the tools, and at the moment in time that the measurements are made. Logging while drilling tools often measure while the invasion process is still in progress, and wireline logging tools after a steady state is reached. Environmental corrections charts and algorithms for individual tool have been available for a long time, but corrections are usually still based on simplistic invasion profiles, and seldom take the time factor into account The sophistication of modem tools is apparently not matched by improvements in invasion correction programs. This paper reviews the invasion process and discusses the effects of various mud types on most common logging tools. Logging suites are recommended for certain mud types that allow the tog analyst to alleviate invasion and borehole effects. Several field cases are used to illustrate the invasion process. A MWD log with 7 repeat runs was analyzed with forward models, in which the user can speci~’ the invasion parameters interactively. The results prove that the piston invasion model is usually adequate for deep resistivity calculations, but a more detailed invasion profile is required for porosity log corrections. It is concluded that borehole and invasion effects can probably never be completely eliminated, but by combining the responses of various tools and forward modeling, the most likely solution can be found, and erroneous interpretations avoided. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER B ON THE QUEST FOR DEPTH. ir A.M. (Ton) Loermans Nederlandse Aardolie Maatschappij The Netherlands S. (Steve) Kimminau Shell International Exploration and Production, The Netherlands H. (Harald) Bolt TechMan Consultants, The Netherlands SUMMARY. Depth is the most important logging parameter. Two main problems occur: (i) sometimes the depth reported by the logging contractor is seriously wrong and (ii) in field studies occasionally petroleum engineers just assume the (correctly measured) depths to be in error and “correct” the well data to fit an erroneous reservoir model. Illustrations of both problems are presented. While the cases shown were “near misses”, i.e. the errors were picked up just prior to major investment decisions, the damage and financial loss could have been many million dollars indeed. A further analysis of these and similar incidents, proves indeed that there is need for two things: (I) a tightening up of the depth control procedures and (2) an improved audit trail made available to the client. Following the initiatives triggered at the 1996 EFES symposium and the discussions on the SPWLA special workshop in Taos 1997, several projects were started. Presently at least one major logging company is routinely providing a full audit trail on the log header. Technical reviews of actual depth control practices are increasing the awareness with both logging companies and Operators and have led to further improvements in the procedures being used. Currently, apart from the continued requirement for dedication to Quality Management, the principal issues outstanding are (1) improve stretch corrections and (2) report actual depth uncertainties on every log. The potential for drill pipe and casing measurements to become of more value, by a similar increased focus on quality management of the measurements can be done along the same lines. With the QM measures proposed for conventional depth measurement systems, there would be no firm reason to abandon the traditional methods of depth control, unless a drastically new method would yield the same accuracy at significantly lower costs. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER C CHARACTERIZATION OF DUAL POROSITY SYSTEM IN CARBONATES USING SONIC AND RESISTIVITY MEASUREMENTS D J Dutta ONGC-Schlumberger Joint Research Center, New-Delhi, India. S Madhavan, KM Sundaram Oil and Natural Gas Corporation Limited, Mumbai, India ABSTRACT Multiporosity systems are quite common in carbonates in contrast with clastics in which most of the porosity is interganular in nature. An important portion of porosity in carbonates is often contributed by vugs (solution cavities) and intercrystaline pores in addition to intergranular porosity. Carbonate rocks also contain solution channels/path of solution migration which have random orientation, random distribution and therefore lack coherence with current distribution generated by electrical devices and therefore appear indistinguishable from the electrical behavior of a porous fabric. Similarly, their high aspect ratios and randomness in respect of orientation and distribution make them amenable to be modelled acoustically akin to rock fabric rather than as a separate system. It therefore appears that modeling a carbonate rock as a modified fabric consisting of intergranular porosity and randomly distributed solution channels, and spherical to near spherical inclusion constituted by vugular and moldic porosity is a fruitful way of evaluating carbonates. In this paper we have used the above model and adopted the Kuster-Toksoz equation for the acoustic domain and the low frequency conductivity analog of the Maxwell-Garnet equation in the electrical conductivity domain to evaluate carbonate intervals. The transform of the acoustic parameters to the primary porosity (of the modified fabric) and the inclusions (secondary porosity due to vugs and molds) has been performed using both the Gassmann equation approach of rock physics as well as Wylie empirical equation approach. Both the approaches give approximately identical partitioning of the total porosity into the two porosity classes referred to above. Excellent match of the forward modeled resistivity of water bearing intervals (with the Maxwell-Garnet equation analogue used for transforming porosity to resistivity for non-contiguous vug cluster), with field resistivity curve validates the usefulness of the rock model and equations employed, in evaluating the carbonate formations of the type discussed. Partitioning of the secondary porosity on the basis of hierarchical filling of vugs and primary porosity by hydrocarbon is attempted in the transition zone near oil-water contact. Two field examples in different areas in western offshore India are presented and discussed in this paper. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER D GAS WHILE DRILLING (GWD); A REAL TIME GEOLOGIC AND RESERVOIR INTERPRETATION TOOL G.Beda and R.Quagliaroli EM Agip Div., Milano, Italy G.Segalini, B.Barraud, and A.Mitchell ELF EP, Pau, France ABSTRACT The acquisition of gas in mud data while drilling for geological surveillance and safety is an almost universal practice. This source of data is only rarely used for formation evaluation due to the widely accepted presumption that they are unreliable and unrepresentative. Recent developments in the mud logging industry to improve gas data acquisition and analysis has led to the availability of better quality data. Within a joint ELF/ENI-Agip Division research program, a new interpretation method has been developed following the comprehensive analysis and interpretation of gas data from a wide range of wells covering different types of geological, petroleum and drilling environments. The results, validated by correlation and comparison with other data such as logs, well tests, PVTs etc. enable us to characterise: • lithological changes • porosity variations and permeability barriers • gas/oil and hydrocarbon/water contacts • vertical changes in fluid over a thick mono-layer pay zone The comparison between surface gas data and PVT data clearly confirms the consistency between the gas show and the corresponding reservoir fluid composition. The near real time availability, at no extra acquisition cost, of such data has led to: • the optimisation of future well operations (logging, testing • a better integration of while drilling data to the well evaluation process • a significant improvement both in early formation evaluation and reservoir studies especially for the following applications where traditional log analysis often remains inconclusive: • very low porosity reservoirs •. thin beds • low resistivity pay • light hydrocarbons th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER E WHICH RESISTIVITY SHOULD BE USED TO EVALUATE THINLY BEDDED RESERVOIRS IN HIGHANGLE WELLS? Jacques R. Tabanou, Barbara Anderson, Sue Bruce, Ted Bornemann, Kyel Hodenfleld, and Peter Wu Schlumberger ABSTRACT Thinly bedded sandstone reservoirs are being developed extensively offshore. Most wells in these developments are high-angle or horizontal and logging-while-drilling (LWD) is the common choice of data acquisition. These reservoirs reveal a large resistivity anisotropy when logged with 2-MHz tools in high-angle wells. Log data can be inverted to compute horizontal resistivity (Rh) and vertical resistivity (Rv). However, the anisotropy effect can be hidden if invasion is sufficiently deep. Furthermore, relative dip is required for accurate inversion of Rv . But even when an accurate Rh and Rv estimation is made, the question remains “which resistivity should be used to calculate hydrocarbon reserves?” Firstly, with three-dimensional (3-D) modeling, this paper establishes the conditions of invasion under which Rv and Rh estimates can be used reliably. Density images from a LWD tool provide a means to calculate the relative dip for use as required input to the Rv Rh inversion. Secondly, traditional methods for determining water saturation that use only Rh require accurate inputs of shale volume and bound water resistivity. This paper proposes a method of using Rv , Rh and an input for the resistivity of the shale laminations to derive the resistivity of the clean sand layers and the net-to-gross ratio. Hydrocarbon volume is similarly calculated by either the traditional Rh method or by the proposed method using both Rh and Rv. A universal interpretation chart is presented for Rv Rh space to facilitate quick-look detection of thinly bedded pay sands and the estimation of oil in place. Log examples demonstrating the use of these techniques are presented. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER F A NEW MWD FULL WAVE DUAL MODE SONIC TOOL DESIGN AND CASE HISTORIES Georgios L. Varsamis, Laurence T. Wisniewski, and Abbas Arian SensorWise, Inc., Houston Gary Althoff Sperry-Sun, Halliburton Energy Services. Inc., Houston ABSTRACT As MWD/LWD formation evaluation applications evolve, the need for a full suite of logging sensors becomes evident. A key aspect of any complete logging suite is a full-wave sonic tool. In a drilling environment a sonic tool has applications that extend beyond standard porosity-type evaluations. MWD/LWD sonic logs can also provide real-time seismic tie-in as well as assist in drilling optimization. Conventional perceptions are that there are two key aspects that could minimize or preclude the use of sonic technology in an MWD/LWD environment First, drilling noise is assumed to exist within the frequency range of the measurements, and minimization of it was considered impractical. Second, the interference of the tool-body mode (acoustic waves coupled to the receivers as a result of tool body waves) is thought to be dominant, to the extent where significant manual post-processing is required to generate reasonable logs. This paper presents some fundamental design choices and several case histories from a new Full-Wave Dual-Mode MWD/LWD sonic tool. The tool is constructed such that it operates in two modes (operator selectable and available as either/or or both): (i) classical Compressional Wave Sonic tool (the equivalent to the traditional monopole source mode); (ii) Enhanced Shear-Wave Detection mode (multipole source mode). The tool has multiple transmitters and multiple receivers (arranged in a combination of arrays) that allow operation in both modes. The tool utilizes a unique set of isolation sections that minimize the effects of drilling noise and tool-body mode. The design of these isolation sections was based on extensive lab testing with full size models, as well as the use of sophisticated 3-D analytical models. These studies led to a comprehensive understanding of the dynamics of the tool body mode and drilling noise and the best ways to reduce them. Implementation of the final design resulted in a negligible tool mode and nearly complete isolation from drillstring noise. The case histories include data from a variety of geographic locations including both fast and slow formations and from the AMOCO test well in Catoosa. Through the case histories the functionality of the tool and its unique characteristics are demonstrated. Emphasis is placed in the presentation (through case histories) of the tool’s capability to measure formation shear velocity in slow formations. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER G INTERPRETATION OF ASYMMETRICALLY INVADED FORMATIONS WITH AZIMUTHAL AND RADIAL LWD DATA Darrel E. Cannon and Craig Kienitz Schlumberger ABSTRACT In permeable, dipping formations, invasion of drilling fluid is often asymmetric because of gravity slumping of the filtrate. This effect can be observed within less than an hour of the bit penetrating a highly permeable, gas-bearing formation. Accurate interpretation of log data in such environments requires a technique that accounts for both azimuthal and radial distribution of filtrate. Logging-while-drilling (LWD) measurements, when rotated through the zone of interest, offer the data necessary to evaluate such formations. The interpretation process used combines azimuthal nuclear data with azimuthal and radial resistivity data to compute accurate values of porosity, water saturation and mineralogy. First, resistivity data are inverted for Rx0 (flushed zone resistivity), Rt (true resistivity) and Di (diameter of invasion) in four directions. Next, the different values of Di are used to compute individual invasion scalars for density and neutron log data based on their radial response functions. This is possible because the resistivity measurements used are similar in radial response to the density and neutron radial responses. Then, the log data and invasion scalars are entered into a petrophysical solver for the final results. The invasion scalars assist the solver in determining the magnitude of the hydrocarbon correction required, which is especially important in gas zones. Full correction of log data provides results that are in close agreement with core data. Understanding the effects of gas in three dimensions helps explain “lazy” neutron curves. In gas zones, the density log is highly affected by varying invasion, where as the neutron has an almost constant gas effect that is relatively independent of invasion as predicted by modeling. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER H JOINT INVERSION OF MPR AND HDIL MEASUREMENTS Raghu K. Chunduru, Alberto G. Mezzatesta, Hal W Meyer Baker Hughes Incorporated Tom Maher Shell Offshore Inc. ABSTRACT Traditionally, measurement-while-drilling (MWD) data are used primarily for geosteering purposes and drilling decisions such as monitoring of hole direction, deviation, and delineation of abnormally pressured zones. Wireline resistivity measurements, galvanic and induction, play a fundamental role in identiftying and delineating oil- and gas-bearing formations. In recent years, MWD logs have been used not only for drilling decisions but also for realtime formation evaluation. The availability of both MWD and wireline data not only provides the interpreter with abundant information about subsurface formations but also poses a new challenge to generate a unique model(s) that better explains both data sets. In this study, we combine MWD and wireline measurements in order to take advantage of the best features of the data in finding a consistent earth model. The proposed interpretation is applied to synthetic and field data examples. MWD and wireline data obtain responses from similar subsurface formations at different times, reflecting different borehole and invasion conditions. Each of these data offers distinct advantages when compared to the other. In general, MWD data are not affected by invasion, allowing better interpretation of formation resistivities. On the other hand, wireline data allow us to characterize the, invasion profile, resulting in identification of permeable and impermeable zones, and thereby facilitating the evaluation of movable and residual hydrocarbons. Conventionally, MWD and wireline data are interpreted independently to estimate formation resistivities that might result in inconsistent earth models. In our study, a dual earth model that describes the appropriate logging conditions of both wireline and MWD is considered. This model contains a set of common parameters including the true formation resistivity and bed boundaries. An inversion process that handles the dual model has been developed and implemented that deals with any combination of wireline and MWD resistivity measurements. The proposed scheme is implemented using Multiple Propagation Resistivity (MPRSM) and High Definition Induction Log (HDILSM) measurements. The value of the joint inversion process is demonstrated by the estimation of reliable and consistent formation parameters for both synthetic and field data examples. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER I SEISMIC TIE USING SONIC-WHILE-DRILLING MEASUREMENTS Mohamed Hashem Shell Offshore Donald Ince Arco Alaska Kyel Hodenfleld and Kai Hsu Schlumberger ABSTRACT Surface seismic has guided geoscientists in the exploration and development of hydrocarbon reservoirs. After a geological target has been identified for drilling, the surface seismic is used to develop a drilling plan. However, uncertainties in seismic velocities can result in depth errors on seismic sections to geological targets of interest and to drilling hazards. Depth calibration of the surface seismic data includes acquiring wireline sonic or seismic checkshot data or both to convert the seismic two-way traveltime to depth. In some cases, acquiring such data after the well has been drilled may be too late to avoid a costly sidetrack or properly prepare the wellbore for a drilling hazard. Real-time compressional at acquired while drilling has been used to put the drill bit on the seismic map. Thus it is possible to determine if a given reflector has been crossed or how far the reflector is from the current bit position. This paper shows sonic-while-drilling measurements obtained over a wide range of borehole, environmental and petrophysical conditions. It also demonstrates that correlating a real-time synthetic seismogram with the surface seismic traces allows the placement of the well path on a seismic map while drilling. Furthermore, the time-depth relationship obtained from integrating the & measurements provides an opportunity for the petrophysicist or geologist to tie surface seismic features with other logging-while-drilling (LWD) logs for real-time formation evaluation and interpretation. The advantages of seismic tie-in demonstrated by the examples include (1) confirming/modifying well plans and trajectories to reservoir targets; (2) refining depth to drilling hazards to allow for proper preparation and improve drilling safety; and (3) eliminating the uncertainty of acquiring wireline data in difficult wellbore conditions and possibly reducing the cost and time involved with wireline acquisition. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER J PROPAGATION RESISTIVITY TOOLS W. Hal Meyer Baker Hughes INTEQ ABSTRACT Statements in recent literature have indicated that the apparent resistivity is a strong function of frequency at 2 MHz (the most common frequency for propagation resistivity tools). If this were true, resistivity dispersion should be a major factor in the interpretation of MWD resistivity data. Some published laboratory data even suggest that water saturations measured at 2 MHz are grossly inaccurate due to dispersion effects. However, the analysis of hundreds of logs with data at two frequencies (400 kH.z and 2 MHz) demonstrates that very little resistivity dispersion occurs in oilfield formations at these frequencies. Analysis of practical oilfield formations using theoretical models shows that resistivity dispersion should not become apparent unless the frequency is much higher than 2 MHz. This resistivity dispersion can be seen clearly in wireline field data when high frequency dielectric tools are used in the same formation as low frequency induction tools. However, when 2 MHz tools are run with induction tools the results are usually comparable. When the results are different, some effect other than resistivity dispersion is usually to blame. The same theoretical models indicate that dielectric dispersion should be significant at 2 MHz, which is in agreement with the two-frequency MWD field data. Hundreds of two-frequency logs through thousands of formations have been analyzed and several typical examples are shown in this paper. The two different measurements are made almost simultaneously using the same antennas. Separation between the resistivity data at the two frequencies is common. However, other effects such as anisotropy, invasion, dielectric effects, and differences in vertical resolution are found to explain these differences more accurately than resistivity dispersion. All of these other effects cause separation of attenuation and phase resistivity data. separation of data from different spacings, or both. Therefore, these effects can be distinguished from resistivity dispersion when all available data are compared with the theoretical formation responses. Only one clear case of resistivity dispersion has been found in any of the two-frequency data sets. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER K FORMATION EVALUATION FROM LOGGING-WHILE-DRILLING DATA IN A 6.5 INCH HORIZONTAL WELL - A NORTH SEA CASE STUDY Carlos Macso Schlumberger Oilfield Services Nikolai Sudaldewicz Oryx U.K. Energy Company Philip Leighton Topaz Consultants Ltd ABSTRACT Horizontal wells are being increasingly drilled in the development and production cycles of oil fields. It is important to be able to maximise the amount of information obtained from relatively restricted data sets offered by LWD tools with respect to conventional wireline tools. This paper illustrates the type of geological, petrophysical and borehole information that can be obtained from a thple combination Logging-While-Drilling (LWD) data set in a 6,5-inch horizontal hole. A continuous LWD data set was obtained over an interval of 3930’ of measured depth (MD) from the BRENT Group of the Murchison Field, UK Northern North Sea. The data set comprised gamma ray, ten resistivity curves, azimuthal density and neutron data; no other logging tools were used in the well. Interpretation of the data set has provided invaluable information concerning the borehole condition, geology and petrophysical properties. Full integration with available wellsite data (e.g. palynofacies analysis) enabled the operator to better characterise the target Rannoch Formation (Mica Sand Unit). Geological information obtained by the use of density images (although the tool was run slick) helped differentiate nodules/concretions from beds and the direction of drilling relative to the bedding. Borehole spiralling could also be recognised from images and was directly related to changes in the bottom hole assembly. Further geological and petrophysical information was obtained by close examination of the resistivity curves. A number of zones exhibit resistivity curve separation. The order of resistivity curve separation varies with both shallow to~p and deep to shallow tends being recognised. Where a deep to shallow trend is present, oil-based mud invasion is the conventional interpretation. This can be confirmed in zones with time lapse logging. Where a shallow to deep trend is recognised interpretation is more complex, requiring both the phase shift and attenuation curves to be examined. Here combinations of bed boundary effects and anisotropy are recognised and confirmed by modelling. Resistivity anisoiropy in certain zones can be related to sedimentary character while tool response across boundaries can help determine the dip and lateral continuity of the beds. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER L A NOVEL APPROACH FOR COMPENSATING NEUTRON POROSITY LOGS FOR BOREHOLE EFFECTS Michael Evans, David Best and Kyel Hodenfield Schlumberger ABSTRACT Neutron porosity logging techniques have been used for many years to measure the porosity of a formation surrounding a borehole. It is well known that the measurement is adversely affected by changes in measurement geometry such as borehole size, shape and tool position within the borehole (tool standoff). Compensation techniques attempt to overcome these perturbing effects by using two detectors — one located far from the source to measure formation porosity and a second located near the source to measure the effects of changing measurement geometry. Taking the simple ratio of near-to-far (N/F) counting rates reduces the effects of changing geometry. However, this procedure does not provide complete compensation and the resulting porosity values must still be corrected for borehole size, shape, and tool position within the borehole. Much improved compensation for borehole geometry effects is achieved by modifying the simple near-to-far ratio. A function of the far-detector count rate can be determined that results in the two detectors having nearly identical radial responses in the proximity of the tool. The ratio of the near-detector count rate to this function of the fardetector count rate yields a modified ratio that is insensitive to geometric perturbations that occur near the tool. This modified ratio results in a porosity measurement that is borehole invariant — a measurement that virtually needs no correction for washouts, rugosity, borehole shape or tool standoff. The technique is applicable to both wireline and logging-while-drilling (LWD) neutron porosity measurements. The benefits of this new compensation technique will be described and illustrated with laboratory data and Monte Carlo simulation results. The ability of the technique to implicitly account for changing borehole geometry will be demonstrated with several well log examples. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER M A NEW ALGORITHM FOR CORRECTING NEUTRON DECAY LOGS FOR BOREHOLE AND DIFFUSION EFFECTS Medhat W. Mlckael Baker Atlas, Houston, Texas ABSTRACT Pulsed neutron capture (PNC) logs have been in routine use for decades for the measurement of formation capture cross section. The measurement is robust, accurate, and statistically precise. However, borehole and diffusion effects on the measurement are difficult to characterize and use since they require accurate knowledge of not only borehole size but also borehole capture cross section. This paper presents an innovative approach to correct pulsed neutron logs for borehole and diffusion effects without the knowledge of borehole capture cross section or water salinity. This is accomplished through the utilization of all the information available in the time decay spectra of both the near and far detectors. The correction to the apparent formation capture cross section of one of the detectors is described as a mathematical model of the relative counts in different gates of the time decay spectra of both detectors. The coefficients of the model are obtained from a nonlinear least-squares fit of the model to data from different borehole and formation conditions. The model was optimized using over 5000 data points generated from accurate Monte Carlo simulations covering a very large range of down-hole conditions. The accuracy of the new algorithm when tested on modeling data was found to be on the order of 0.5-1 on over the entire range of formation and borehole capture cross sections. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER N CONTINUOUS OIL, GAS AND WATER HOLDUP USING PULSED-NEUTRON SPECTROSCOPY TECHNIQUES Frank Morris and James Hemingway Schlumberger ABSTRACT In recent years, as horizontal wells have become more prevalent, conventional production logging services have proved inadequate for segregated fluid regimes. Many new techniques utilizing nuclear tools have been developed to overcome the problems with older measurements. One such technique utilizes pulsed-neutron spectroscopy for making continuous oil, water and gas holdup measurements in producing wells, regardless of deviation. This measurement uses yields-based carbon-oxygen ratios at two detectors that are especially sensitive to oil versus water and a net inelastic count rate ratio between the detectors for providing information about gas. Relatively linear sensitivity to all three borehole fluid phases is achieved with centralized tool operation. The technique is based on a set of response equations giving detector carbon-oxygen ratios and the net inelastic count rate ratio as functions of the quantities of oil, water and gas in the borehole and oil and water in the formation. Inverting these equations gives a measure of holdup and its precision. A solution can be obtained with no a priori information about fluids present. However, the solution algorithm permits the input of any knowledge about borehole and formation fluids that will enhance the accuracy of the results. To ensure the accuracy of the technique, measurement response for oil and water has been characterized in limestone and sandstone test formations with a variety of typical completions. Response characterization in gas, however, cannot be done with laboratory measurements because of safety issues. Instead, Monte Carlo modeling was used to expand the characterization to gas as well as to other special completions. This technique is now being used routinely in production logging. Holdup measurements with an accuracy of 5 holdup units or better are achievable at logging speed of 200—500 ft/hr. Comparisons between holdup measured by this method and probe-based techniques have verified the overall accuracy of this technique. Case studies demonstrate the value of holdup measurements made in this manner for identifying and quantifying points of unwanted fluid entry. In addition, combining these holdup measurements with oil and water flow velocities also measured using nuclear techniques gives production rates in wells where conventional techniques do not work. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER O INTRODUCTION OF ENHANCED CARBON-OXYGEN LOGGING FOR MULTI-WELL RESERVOIR EVALUATION J. Hemingway, R. Plasek, J. Grau, T. Das Gupta and F. Morris Schlumberger ABSTRACT Recent advances in spectral processing techniques for pulsed neutron spectroscopy (PNS) tools have improved accuracy and precision of measured carbon-oxygen (C/O) ratios. Improved elemental standards, combined with the development of a full spectrum calibration, provide better tool-to-tool accuracy and precision in a wider range of porosities and in the presence of gas. Using the C/O ratio to compute water saturation offers many advantages over conventional techniques that depend on formation water salinity. The C/O ratio relates directly to the volumes of oil and water in the formation. The conversion of CO ratio to oil saturation is based on a very large database acquired using laboratory formations with a wide range of wellbore environments. This database was acquired over several years at the Environmental Effects Calibration Facility (EECF) in Houston, Texas. Previous conversion of C/O ratio to oil saturation was based on a simple “nearest neighbor” interpolation technique in the database. The new method is based on a weighted multiple linear regression (WMLR) technique in an expanded database. The WMLR technique utilizes all the points in the database so that the tool response across the entire database is smoother and more robust. WMLR is used to compute oil saturation from two sets of C/O ratios. Combining the oil saturation computed from the spectrally derived CO ratio with the oil saturation inferred from a windows-derived C/O ratio produces an accurate as well as a precise oil saturation. A new technique is presented for computing changes in oil saturation more precisely over a period of time using the windows CO/ ratio combined with an environmental factor, alpha. The improved precision of the C/O ratio measurement allows the tool to be used for monitoring oil saturation in existing fields where secondary or tertiary recovery techniques are in use, thereby maximizing production and minimizing cost. Shutting off depleted zones can be just as important as identifying by-passed reserves. Experience with several large field studies has led to the validation of this new CO logging technique. Comparisons of CO-derived oil saturations from several field tools with core-derived oil saturation have established confidence in the new technique. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER P INVESTIGATIONS INTO IMPROVED METHODS OF SATURATION DETERMINATION USING PULSED NEUTRON CAPTURE TOOLS P.Cowan and G.A.Wright, AEA Technology plc, UK ABSTRACT One of the main uses of Pulsed Neutron Capture logs is to monitor well production by determining saturation behind casing using ‘time lapse’ sigma analysis. This method works well when there is a large contrast between the intrinsic formation sigma when saturated with water and when saturated with hydrocarbons, but breaks down in low contrast conditions. This occurs when the formation water is fresh or slightly saline and porosity is low. Amerada Hess, BP Exploration, Norsk Hydro, Saga Petroleum and Baker Atlas sponsored a programme of work to investigate new methods of saturation determination for low contrast conditions. Monte Carlo modelling was used in these investigations to identify potential methods and to demonstrate them theoretically. Finally the new methods were successfully applied to real log data. The MCBEND Monte Carlo code was used for this study. Developed by AEA Technology, MCBEND provides an accurate simulation of radiation transport for sources of neutrons, gamma-rays or electron/positrons as well as coupled (n,y) processes. It has been successfully applied to all types of nuclear logging applications over a number of years as it performs an accurate computer simulation of the operation of the nuclear logging tool. Detailed and validated computer models of commercial Pulsed Neutron Capture Tools were established in MCBEND, using the ‘secure geometry’ option which protects sensitive tool information. Numerous MCBEND calculations were performed for these tools in low contrast conditions. Analysis of the tools’ response concentrated on utilising detector counts rather than the sigma response. New methods of water saturation determination were derived using detector counts and detector count ratios. These new methods theoretically proved to be considerably more accurate than the traditional time lapse method. The next stage was to demonstrate the new methods of saturation determination by applying them to log data. Log data were provided by the project sponsors. Results show that the new methods can be successfully applied to real log data - giving a viable alternative to time lapse sigma analysis in low contrast conditions. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER Q SIMULTANEOUS CORRECTIONS FOR MUTUALLY-DEPENDENT ENVIRONMENTAL EFFECTS OF COMPENSATED NEUTRON MEASUREMENTS Medhat W. Mlckael and PingJun Guo Baker Atlas, Houston, Texas ABSTRACT Traditional applications of environmental corrections for compensated thermal neutron measurements assume that environmental effects are independent of each other and that the final correction is a sum of the individual corrections. While this assumption is valid for many corrections, it was found that some corrections are mutually dependent and require a different treatment. A study has been carried out to isolate the corrections that are mutually dependent and devise new corrections to treat the combined effects simultaneously rather than sequentially. The two major effects that exhibited large mutual dependency were 1) borehole fluid salinity and tool standoff, and 2) formation fluid salinity and temperature. The results of the modeling study are presented to show the interaction between these effects. New corrections treating these effects simultaneously were developed and tested. The application of these corrections on an extensive set of simulated and experimental data is also presented to show the improvement in the accuracy of the corrected porosity over the traditional sequential corrections. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER R ANALYSIS OF NON-LINEAR RESISTIVITY INDEX VS SATURATION DATA USING A BINARY ARCHIE MODEL G. A. Brown Sensor Highway Ltd. ABSTRACT Since Archie published his saturation equation the exponent “n’ has been derived from the analysis of core samples using a logarithmic Resistivity Index-Saturation (RI-Sw) cross-plot. However, not all core sample data falls on straight lines corresponding to Archie’s theory. This paper develops a Binary Archie electrical Model of the formation based on the concept that the rock fabric is composed of two (or more) individual morphologies co-existing independently at the microscopic level. A fitting technique is employed to derive the saturation and porosity exponents (n & m), and the percentage contribution of each morphology from non-linear RI-Sw data. In the case of shaly data the parameters n*, m* and the Cation Exchange Capacity (Qv) can also be determined. This allows the electrical properties of the rock to be described in terms of conventional Archie concepts, without the need for additional empirical parameters. The BAM model is demonstrated to be a good fit to both Shells Group UI shaly sand data set and a wide variety of other clean and shaly non-linear RI-Sw data. Evaluation of shaly sand data indicates that the excess conductive component is primarily associated with the micro-porous morphology and that the tortuosity of this morphology, characterised by the porosity exponent, has a significant effect on the shape of the RI-Sw curve. The technique has been found to derive consistent parameters from groups of core samples which normally produce a spread of conventional saturation exponent values. In addition, the model can be used to understand the differences between equilibrium and transient RI-Sw data measured on the same core sample using the Porous.Plate and Continuous Injection techniques. The results demonstrate that pessimistic calculations of oil or gas in place (0IIP or GIIP) will often be made when the true binary character of rocks has not been recognised and properly taken into account. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER S INTEGRATED PETROPHYSICAL ANALYSIS AND SEQUENCE STRATIGRAPHY OF THE PALAEOCENE INTERVAL, FAROES-SHETLAND BASIN A. Griffin, M. Andrea, and P.F. Fish Enterprise Oil plc ABSTRACT Enterprise Oil has developed a sequence stratigraphic framework for the Flett and Erlend sub-basins, West of Shetlands. Three regionally correlatable, unconformity bounded mega-sequences have been identified. These megasequences are further subdivided into depositional sequences recognisable at well and seismic scales and form the framework of a detailed play fairway analysis which focuses on the identification of reservoir and seal horizons. Petrophysical data is a key element of the play fairway analysis. The aim of this paper is to present a predictive reservoir effectiveness model that has been developed to aid exploration of the Palaeocene play. The model addresses the distribution of reservoir porosity and net effective reservoir thickness with burial depth, and provides a dataset for reservoir parameter input to exploration prospect evaluation. The model indicates that effective reservoir quality is significantly reduced at burial depths greater than —10,000 ft. of overburden. In addition, the model identifies a sub-set of wells with anomalously high reservoir quality at burial depths >10,000 ft. These high porosity sands have distinctive amplitude versus offset (AVO) and acoustic impedance (AI) signatures. Petrophysical evaluation of the Palaeocene reservoir quality is further complicated by low salinity formation waters giving the potential for low contrast pay zones. Generally, salinities are <25,000 ppm NaCI equivalent, and decrease with increasing stratigraphic age and burial depth. Vertical salinity variations are discussed. Log analysis has been calibrated, where possible, to available core data. The development of a predictive reservoir effectiveness model, integrated within a sequence stratigraphic framework, has highlighted the importance of petrophysics and rock physics to the understanding of reservoir quality distribution in the Flett and Erlend Basins. The dataset is currently being used as a tool for acreage evaluation, prospect generation and volumetric modelling of Palaeocene reservoirs, West of Shetland. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER T VEOLOCITY-POROSITY-MINERALOGY GASSMANN COEFFICENT MIXING RELATIONSHIPS FOR WATER SATURATED SEDIMENTARY ROCKS Phil Holbrook, Igor Goldberg and Both Gurevich ABSTRACT Clay mineral diagenesis and PVT-x Sodium Chloride brine relationships play pivotal roles in the full implementation of Gassmann equation forward and inverse modeling. The velocity of Sodium Chloride brine varies from 1420 M/sec to 1960 M/sec over the temperature - pressure - salinity range that exists in the porous sedimentary crust. The bulk modulus and density coefficients of Sodium Chloride brines were extracted from voluminous measured velocity data by Archer (1992). Archer’s equation of state thermodynamic molecular interaction coefficients were re-cast as third order functions of NaC1 brine density, pressure, temperature and molality. Sodium Chloride brine bulk modulus, density and velocity are linked through Archer’s equation of state. The third order PVT – x NaCl regression of NaC1 brine equation of state provides very accurate physically consistent fluid coefficients for Gassmann equation forward and inverse modeling. Clay minerals have been the mineralogic stumbling block for general Gassmann equation forward and inverse modeling. The Vp of clay minerals varies from 2600 M/sec for hydrated sedimentary claystones to 9400 M/sec for dry low-grade metamorphic claystones. However, it has been found that the variability of claystone Vp, Vs, and elastic coefficients is highly systematic. The elastic coefficients extracted from velocity measurements through Hashin-Shtrikman inversion are closely comparable over the entire observed velocity — porosity - grain surface wetness range. Clay mineral particles are mechanically well behaved when the elastic coefficients of the surrounding solid and fluid media are properly accounted for through weighted average Hashin-Shtrikman inversion. Low magnesium sedimentary clay minerals have a mean Poisson’s ratio of 0.29 ± 0.3. The independently derived system measurement error is 10%. The dry rock bulk and shear modulus ratio for sedimentary clay minerals can be taken as a constant Poisson’s ratio for inversion purposes. Hashin-Shtrikman inversion indicates that sedimentary clay mineral particles are apparently hydrated with an electrostatically bound water layer over 90+% of their total porosity range. The elastic properties of clay mineral particles are uniformly low when the bound water layer is present. The velocities and elastic properties of claystones increase rapidly to that of dry clays with the loss of the last few units of water filled porosity. Self weight consolidation to zero porosity and total clay mineral dryness occurs before the low grade metamorphic clay mineral transition. Gassmann equation forward and inverse modeling is tractable when variability of fluid and clay mineral elastic coefficients are taken into account. The elastic coefficient mixing laws for the Gassmann and Woods equations are linear. Optimum average (α & β) pore compliance coefficient dependencies were determined through Gassmann equation porosity inversion. A true statistical test of coefficient extraction and forward modeling application was performed on separate, but nearby datasets. Using the average (α & β) pore compliance porosity dependence, the measured vs. predicted accuracy of Gassmann equation forward and inverse Vp, Vs and porosity modeling is about 5%. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER U MAJOR CHALLENGES IN OLD OIL FIELD SOLVED. ir A.M. (Ton) Loermans Nederlandse Aardolie Maatsehappij The Netherlands. SUMMARY. The object of this study is a one billion barrel STOIIP field, discovered during the Second World War with currently some 600 wells. Yet, after more than 40 years of drilling, production and petroleum engineering studies, there were still a few fundamental problems: (i) paradoxes on the original fluid contacts, (ii) major uncertainty on the net to gross ratio (N/G) in an important part of the field and (iii) some unexplained log responses. As part of a larger petroleum engineering study, the above petrophysical issues were tackled successfully. First of all the original fluid contacts. By the combination of (i) a rigorous well by well evaluation, and (ii) making the principal differentiation between Oil Water Contact (OWC) and Free Water Level (FWL) the paradoxes were solved. As a “by product” some valuable clues on the structural geological history of the field were obtained and post charge faulting could be quantified based on the observed variations in contacts. Secondly, building further on the work done on the fluid contacts, the N/G problems could be solved. It was concluded that the apparent very low N/G ratio’s initially derived from the standard methods of log evaluation were erroneous due to the special effects of the peculiar crude present in this field. Finally, some other peculiar log responses seen in several parts of the field during various stages of the. field’s history were explained satisfactorily, providing a solid basis for a complete and coherent petroleum engineering model for the field. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER V CASE STUDY OF AN ANISOTROPIC CROSSBEDDED RESERVOIR—IMPROVING PREDICTIONS OF RESERVOIR PROPERTIES THROUGH MODELING THE RESISTIVITY Tom Walsgrove Amoco Barbara Anderson and Tom Barber Schlumberger ABSTRACT The Davy Field, situated at the south-eastern edge of the UK sector of the Southern North Sea, was discovered in 1970 by the Amoco-operated 49/30-2 well, which tested gas from the Permian Rotliegend Sandstone. Development of the field occurred during late 1995 to early 1996 with the drilling of four horizontal wells. The producing reservoir, the Leman Sandstone, developed across the Davy area, consists of 300-400 feet of predominantly high quality aeolian dune facies sandstones (porosity 17%; permeability 25-250 md). Within individual dune facies, there is considerable fine-scale variation in sorting and grain size, which provide the major controls on reservoir quality. Permeability is very anisotropic, with production flow parallel to crossbedding. Although comprehensive well data exist for both vertical and horizontal wells, reliable water saturation determination has been problematical. The observed mismatch between log-calculated and core-derived capillary pressure water saturations appears to be related to formation anisotropy, resulting from the cross-bedded nature of the aeolian dune sandstones. This mismatch becomes more apparent in the horizontal wells. Computer modeling of multiarray induction tool sensitivity to typical Rotilegend crossbedding anisotropy shows that the apparent resistivity read by the tool can vary by a factor of 2 to 5 depending on the well deviation and the crossbedding angle. Using horizontal and vertical resistivities within the modeled range and a knowledge of the local fine structure, we are able to reconcile the calculated saturations and the log values. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER W USING PETROPHYSICS TO IMPROVE RECOVERY: WHITNEY CANYON-CARTER CREEK FIELD, WESTERN WYOMING THRUST BELT, USA Pamela A. Logan, Gary W. Gunter, Charles R. Smart, and Mike A. Miller BPAmoco ABSTRACT Whitney-Canyon Carter Creek, is a significant North American gas field discovered in 1978 which has produced 1.3 TCF of gas to date of the approximately 4.0 TCF of gas in-place. It produces from the Madison Group, a complex carbonate. This case study highlights how probabilistic formation evaluation techniques along with an integrated three-stage petrophysical process were applied to examine the significant differences in well performance and provide recommendations to increase recovery. Stage I defines petrophysical rock types for the Madison carbonates by relating geologic framework, lithofacies and petrology to porosity, permeability and capillarity. Stage 2 for the Madison carbonate produces petrophysical models of pay and flow behavior. Probabilistic formation evaluation techniques were extended to develop petrophysical rock types models. Using nuclear modeling was a key step in understanding a complex gas correction and the effects of boron on porosity. Followed by applying the Modified Leverett J Method to identify the free-water level and confirm the wireline calculations in this highly resistive reservoir system. Applying this integrated approach was important in recognizing additional upside potential Stage 3 integrated the reservoir description from the previous stages with actual well production. Pressure transient analysis was used to evaluate fluid sensitivity and formation damage mechanisms. Reservoir simulation analyzed the impact of flow units on ultimate economic recovery. Recommendations on possible remedial actions were. a crucial product of this integrated field study. In summary the Model (PLPM) was successfully applied to characterize the reservoir and the drivers of the reservoir performance. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER X DETERMINATION OF EARTH STRESSES, AZIMUTHAL ROCK-MECHANICAL PROPERTIES AND FRACTURE PERMEABILITY USING DIPOLE SHEAR SONIC LOGGING IN THE GULF OF SUEZ, EGYPT T.Klimentos and M.Badri Schlumberger S. Hassan Petrobel A.Zeid Schlumberger ABSTRACT Regional stress patterns, such as the one in the Gulf of Suez, which is attributed to the Red Sea Rift, give a general indication of the likely fracture orientation or maximum stress trend. However, local variations and the effects of localized structures, such as large faults, can alter the stress pattern completely, counteracting or adding to the regional stress. Such a local stress information can be very important to many petroleum exploration and development related aspects, i.e., hydrocarbon migration, hydraulic fracturing, optimum well placement, welibore stability, and sand production. In this case-study, shear-wave anisotropy data obtained from dipole shear sonic logging over fractured reservoirs in the Gulf of Suez, Egypt, was used to determine the orientation and magnitude of the principal horizontal stresses. The cross-dipole shear data from the DSI Dipole Shear Sonic lmager tool was processed to obtain the oriented fast and slow shear waves. This information was then used to determine the percentage of shear-wave anisotropy, azimuthal rock mechanical properties, the direction and magnitude of the in-situ earth stresses, and the orientation of fractures. Two major zones of anisotropy~ were identified. Zone I showed significant shear-wave anisotropy with a trend along North-South. consistent with the Nubia stress trend. The upper interval of this zone was detected as an open fracture system, using the shear-wave anisotropy data in conjunction with the Stoneley-wave fluid-mobility evaluation. This interval was subsequently perforated and produced hydrocarbons. Zone II exhibited an anisotropy azimuth trending along a NW-SE direction, consistent with the known tectonic regime of the Gulf of Suez stress trend (Clysmic fault trend). The stress and fracture information obtained from the dipole shear sonic anisotropy data was subsequently used in the planning of deviated wells, and for fracture treatments. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER Y WL AND LWD SONIC LOGGING OF SOFT FORMATIONS - A CASE STUDY FROM A DEEP-WATER, FRONTIER-AREA FIELD Finn O. Sollie Norsk Hydro ASA, Oslo, Norway. ABSTRACT Accurate seismic interpretation of lithology and porefill relies on good well ties and hence on good quality sonic and density logs. This study examines the Wireline (WL) and Logging While Drilling (LWD) log responses of shallow and soft formations found in the “Ormen Lange” blocks off mid-Norway. Prior to drilling the discovery well in 1997, the geological well prognosis described uncommon formations compared with, for example, North Sea basins. Special deep-water-location formations, such as gas hydrates and oozes, were predicted. Minor indications of gas hydrates were observed. However, ooze-like formations several hundred metres thick with porosities approaching 6O pu, and high pore pressures, were encountered. Problems were experienced in both Hydro operated wells with regard to hole stability and the carrying out of normal open-hole (OH) WL logging. Modern WL sonic logging includes the simultaneous recording of monopole- and dipole-source generated waveforms. Monopole source technology is traditionally used for the detection of compressional and shear slownesses in fast formations, whereas the dipole source allows the determination of shear slownesses in softer formations. Examples illustrate the shortcomings of monopole-based techniques in measurement of compressional slownesses in very soft formations and the advantages of the dipole based methods for the determination of both compressional and shear slownesses in these conditions, even in cased hole. Data from OH, and cased-hole (CH) logging are shown. The quality of OH logged shear slownesses in soft formations are examined and assessed. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER Z SATURATION MODELING AT THE WELL LOG SCALE USING PETROPHYSICAL ROCK TYPES AND A CLASSIC NON-RESISTIVITY BASED METHOD Gary W. Gunter, Charles R, Smart, Mike A, Miller, and Joe M. Finneran BPAmoco ABSTRACT This paper demonstrates how a detailed saturation profile can be constructed that is rooted in capillary pressure theory as opposed to being resistivity based. A case history is utilized. This method ascertains the relationship between petrophysical rock types and capillary pressure. The product is a water saturation profile that is determined at the well log scale. This product is an outcome of making relatively small modifications to the classic Leverett J procedure. The method iteratively integrates petrophysical rock type and rock properties along with permeability/porosity ratio (K/PHI). The final outcome is a powerful interpretation tool. Applying this knowledge enhances business value in exploitation or exploration activities. This method can be automated either through an algorithm or with a spreadsheet. The “Modified J Saturation” method seamlessly integrates routine core analysis, capillary pressure data, formation evaluation, and petrophysical rock types into an interactive graphical method for predicting saturation. As an additional advantage it provides a quick, user-friendly tool for performing various sensitivity analyses and quantifying the uncertainties in saturation prediction, hydrocarbon column length and free water level. This paper discusses data collection, data preparation, and process computations. The example is from a carbonate reservoir with complicated pore geometry. The predicted “Modified 3 Saturation” shows an excellent match with the “ground truth” data of Dean Stark initial water saturation. Laboratory data was collected on 20 carefully selected core plugs. The resulting measured data is the basis for the saturation prediction. The complex pore geometry can be seen in the various capillary pressure mediums used to present the results. This reservoir shows a bimodal pore system with significant micro-porosity. The poster session will illustrate this method in other reservoir systems. In the poster cases the “Modified I Saturation” is used to evaluate the hydrocarbon column length and system dynamics. These examples show how important it is to have an integrated perspective of the petrophysical properties in order to accomplish accurate formation evaluation. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER AA PERMEABILITY CALCULATIONS FROM CLUSTERED ELECTROFACIES, A CASE STUDY IN LAKE MARACAIBO, VENEZUELA Robert Y Elphick Schlumberger William Ray Moore Schlumberger ABSTRACT Two new interpretation techniques have been developed for the purpose of providing more accurate permeability information. The first is an improved approach to determining ElectroFacies using a k-means clustering algorithm, and the second involves using separate permeability / porosity relationships for each ElectroFacies to be used for calculating permeability curves. The study for which these techniques were developed, included wells drilled over the last forty to fifty years with well logs that cover a wide vintage. The earlier ES and IES logs needed to be interpreted along with more modem log suites in order to provide sufficient data for detailed reservoir descriptions suitable for reservoir simulation. The techniques developed proved to be very successful in dealing with the variations in log types and vintages. The clustering technique used was based on log characteristics and the ElectroFacies were determined using a kmeans method. Alternative clustering techniques are described and compared in the paper. The derived ElectroFacies were then compared to core data, particularly to determine “Rock Types” with identifiable fluid flow characteristics. Porosity / Permeability relationships were then assigned to the various ElectroFacies and used to calculate permeabilities during the log analysis. Core and log data are presented for a number of wells to show the success of the method over a variety ci fluvial to shoreface clastic systems with variations in diagenetic overprints. The data presented are from the Alto de Ceuta Field in Lake Maracaibo, Venezuela. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER BB THE USE OF NMR LOGGING FOR GAS RESERVOIR CHARACTERISATION IN A NORTH SEA PLAY WITH COMPLEX AND VARIABLE ROCK PROPERTIES A. Salvignol-Bellando Nederlandse Aardolie Maatschappij B. V Y. Volokitin, W. Slijkerman Shell International E&P RTS R. Bonnie Nederlandse Aardolie Maatwhappij B. V ABSTRACT In 1997 MRIL-C was run in multi-acquisition mode to improve reservoir characterisation in a North Sea gas field. In the absence of core data, reservoir characterisation based on conventional logs only is extremely difficult in this field, due to the variability of matrix properties. NMR logs, which provide “matrix independent” measurements, improve the interpretation of porosity and fluid content compared to conventional logs. The reservoir quality in this clastic play is controlled mainly by the degree of diagenesis. Halite plugging in particular deteriorates reservoir quality considerably. Porosities in the field vary typically between 2 and 30% BV, salt contents range between 0 and 40% BV. With no quantitative information on halite content, the use of conventional logs in a standard deterministic approach for reservoir characterisation, will provide highly unreliable results. MRIL-C was run in a well drilled with OBM over the gas zone and across the GWC. Dual wait time data, diffusionand BVI-data were acquired in three passes. Shell’s proprietary processing software provided, independent from any other log data, porosity, permeability, fluid saturation, the volume of irreducible water (BVI) and indicated the location of the FWL estimate. The presence of the various fluids in the pore space (water, OBM filtrate and gas) and the effects of the variable matrix properties on reservoir quality could clearly be seen in the NMR results. Based on the NMR evaluation, different reservoir intervals and theft properties were identified: • High permeability and porosity in line with the evaluation from conventional logs were calculated in the halite free gas interval. • The large variations in porosity and permeability, related to variations in halite content were quantified in the cemented interval. • A GWC in line with the FWL identified from pressure measurements could be determined, with significant amounts of free water and no gas in the bottom part of the well. • In line with core descriptions from other wells in the field across the GWC, the NMR results also suggested that the diagenesis is different in the water zone as compared to the gas zone. • Significant amounts of invaded OBM filtrate were only observed in the lower permeability gas zones and in the water bearing interval. The absence of invasion in the high permeability gas interval can be explained by the high mobility of the gas. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER CC AN IMPROVED NMR TOOL DESIGN FOR FASTER LOGGING D. McKeon, C. Cao Mirth, R. Freedman, R. Harris, D. Willis, D. Davies, G. Gubelin, R. Oldigs, and M. Hurlimann Schlumberger ABSTRACT An important limitation of pulsed nuclear magnetic resonance (NMR) logging tools has been the slow logging speed required to acquire full NMR data in reservoir rocks, usually between 200 and 900 ft/hr. To overcome this limitation, a prototype NMR tool has been developed that acquires full NMR data at logging speeds to 2400 ft/hr, a speed that is comparable to that of nuclear wireline tools and three to five times faster than previous-generation NMR tools. These improved logging speeds are possible because of advances in NMR technology: the use of long prepolarization magnets and a new method of acquiring CPMG measurements. A new acquisition sequence and processing algorithm has also been developed that improves the precision of the total porosity measurement and the T2 sensitivity limit. Although developed in conjunction with the new prototype, the new acquisition sequence and processing algorithm can be applied to the previous generation of NMR tools. This paper gives a brief history of NMR measurements, their processing and interpretation—followed by a description of the new prototype and processing algorithm and the field test program. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER DD THEORY AND OPERATION OF A NEW, MULTI-VOLUME, NMR LOGGING SYSTEM M. G. Prammer, I. Bouton, RN. Chandler, and ED. Drack NUMAR Corporation, A Halliburton Company ABSTRACT A recently introduced nuclear magnetic resonance (NMR) logging tool offers single-pass operation combined with high logging speeds. The streamlined operation of the new tool reduces overhead in job planning and facilitates automated quick-look interpretation models for delivery at the well site. The new device can acquire data from nine distinct measurement volumes. Extraordinary gains in logging efficiency are possible because logging speed can be scaled proportional to total measurement volume, which has been increased threefold from the previous generation of Nl~1R tools. Alternatively, the measurement volumes can be utilized independently of each other with different multi-wait time, multi-echo time sequences proceeding in parallel. Logging operations are further enhanced by • T1-independence. The tool can fully polarize fluids with longitudinal relaxation times (T1) ranging from 0 to 4 seconds at logging speeds of up to 24 ft/min. Polarization corrections or speed adjustments are not required. • Standardized tool setup and data processing. A single measurement sequence can cover the majority of NMR logging applications. This sequence yields total, effective, clay, bound-fluid, and free-fluid porosities; gas detection; and direct hydrocarbon typing. This paper is a practical guide to understanding and using the measurements made by the new device. The shape of the magnetic field and the dimensions and properties of all measurement shells are discussed. Tool calibrations are shown to be valid both for individual measurement volumes as well as for averages across measurements. Field data are used to demonstrate that data from the new tool are compatible with existing interpretation strategies. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER EE ANOMALOUS NMR RESPONSES IN HIGHLY PERMEABLE SANDSTONE RESERVOIRS: A CASE STUDY Joao de D. S. Nascimento and Paulo S. Denicol PETROBRAS, Rio de Janeiro - Brazil ABSTRACT Two field examples are presented of NMR logs in sandstone reservoirs with high porosity and permeability. The NMR logs show a total of 160 meters of anomalous porosity and T2 distribution response, when compared to conventional logs and formation test results. The NMR porosity is approximately 12 p.u. lower than the density porosity using a 2.65 g/cc grain density. The T2 log mean distribution falls almost entirely below the 33 ms cutoff limit. Consequently these NMR responses would normally indicate the predominance of capillary bound water porosity with little free fluid porosity, suggesting poor reservoir quality. This is in contrast to high resistivities and low shaliness of the sandstones in the study area, which normally suggest good permeability. Indeed, on a drill stem test (DST) the well flowed at the rate of 3,000 bbl/day of oil with a measured permeability of 6 Darcies. In the two examples, the theoretical resistivity, obtained from the relationship between the free fluid porosity and the total NMR porosity, agrees very well with the Rxo measurement, but conversely is 20 times lower than Rt. The NMR measurement may be related to some kind of effect within the flushed zone. Furthermore, the DST interpretation indicates the formation is highly damaged, with the damage ratio varying from 17 to 6. Consequently, mud solids within the reservoir pore space of the flushed zone causes the reduced NMR porosity and the T2 distribution response. In conclusion, it is inadvisable to use the NMR as a replacement for the density and neutron logs in reservoirs where mud solid invasion may occur. This near wellbore effect is almost entirely within the depth of investigation of the NMR tool. Evidence will also be presented that in reservoirs where mud solids invade the formation, the density and neutron logs can be affected, although to a lesser extent. The shallower the depth of investigation of a logging tool the higher the influence of the damage zone. As a result, mud solid invasion results in an underestimation of the volume of oil in place. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER FF RESTRICTED DIFFUSION AND INTERNAL FIELD GRADIENTS Matthias Appel, J. Justin Freeman, Rod B. Perkins Shell E&P Technology Company, Houston, TX, USA Jan P. Hofman Shell International E&P RTS, Riswijk, The Netherlands ABSTRACT Nuclear Magnetic Resonance (NMR) is typically assumed to be lithology-independent since the pore fluids are the only source of the measured signal. However, due to interactions of the pore fluids with rock surfaces, the rock matrix can significantly influence the fluid response. Although these interactions may seriously complicate the interpretation of the data, they also provide useful information about the pore structure of the rock. Two primary mechanisms that complicate interpretation of NMR data (wireline logs or core analyses) are: (1) strong internal gradients; and (2) restricted diffusion. Under laboratory conditions, measured in a homogeneous background magnetic field, internal magnetic field gradients can be calculated from the increase of the transverse relaxation rate of pore fluids with increasing interecho time. Whereas a linear relation between the increase of the relaxation rate and the squared inter-echo time exists for the case of unrestricted molecular diffusion of the pore fluid, deviations from this linear relation can be expected due to effects of restricted diffusion and a pore size distribution causing a distribution of magnetic field gradients. We report results of laboratory NMR experiments on core samples taken from the Gulf of Mexico and from the Far East. For these samples, the calculation of internal magnetic field gradients was complicated by the effects of restricted diffusion. Whereas the reduction of molecular mobility due to restricted diffusion was sufficient to explain the measured data in the range of inter-echo times interesting for standard NMR logging procedures, an additional reduction of the effective internal magnetic field gradient had to be assumed to interpret the measured data for a wider range of inter-echo times. This observation was supported by separate Pulsed Field Gradient NMR diffusion measurements using pulse sequences that minimize the effects of internal gradients. Our measurements confirm recent theories published in the open literature and may be of significant importance for the interpretation of wireline NMR data that is based on the diffusion of pore fluids over extended diffusion times. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER GG COMBINING NMR AND FORMATION TESTER DATA FOR OPTIMUM HYDROCARBON TYPING, PERMEABILITY AND PRODUCIBILITY ESTIMATION C. Castelijns, R. Badry, E. Decoster, and C. Hyde Schlumberger ABSTRACT This paper describes the joint interpretation of NMR and Formation Tester data. We show that combining the data from these two different measurements leads to a robust hydrocarbon typing. The two measurements complement each other. A primary application is the use of the NMR log for the optimization of the Formation Tester acquisition. Pressure testing and sampling depths are chosen based on the permeability estimate from the NMR measurement. Field examples are included showing the identification of oil, gas and tar zones in diverse environments. In one well the data is used to evaluate changing oil properties. The combination of data allows for the evaluation of the production potential of different zones. One of the examples demonstrates how the comparison of NMR log derived permeability with Formation Tester drawdown mobility leads to an improved determination of the NMR T2 cut-off parameter. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER HH USING NMR AND ELECTRICAL LOGS FOR ENHANCED EVALUATION OF PRODUCIBILITY AND HYDROCARBON RESERVES IN GAS RESERVOIRS WITH HIGH IRREDUCIBLE WATER SATURATION S.M. Hussein and S. Hassan Petrobel T. Klimentos, A. Boyd and A. Zeid Schlumberger ABSTRACT Often. in the Nile-Delta sandstone formations. the space surrounding the pores can be occupied by a variety of clay minerals, such as. chlorite. smectite. illite and kaolinite. Water attaches itself to the surfaces of clay particles, and since clays have large surface-to-volume ratios, the relative volume of clay bound water is large. Due to their extremely high specific-surface area and cation exchange capacity values both water saturation (using conventional saturation equations) and permeability, may be miscalculated. In this paper an integrated NMR formation evaluation study of a Nile-Delta gas reservoir with high Swir is presented. The NMR data was acquired using the CMR Combinable Magnetic Resonance Tool. The main scope of this study was to use the CMR-APS (neutron)-LDT (Density) and Electrical resistivity logs in order to estimate accurately hydrocarbon reserves, effective and total porosity, type of pore-fluids and permeability. The CMR permeability estimates after having been corrected for the gas-effect compared favorably to the MDT derived mobility values. Moreover, the CMR data of bound-fluid volume and irreducible water saturation were used as inputs in the ELAN evaluation. Such a combination of CMR and ELAN provided us with reliable and accurate answers regarding lithology. effective and total porosity, bound and movable fluids, Sw. hydrocarbon reserves and producibility. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER II BOUND FLUID ESTIMATES Peter Speier, Steve Crary, Robert L. Kleinberg, and Charles Flaum Schlumberger Oilfield Services ABSTRACT One of the most important magnetic resonance (MR) parameters measured downhole is bound fluid volume (BFV). BFV represents the water contained in small pores and trapped by capillary forces. BFV is used to improve resistivity interpretation and to estimate the permeability of the formation. In standard wireline MR logging, a long echo train is recorded and transformed into a distribution of transverse relaxation times (T2 spectrum). BFV is estimated from the fast decaying part of the T2 spectrum using a free-fluid cutoff, fixed or tapered cutoffs are employed, depending on the petrophysical situation. Lateral tool motion during the MR measurement significantly alters the T2 spectra. Motion during the long echo train acquisition acts as an additional relaxation effect with the result that even when the total porosity estimate is correct, the estimated BFV can be too high. This effect has been demonstrated with simulations that show the lateral motion effect depends on the detailed trajectory of the tool in the hole. This complex dependence makes it difficult to correct BFV for the motion effect. We have developed an acquisition technique that minimizes motion effects. BFV is obtained by measuring the signal amplitude of incompletely polarized nuclear spins. The effect of this partial polarization is compared to the application of a tapered cutoff to the relaxation time distributions. The partial polarization response can be matched to a specific tapered cutoff function by adjusting the wait time of the measurement. Then, only the bound fluid portion of the signal is polarized during the acquisition and, consequently, the amplitude of the MR signal represents the BFV of the formation. The amplitude of the signal (i.e. apparent porosity) can be determined from a few echoes so that a traditional T2 analysis, and thus the acquisition of long echo trains, is not required to obtain BFV. This novel and efficient approach to the acquisition I BFV using only partial polarization of the MR signal has wide utility. The short wait time allows rapid acquisition of high-resolution BFV data with improved signal-to-noise ratio. The technique is also useful for acquiring BFV in situations with large lateral tool motion, such as while drilling. Comparisons between core measurements, and NMR data acquired downhole with a wireline tool demonstrate that with a properly chosen wait time results similar to a tapered cutoff are obtained. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER JJ PULSED NEUTRON DENSITY MEASUREMENTS: MODELING THE DEPTH OF INVESTIGATION AND CASED-HOLE WELLBORE UNCERTAINTIES Richard C. Odom, Shawn M. Bailey Computalog U.S.A., Inc Robert D. Wilson San Juan Scientific Michael P. Archer Chevron, U.S.A., Inc. ABSTRACT The development of a density-based porosity derived from the transport of inelastic gamma rays created by a pulsedneutron system has been discussed in several recent publications. Accurately applying this technology to cased reservoirs will require an appreciation of such measurement dynamics as the depth of investigation and the effect of cement-sheath thickness. Using Monte Carlo computer simulations of radiation transport, the depth of investigation of the pulsed neutron measurement can be modeled for several formation and borehole conditions. The computer model is structured in a high-porosity sandstone formation segregated into concentric zones similar to the benchmark experiments that established the depth of investigation for the open-hole density measurement. To determine the depth of investigation for the cased-hole measurement, the pulsed-neutron model uses 5 ¼-inch casing centered and cemented in an 8-inch borehole. Modeling results indicate that the cased-hole, pulsed-neutron density measurement has roughly twice the penetration into the formation as compared to the open-hole measurement where the pad is in contact with the formation surface, Computer simulations of the pulsed-neutron density were extended to model how the depth of investigation changes with formation density. Further, a series of computer simulations were run with various (3/4 and 1 3/4-inch) cementsheath thicknesses to assess this uncertainty in the cased-hole measurement. To test the validity of results from the computer simulations, open-hole density and caliper measurements were compared with the cased-hole density measurement in the same well. Along with the ability to measure the density behind casing, the deeper penetration of the pulsed-neutron measurement and the centric radiation pattern of the through-tubing sonde mean that minor borehole rugosity is not a critical parameter. Well log examples demonstrate how the cased-hole measurement can be applied to check the quality when borehole rugosity has affected the openhole density porosity. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER KK EFFECTIVE PULSED-NEUTRON LOGGING IN BOTH TUBING AND CASING FOR BROWN FIELDS M L Sanni, S McFadden Shell Pet Dev Co Nigeria S Kimminau Shell International P Wanjau, L. Silipigno Schlumberger Sugar Land B. Roscoe Schlumberger Ridgefield ABSTRACT Reliable logging in both tubing and casing in Niger-Delta oilfields has allowed bypassed oil to be found and produced for very low costs, typically less than 1 dollar per barrel. Fields in the Niger-Delta often consist of a stack of many “completable” reservoirs, each seldom more than a few hundreds of feet thick, but together stretching over several thousands of feet vertically. In most completions many potential reservoir intervals are behind both casing or liner and at least one string of tubing. Cased-hole logging provides information on fluid contact movements and thus reservoir connectivity during the production life; it has identified bypassed oil previously thought to have been developed by existing producers and, conversely, unperforated sands being drained via neighboring reservoirs. Carbon-oxygen logging through both casing and tubing became possible in 1991 with the introduction of a 1 11/16 in. OD tool. However, the tool was originally intended t0 be conveyed through tubing on its way to an interval with a single casing or liner below the tubing shoe, and it had not yet been characterized for logging inside two steel tubulars. Initial field trials inside two tubulars concentrated on establishing fluid contacts on an empirical basis, while a research program provided both Monte-Carlo modeling data and a laboratory characterization database. Laboratory data have been acquired in 68 conditions, characterizing tubing sizes of 2 3/8 and 3.5 in. within 8.5 and 12 in. boreholes. A software job planning tool calculates carbon-oxygen yields and statistical precision under a variety of completion conditions, allowing the optimal data acquisition program to be planned. Interpolation between different completion cases is not automated because of the sparsity of the measurements and the complexity of the completion geometry, but a library of well-characterized cases allows inter-comparison. Field results have been encouraging, with more than 70 percent of cases clearly indicating fluid contacts that have subsequently been proven by recompletion. The new characterizations and their inclusion in the planning database are currently under field test. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER LL WATER SATURATION IN BREAKTHROUGH INTERVALS, EKOFISK FIELD, NORWAY P.E. Van DeVerg, J.J. Howard, M. Fidan and R.L. Nesvold Phillips Petroleum Co. Norway ABSTRACT Resistivity based water saturation estimates in Ekofisk Field waterflood breakthrough intervals are affected by changes in temperature, water resistivity, and Archie parameters introduced by the injection of cold seawater into the reservoir. Reservoir temperatures in wells adjacent to seawater injectors can drop below 100° F with strong negative geothermal gradients associated with breakthrough intervals. Borehole temperatures are elevated compared to reservoir temperatures in these zones because thermal energy is provided by the overburden to the drilling mud. It is observed that well intervals at early breakthrough stages do not have any negative temperature gradients, so temperature does not always indicate breakthrough. Radial temperature distributions were generated from wireline temperature logs run at different times. Cased-hole temperature measured 7-20 days after mud circulation ceased was used as the equilibrium formation temperature. The choice of water resistivity is between original formation water salinity, injected seawater, and some intermediate mixture. When produced water information is not available, the decision to use injection water salinity is based on the appearance of a negative temperature anomaly, This criterion is based on the observation that during waterflood the formation water bank precedes any thermal effects created by the injection of colder water. Laboratory experiments indicate that Archie saturation exponent at the imbibition end-point is greater than that measured at the drainage end-point. The resistivity index hysteresis associated with changing distributions of the conducting and insulating fluid phases in the pore space underlies the problem of what Archie parameters to use when evaluating water saturations at intermediate waterflood conditions. An empirical method that accounts for this variable saturation exponent was evaluated. This method used a constant exponent value with linear offset from RI=1 at fully water-saturated conditions. All saturation models generate similar end-point residual oil saturations, as does the Archie equation with imbibition end-point saturation exponents. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER MM AN EVALUATION OF HORIZONTAL WELL PRODUCTION LOGGING TECHNIQUES - NORTH SLOPE, ALASKA R. J. North, Schlumberger, D. W. Warner, and J. L. Brady ARCO Alaska Inc. ABSTRACT Two new, independent production logging techniques have been used to determine downhole flow profiles in horizontal wells on the North Slope of Alaska. Each method presents certain advantages and drawbacks in operational procedures and answers provided. This paper evaluates and compares results from the two techniques to determine the best applications for each. The first technique uses a traditional production logging string in combination with pulsed neutron log (PNL) capture data. Specifically this technique combines temperature, pressure and spinner measurements with PNL data collected while sea-water and borax water are pumped into the well. The resulting change in the formation sigma data provides a technique to determine the liquid productivity of the well. The liquid productivity data are combined with the traditional sensor data to determine the liquid and gas production profile. The second technique combines the traditional production logs with oil and water velocity measurements and threephase holdup from PNL inelastic data. Additional input is supplied from an array of local electrical probes that provide an image of phase distribution in the borehole. The individual phase velocities and holdups are combined with the traditional sensor data to determine the down-hole flow profile of each phase. Both techniques provide answers in the difficult environment associated with the tortuous flow paths in horizontal wells. Both methods generate liquid and gas production profiles. The velocity/holdup technique provides the additional determination of individual oil and water rates from the total liquid profile. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER NN THIN BED RESERVOIR CHARACTERIZATION IN THE BALDER GAS CAP, NORWEGIAN NORTH SEA F. M. Haynes, D. Bergslien, O. M. Burtz Esso Norge AS M. S. Munkholm Baker Atlas Geosciences ABSTRACT The gas cap of the Balder Field is contained in thinly bedded, deepwater Tertiary sands of variable facies and reservoir quality. Although the gas cap has now been penetrated by numerous exploration and development wells, uncertainty remains as to its size, distribution, and ultimate effect on reservoir performance due to internal geologic complexity (thin beds, injected sands, multiple facies, and faulting). With a depletion plan involving both gas and water injection, geologic/reservoir characterization of the complex gas cap is considered critical for prudent resource management during depletion. To this goal, resistivity modelling and core-based thin bed reservoir description from the first phase of development drilling have been integrated with seismic attribute mapping. Detailed core descriptions, core permeability and grain size distribution data delineate six facies and help in distinguishing laterally continuous massive and laminated sands from potentially non-connected injection sands and non-reservoir quality siltstones and tuffs. Volumetric assessment of the gas-cap, thin sand resource has been enhanced by 1D forward modelling of induction log response using RtBAN. After defining beds and facies with core and high resolution log data, the deep induction log (AIT-H:AHF6O) was modelled as a 6FF40 device. Because many of the beds were thinner than 6FF40 resolution, the modelling is considered to provide a lower bound on Rt. However, for most beds this model-based Rt is significantly higher that provided by one foot resolution shallow resistivity data (AHOl0), and is thought to be the best available estimate of true formation resistivity. Sensitivities in STOOIP were accessed with multiple Rt earth models which can later be tested against production results. Sand thickness models constrained by this log- and core-based petrophysical analysis were used to build impedance seismic synthetic sections from which seismic attributes could be extracted and calibrated. The model-based attribute calibration was then applied to the real impedance 3D cube permitting sand thickness to be mapped and reservoir geology to be modelled with significantly more detail than previously possible. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER OO A MODEL FOR HYDROCARBON SATURATION DETERMINATION FROM AN ORTHOGONAL TENSOR RELATIONSHIP IN THINLY LAMINATED ANISOTROPIC RESERVOIRS R.A. Mollison Baker Atlas J.S. Scion Joanneum Research O.N. Fanini, B. Kriegshäuser Baker Atlas W.H. Meyer Baker Hughes INTEQ PlC. Gupta Roorkee University, India ABSTRACT The macroscopic effect of a laminated sand-shale sequence with different resistivities results in electrical anisotropy. This effect transforms the classic scalar, low contrast laminated shaly sand problem into an orthogonal tensor relationship with the addition of a vertical resistivity parameter to the model. The horizontal tensor component of the saturation model demonstrates that the laminar shale component is a purely additive conductivity term and the conventional bulk formation resistivity (Rt) is the inverse weighted sum of the sand and shale resistivities. The vertical tensor component represents a series resistivity relationship where vertical Rt is the weighted sum of the sand and shale resistivities, Combining both vertical and horizontal tensor components yields a system of coupled, nonlinear equations that provide a more robust solution for hydrocarbon saturation than a scalar model. A sensitivity analysis as a function of shale and sand conductivity ratio and shale volume suggests a more confident hydrocarbon saturation estimation as a result of using both horizontal and vertical formation resistivities. Expansion of the shale conductivity term within the saturation equation corrects for dispersed clay conductivity using Qv in this two-step tensor model approach and results in an easily implemented Patchett-Herrick model. This new solution was applied to a petrophysical model derived from MWD gamma ray and propagation resistivity data from a Gulf Coast well. Additional model parameters for shale and sand were estimated from adjacent well wireline data. Anisotropy from MPRTM data was used in the direct orthogonal tensor solution. This study concludes that in laminated shaly sands, the new orthogonal tensor model provides an improved solution for hydrocarbon saturation as a function of resistivity anisotropy when compared to traditional models. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER PP REVEALING THE PETROPHYSICAL PROPERTIES OF A THIN-BEDDED ROCK IN A NORWEGIAN SEA RESERVOIR BY TIHE USE OF LOGS, CORE, AND MINIPERM DATA Lars Helge Flolo Statoil Wendell P. Menard and Ken W. Weissenburger Conoco Jostein M. Kjerefjord and Dag M. Arnesen Statoil ABSTRACT An unconventional thin-bed analysis based on logs, core and miniperm data was needed to calculate the petrophysical properties of a reservoir under development in the Norwegian Sea. More than half of the reservoir section under investigation is composed of heterolithic facies: thinly interbedded sandstone and mudstone layers from less than one to several centimeters in thickness and of variable quality. By using miniperm measurements with I-cm spacing on slabbed core, it was possible to resolve the properties of the rock far below the vertical resolution of conventional wireline logs and relate them to the bulk log measurements. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER QQ IDENTIFYING AND ESTIMATING FORMATION STRESS FROM BOREHOLE MONOPOLE AND CROSS-DIPOLE ACOUSTIC MEASUREMENTS X. M. Tang, N. Y. Cheng, and Arthur C. H. Cheng Baker Atlas, Houston, TX ABSTRACT This paper describes an effective method for determining the formation stress field using borehole acoustic monopole and cross-dipole measurements. This method is based on a stress-velocity relation. Testing this relation using published laboratory data not only validates the relation but also determines the essential parameters required by the relation. Application of the stress-velocity relation to the borehole environment results in a theory that lays the foundation for determining formation stress from borehole acoustic monopole and cross-dipole measurements. The theory predicts that the two principal stresses perpendicular to the borehole produce a splitting in the crossdipole measured dipole-shear wave data. The stresses also produce an even greater splitting in the monopole-shear wave data. The latter splitting provides a criterion to verify whether or not the determined anisotropy is caused by the stress field. Thus, by combining the two measurements, one can detect the stress-induced shear-wave anisotropy and estimate both the orientation of, and the difference between, the two principal stresses. Furthermore, using a cross-dipole measurement, one can define a stress indicator that is related to rock deformation and formation shear stress magnitude. This indicator may be used to indicate impending rock failure or yield, thus being potentially useful in sand control and borehole stability applications. A field data example demonstrates the application of the method. This data set shows that stress-induced shear wave velocity change and anisotropy are significant in sand but negligible in shale, consistent with the laboratory testing results. Applying the proposed method to the acoustic logging data yields the maximum stress orientation and the shear stress magnitude. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER RR STRESS-INDUCED DIPOLE ANISOTROPY: THEORY, EXPERIMENT AND FIELD DATA T.J. Plona, BK. Sinha, M.R. Kane, K.W. Winkler, and B. Frignet Schlumberger ABSTRACT Dipole sonic logging tools provide the capability to measure formation anisotropy. Anisotropy is important in seismic AVO applications. However, one needs to distinguish between two forms of anisotropy: intrinsic or stressinduced in order to fully utilize anisotropy information. Intrinsic anisotropy is caused by bedding, microstructure, or aligned fractures, Stress-induced anisotropy results from the application of tectonic or overburden stresses. Stressinduced anisotropy and stress direction determination is important for geomechanics applications including oriented perforation for optimized fracturing in both hard and soft formations and placement of horizontal wells for maximum production. We have developed a new acoustic method which allows us to identify stress-induced anisotropy (and thus, stress directions) even in the absence of breakouts or hole deformations. We have shown theoretically that stress-induced dipole anisotropy exhibits a characteristic crossover in frequency of the two dipole fiexural wave slowness dispersion curves polarized parallel and normal to the far-field uniaxial compressive stress direction. This crossover phenomenon is a result of borehole stress concentrations. At low frequencies, the flexural wave field penetrates deep into the formation and senses the far-field stress. At higher frequencies, the flexural waves are confined to the nearwellbore region and are dominated by the stress concentrations near the borehole. In contrast, for intrinsic anisotropy, the two flexural mode dispersion curves do not cross as a function of frequency for a transversely isotropic medium. Hence, the existence of a dispersion crossover can be employed in a new technique for distinguishing stress-induced from intrinsic anisotropy. We have conducted laboratory measurements on a large block of Berea sandstone subjected to uniaxial compressive stresses up to 5 MPa. Waveforms are recorded for dipole polarizations both parallel and normal to the stress direction and then processed by a modified matrix pencil algorithm to yield flexural dispersion curves. The measured dipole flexural dispersion curves indicate that the stress-induced crossover behavior occurs with as little as 0.5 MPa applied stress. Theory and experiments are in excellent agreement. The fast shear direction is the maximum stress direction. We have investigated dipole dispersion curves in one horizontal well using a cross-dipole sonic logging tool. There are both isotropic and anisotropic sections. We showed anisotropic examples where dipole dispersion curves crossover (indicating stress-induced anisotropy) which are consistent with our theoretical models. Dispersion analysis offers excellent opportunities to derive additional information from sonic waveforms. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER SS HORIZONTAL STRESS ORIENTATION ANALYSIS USING TILE CROSS-DIPOLE ACOUSTIC LOG IN THE ELDFISK FIELD Doug Patterson Baker Atlas Geir Skjong and James M Wade Phillips Petroleum Co. Norway ABSTRACT Horizontal stress orientation is an important aspect of formation evaluation and a governing factor in the optimization of the development and drainage of a reservoir. This is especially true if the formation is hydraulically stimulated or is naturally fractured and stress governs the directional aspect of permeability and hence production. The standard methods for determining stress include the evaluation of borehole breakout from image logs, anelastic stress relaxation from whole core, and the monitoring of surface displacement through the use of surface tiltmeters, Increasingly, the use of cross-dipole acoustics is demonstrating its value in assessing the horizontal. stress orientation. The ability of the cross-dipole acoustic log to identify large-scale fractures and fracture networks has been well documented. On a smaller scale, a naturally fractured reservoir will respond to azimuthal stress in like fashion. This response will induce a shear-wave anisotropy, which can be measured and evaluated by cross-dipole analysis. This paper presents a case history where cross-dipole measurements were made to determine the horizontal stress orientation of the Eldfisk Field in the North Sea. The study was conducted in conjunction with an extensive stress analysis done prior to a planned waterflood. The results are compared with borehole image logs that were run before and after micro-frac tests to record the direction of the induced fractures. In addition, whole core was taken across the section and anelastic strain recovery measurements were made. These differing methodologies yield the same results and demonstrate the value of cross-dipole acoustic measurements in determining the stress orientation in the Eldfisk Field. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER TT COMPARISON BETWEEN STONELEY, NMR, AND CORE-DERIVED PERMEABILITIES T.W. Geerits ,M. Altunbay , X.M. Tang Baker Atlas Logging Services, Houston, TX K.A. Lehne and O. Kelder Statoil, Stavanger, Norway ABSTRACT Formation permeability is the key to reservoir characterization and management. Wireline acoustic and nuclear magnetic resonance (NMR) logging can now provide continuous permeability index profiles. With the proper calibration of formation fluid parameters for the acoustically obtained permeability index and petrophysical model parameters for the NMR obtained permeability index, we can now map the permeability index to laboratory derived permeabilities. In the Smørbukk field, Mid Norway, the reservoir is located at a depth from 3500 m to 5000 m. The reservoir quality is controlled by sedimentary facies variation, quartz cementation and primary sand quality, secondary porosity, and the presence of illite and grain coating chlorite. This makes the reservoir very heterogeneous, resulting in strong permeability changes on a vertical scale much smaller than the vertical resolution of NMR and Stoneley derived permeability. In this paper we will analyse three wells of the Smorbukk field and demonstrate that after applying proper vertical avenging to the core data, we obtain a good match between Stoneley, NMR and core permeability. However, one stratigraphic unit appears in all three wells, which was cored in two of these wells. This unit shows a consistent mismatch between core and NMR/Stoneley derived permeabilities. This mismatch will be explained by a “borehole damage” theory. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER UU CONTINUITY LOGGING EXPERIMENTS FOR RESERVOIR CHARACTERIZATION AT THE STRATTON GAS FIELD, SOUTHEAST TEXAS J.O. Parra Southwest Research Institute A.W. Gorody Universal Geoscience Consulting C.L. Hackert Southwest Research Institute D. Howlett and D. Melton Texaco E&P Technology ABSTRACT Guided seismic waves can be used to map reservoir structures between wells, with the low velocity beds acting as waveguides. The detectability of guided waves between wells is an indication of the continuity of the layered formation examined. Experiments were conducted between three wells at the Stratton gas field in southeast Texas to establish proof of concept for continuity logging in a producing field. Tertiary reservoirs in the Gulf Coast were targeted for the experiments because prior study had recognized significant reserve growth potential in these deposits. The Stratton gas field is located in Kleberg and Nueces Counties, approximately 30 miles southwest of Corpus Christi. This field was selected because it had been extremely well characterized. The optimum low velocity zones chosen for the experiments were selected on the basis of available sonic logs from one of the detector wells, and on the basis of along-spaced, through casing sonic log run on the source well. Practical safety considerations made it necessary to isolate the planned experiments from all producing intervals and perforations. This limited the intervals available for the experiments to the uppermost section of the Frio Formation. Three low velocity intervals were selected by integrating well logs, petrophysics, and geology, as well as computer modeling and signal analysis. Characteristic guided seismic signatures formed by head waves, leaky modes and normal modes are observed in the processed seismic data. The data analysis confirms that low resistivity shale markers are truly continuous stratigraphic markers. The results of the experiments and data analysis performed at the Stratton field demonstrate that each of the low velocity zones can be identified on all the well logs. Specifically, the analysis of the results demonstrates that the V2 shale interval is continuous between detector well 145 and source well 151, and between detector well 159 and source well 151, at the interwell spacings of 1730 ft. and 2744 ft., respectively. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER VV PETROPHYSICAL APPLICATIONS OF NEW SEISMIC-WHILE-DRILLING TECHNOLOGY IN DEEP WATER Greg Myers, Dave Goldberg, and Alex Meltser Lamont..Doherty Earth Observatory of Columbia University Erich Scholz Downhole Systems, Inc and the Ocean Drilling Program Leg 179 Shipboard Science party ABSTRACT During the Ocean Drilling Programs Leg 179, we recorded drill string vibration data to investigate the sub-seafloor environment as part of two Seismic-While-Drilling experiments in May, 1998. The Ocean Drilling Program’s drill ship the JOIDES Resolution drilled two holes in 714 m and 1660 m water depth in the Indian Ocean to conduct these experiments. To our knowledge, such measurements have never before been recorded in the Ocean Drilling Program or on other deep-water drilling rigs. By comparison of vertical and horizontal drill string acceleration with wireline logs and core data, variations in the petrophysical properties are correlated to drilling parameters. Drill string acceleration varies inversely with porosity from logs and core data which correlate to fracturing and lithologic changes in these examples. Determining petrophysical properties while drilling, therefore, may assist both log analysts and drillers in identifying lithologic contacts and sediment/rock interfaces even in cases where core and log data are not available. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER WW STRESS AND STRESS RELEASE EFFECTS ON ACOUSTIC VELOCITIES FROM CORES, LOGS AND SEISMICS Erling Fjaer SINTEF Petroleum Research, Trondheim, Norway Rune M. Holt NTNU Norwegian University of Science & Technology, Trondheim, Norway ABSTRACT A technique for producing synthetic sandstone under controlled stress conditions has been used to study stress and stress release effects on acoustic velocities. Laboratory experiments were performed to simulate seismic, log and core measurements. The results show that seismic studies may to some extent be used to monitor stress changes, however the relations between stress changes and changes in the seismic velocities will not be linear and will depend on the stress path. In the vicinity of a wellbore there will be a zone with reduced sonic velocities. This zone is thin enough that conventional long-spaced sonic tools will not be affected. If specially designed tools are used to log velocities of the damaged zone, the data may be interpreted to yield stress information. Unloaded cores will show a velocity anisotropy that reflects the forming stress state. Cores that have been loaded back to the in situ stress state will not have frilly recovered velocities, and the stress dependency of the velocities will be larger than in situ. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER XX RESERVOIR MONITORING WITH INTERWELL ELECTROMAGNETIC IMAGING Q. Zhou, A. C. Reisz Chevron Petroleum Technology Company P. Henshaw Chevron USA ABSTRACT One of the major tasks in reservoir monitoring is to determine where the injected fluid is and in which directions it is moving. Interwell electromagnetic (EM) imaging is an emerging technology for reservoir management, especially applicable to mapping and monitoring the fluid movement in the interwell space during enhanced oil recovery (FOR). Interwell EM differs from conventional logging in that it maps the interwell space instead of the small volume surrounding a borehole. The electrical conductivity (reciprocal of resistivity) imaged by the interwell EM technology is directly associated with the reservoir characterization parameters such as water saturation, salinity, and porosity. The interwell conductivity images can reveal the formation details in the previously inaccessible area between the monitoring wells. The conductivity images are especially sensitive to the variations in rock pore fluid (water or oil) and therefore differ from seismic velocity and attenuation measurements which are more sensitive to the physical property variations of the rock matrix. EM and seismic imaging methods can complement each other for improved reservoir monitoring. Significant progress has been made in both theory and instrumentation for interwell EM. For flood monitoring, conductivity images taken at different times during EOR can provide information about the movement of oil, water, and steam. A field experiment at a steam injection site of Chevron was conducted recently in cooperation with Lawrence Livermore National Lab (LLNL). The time-lapse measurements clearly show the progressing anomalous conductive volume which results from the injection. This paper concentrates on the application, physical understanding, and data processing of the method. It is found that resolution depends on the data sampling, spatial coverage, and operating frequency. In addition, constraints based on well logs and other geological information should be used whenever available in the inversion to optimize the image quality. From the theoretical and field experiments, interwell EM imaging is found to be a potentially valuable technology in reservoir monitoring. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER YY INTERPRETING MULTIARRAY INDUCTION LOGS IN DIFFICULT ENVIRONMENTS Torn Barber, Zlatko Sijercic, Harold Darling, and Xiaowen Wu Schlumberger ABSTRACT There are several environments that can potentially create trouble for induction tools. The most important is a well with very saline mud. In addition to making the borehole signal large and difficult to correct, borehole rugosity becomes much more problematic. Saline mud also means conductive invasion in hydrocarbon-bearing formations, which can cause artifacts on high-resolution logs. Some drilling techniques produce periodic borehole rugosity, which in turn produces periodic noise on the logs. This is aggravated by very saline muds. Outside mud salinity issues, the most important environment is that of very resistive beds with very conductive shoulders. These formation contrasts can be outside the mathematical approximations used in the log formulation, causing the deep logs to read incorrectly. Determining Rt at high dip angles has also been problematic. All these difficult cases have, in the past, proved to be more than the processing or environmental corrections could handle, leaving the interpreter to guess at Rt. Although these effects may seem unrelated, the common thread is to flag potential environmental problems so that the log analyst will be aware of limitations or can choose alternative processing methods. In addition, successfully acquired logs require prior knowledge of the environment to make sure the tool is run properly. To that end we have developed a list of guidelines for anticipating problems in running the tools and obtaining the best performance in most situations. These guidelines are contained in a web-based job planner. Real-time environmental log-quality flags are now available to identify environmental problems, allowing the optimum resolution to be presented. In the case of periodic borehole rugosity, an algorithm has been developed to reduce the errors associated with this very difficult situation, as well as guidelines for minimizing the effect. Finally, alternative processing schemes have been developed that give robust logs at any contrast and any shoulder conductivity as well as al high dip angle. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER ZZ IDENTIFYING AND QUANTIFYING RESISTIVITY ANISOTROPY IN VERTICAL BOREHOLES T. Hagiwara and H. Zea Halliburton Energy Services ABSTRACT Resistivity anisotropy is not uncommon in earth formations, and the knowledge of its presence and degree can be valuable in petrophysical analysis, well planning, and rig-site safety. The knowledge gathered in vertical wells is important for identifying and evaluating thinly laminated formations, establishing reliable well-to-well correlations when some of the wells are highly deviated, estimating geopressures in deviated boreholes, and geosteering during drilling. Although resistivity anisotropy is readily observable and quantifiable with 2-MHz, logging-while-drilling (LWD), resistivity devices in highly deviated wells, its detection and assessment in vertical holes has been elusive until recently. A method has been developed to identify and estimate resistivity anisotropy in vertical holes. The method is derived from two-dimensional (2D) modeling of electric- and induction-log responses. From the difference between electricand induction-log resistivity measurements or between deep- and shallow-investigating, electric resistivity measurements, resistivity anisotropy can be estimated. The method is especially useful when it employs data from a high-resolution induction device and a shallow, focused-resistivity tool, which form a routinely run logging combination. This paper describes the new method, presents logging data that illustrate its application and support its results, and discusses its application to invaded formations. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER AAA A NOVEL INVERSION METHOD FOR INTERPRETATION OF A FOCUSED MULTISENSOR LWD LATEROLOG RESISTIVITY TOOL Qiming Li, John Rasmus, and Darrel Cannon Schlumberger ABSTRACT Multisensor and array logging-while-drilling (LWD) tools are becoming routinely available. These tools can provide sufficient information to compute accurate two-dimensional (2D) characterizations of the formation environment surrounding the borehole. However, reducing these measurements to petrophysical parameters requires full knowledge of the tool response and substantial computation. Such processing may require a significant amount of time after data acquisition. An alternative one-dimensional (lD) approach that is simple, robust, fast and accurate is highly desirable. This paper presents a novel approach to 1D inversion based on an intuitive equivalent circuit model that can accurately reproduce the response of a focused array laterolog tool. This approach is well suited to focused, high resolution measurements, which have minimal shoulder bed effects. The equivalent circuit model of these measurements offers a representation of the response to the borehole and invaded zone. The parameters of the model are obtained from finite-element (FE) modeling of the tool response with the tool centered in a circular borehole with a step invasion profile. Since the tool response is entirely characterized as a function of physical parameters, a conjugate-gradient method is employed to minimize the cost function in the inversion. This model permits simultaneous inversion for borehole diameter, invasion diameter, true formation resistivity and invaded zone resistivity. In certain instances, this same technique may be applied to azimuthal measurements to provide useful information on borehole shape and invasion profiles. The computed borehole shape can be useful for wellbore stability analysis and log quality control. The inversion algorithm is verified with synthetic logs from modeling and its robustness is proven with several applications from actual field logs from an azimuthal LWD array laterolog tool. It is well known that laterolog measurements can be affected even if the invasion is relatively shallow. These examples illustrate the need of computing an R1 for both shallow and deep depths of invasion. The technique of automatically solving for the borehole diameter leads to more accurate R, values in overgauge and oval boreholes. An example of using the computed borehole shape to characterize breakouts is also provided. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER BBB INVERSION OF RESISTIVITY LOGS - THE GLORY OF TURBO BOOSTING R.G. Hakvoort Shell International Exploration and Production B.V., The Netherlands ABSTRACT Classical resistivity inversion methods, based on parametric optimizations, are often too time consuming for real-life applications. In this paper a new and fast inversion method, called the “turbo boost” method, is described. The method inverts for both Rxo and Rt on the basis of a resistivity logging tool response in a pancake-layered formation intersected by an arbitrary well path. The new turbo boost inversion method consists of two steps. In the first step, bed boundaries are determined on the basis of the measured resistivity (laterolog) responses. A new bed boundary detection method is introduced, which includes essential environmental corrections for borehole diameter and apparent dip. The algorithm performs well, both for vertical and for deviated wells. Simulations show that the algorithm yields bed boundary locations with an overall accuracy better than 0.25 ft. This first step requires only a few seconds of computation time to process a complete log. In the second step of the inversion, the actual “turbo boosting” takes place. A resistivity model matching the data is constructed by iteratively “boosting” the resistivity values on the basis of the observed mismatch between the measured and simulated laterolog responses. This iterative procedure converges extremely fast. Typically only five calls of the forward model are required to carry out the inversion. Compared to a parametric inversion method a speed improvement of about a factor 100 is achieved. Typical computation time is less than an hour per one hundred feet of DLL log from a deviated well, using a standard Pentium H PC or IBM R56000 workstation. This speed improvement is obtained without sacrificing accuracy. Typical RMS relative errors of the data fit achieved by the final inversion result, are in the order of 5%. The entire inversion method can be implemented such that minimal user intervention is required (“pushbutton” procedure). The performance of the method is shown using both theoretical and field data acquired in both vertical and deviated wells. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER CCC APPLY 2-D RAPID ITERATIVE INVERSION FOR LATERAL RESISTIVITY LOGS Wei Yang Rock physics Lab, Geoscience Department, University of Petroleum, Beijing, China Qingdong Shi Beijing Graduate School, China University of Mining and Technology ABSTRACT Resistivity is the most important parameter in the evaluation of movable and residual hydrocarbon. Log analysts always try to find the true formation resistivity demanded in Archie formula. But apparent resistivities can only be measured, in most case they are much departure from the true formation resistivity, especially in unfocused lateral logs. However in most oilfield in China, many wells only had lateral resistivity logs until the early 1980s. Because the resistivity suite consists of unfocused lateral logs, they are extremely difficult to interpret without an intimate knowledge of the tool’s response in a complex geologic sequence. Interpretation of lateral logs is affected by several different reasons: 1. borehole drilling fluid and 2. the apparent resistivity measured in a bed can be affected by highly resistive or conductive beds that are up to 150 feet away from the bed being investigated. These problems were very difficult to overcome in the past period, on the other hand, these wells have already been metal-cased, herein, the lateral resistivity logs are the only information we can apply now Since the resistivity response is not linear with the true formation resistivity, we have to correct the resistivity measurements for borehole, shoulder; and invasion effects accurately and simultaneously. Twodimensional inversion techniques are an excellent choice. However these techniques require significant computational time, and in many cases they become impractical. In this paper; we apply a kind of rapid iterative inversion method (RIIM) to simultaneous inversion of invasion radius Ri, invasion zone resistivity Rxo, virgin zone resistivity Rt through lateral resistivity logs. Synthetic and field data illustrate the robust of the method The characters of less dependent on initial values of unknown parameters and a rapid convergence make it more applicable. These results show that we can image the profiles of formations through lateral resistivity logs. It drastically reduces computing time by applying an high efficient forward algorithm named numerical mode match (%VL’W) and reducing the amount of forward model executions required by a 2-D inversion process and rapid convergence to a real 2-D earth model. Furthermore, automated blocking technique and other logs are used to choice layers we concern. It takes about 1 hour per 1000 feet on P11 300 PC . The method was successfully evaluated on synthetic and field data and proved to be an efficient tool for the 2-D interpretation of resistivity logging data. It has been applied in some oil-fields in China such as Daqing, Liaohe, Dagang,etc. The analysts have found many ‘lost” oil layers since they can get more precise Sw by using the Rt the method provides. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER DDD BETTER SATURATION FROM NEW ARRAY LATEROLOG R. Griffiths, J.W. Smits, O. Faivre, I. Dubourg, E. Legendre, J. Doduy Schlumberger-Riboud Product Center. ABSTRACT The history of laterolog device development has been one of modifying tool design and data processing to eliminate artifacts and improve vertical resolution. Despite significant improvements, current dual laterolog tools still have limitations. They do not always supply sufficient information to unambiguously determine the true formation resistivity (Rt), particularly in thinly bedded formations. By increasing the information available from the laterolog device we are able to invert for more complex earth models, thereby yielding a more accurate and robust Rt. Extensive modeling has resulted in a new array laterolog that provides measurements with improved information content and coherence. In addition, the measurement is free of reference influences such as Groningen and drillpipe effects. A mathematical inversion technique is used, in conjunction with a two-dimensional (2D) formation model that more accurately reflects the subsurface. This inversion benefits from the measurement improvements, yielding a more accurate Rt and therefore an improved saturation determination. The tool makes an array of simultaneous, symmetrical and focused laterolog measurements. The measurements are intrinsically depth-aligned and resolution-matched while providing multiple depths of investigation. We show that by providing coherent, information-rich measurements to the inversion, the resulting Rt is more accurate and robust than that available from conventional tools. Realistic errors are introduced to the measurements to demonstrate the stability of the R, estimation. At the welisite, Rt is provided by a fast one-dimensional (1D) inversion that only takes into account invasion effects. Improved information content and vertical resolution results in more accurate 1D Rt determination. In thinly bedded formations a two-dimensional (2D) model is inverted, which simultaneously accounts for both radial and axial resistivity variations. Significant improvements in Rt determination are seen when shoulder-bed influences are taken into account. Finally, examples from recent surveys clearly display the benefits of these methods in real reservoir characterization. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER EEE LOG INTERPRETATION WITH FAST INDUCTION LOG INVERSION Qiang Zhou Chevron Petroleum Technology Company ABSTRACT It is well known that apparent resistivity (Ra) is quite different from the true formation resistivity (Rt) in complex formation environments, Efforts have been made to apply inversion techniques to derive Rt from Ra. The advantages of using inversion are that the method automatically derives a Rt model and that the inverted model is consistent with the logs. Inversion improves bed boundary definition and the water saturation calculation. However, there are two bottlenecks to the method, i.e. the processing speed and solution uniqueness. Because of these problems, inversion still has not been routinely used in log interpretation. The major part of the processing time in a rigorous inversion algorithm is spent on calculating the Jacobian matrix that sets the direction of model adjustment. In this 1-D fast algorithm, the Jacobian matrix calculation is avoided. The fast algorithm first applies a shaping filter to the logs. Equivalently, on a first order approximation, the shaping filter maximizes the diagonal elements of the Jacobian matrix as well as symmetrizes the Jacobian matrix. The shaping filter differs from the conventional focusing filters in that it has no resolution enhancement. It only ensures that for each log point of the reshaped log the largest sensitivity is from the formation at the same depth point. The inversion solution is then updated iteratively according to the difference between the filtered log and the calculated tool response (with the same shaping filter applied) of the predicted model. The stability of the inverse algorithm is achieved by using the fact that induction measurements have little sensitivity to a resistive layer with thickness smaller than the main coil spacing. A simple automatic adjustment in correction step length is built into the algorithm to avoid resistivity over-correction and consequently the instability in updating the model. The algorithm converges to a stable Rt model typically after three to five iteration steps. Since there is no Jacobian matrix calculation, the total computation time is roughly equal to the cost of a single forward run multiplied by the number of iterations. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER FFF INTERPRETING MULTL4RRAY INDUCTION LOGS IN HIGH RT/RS CONTRAST ENVIRONMENTS WITH AN INHOMOGENEOUS BACKGROUND-BASED SOFTWARE FOCUSING METHOD Jiâqi Xiao and Ingo M. Geldmacher Baker Atlas, Houston, Texas ABSTRACT Multiarray induction logging tools measure formation conductivity with multiple three-coil arrays. The measurements from various arrays are processed through a software algorithm, referred to as “software focusing,” resulting in a set of curves with predetermined depth of investigation, vertical resolution, and other optimized 2-D features. This software focusing produces good quality focused logs when the formation conductivity varies with small to moderate contrasts between adjacent formation beds. When the formation is significantly inhomogeneous, however, the focused logs will not depict the correct formation conductivity. The root cause for the problem is the nonlinearity of the induction response and the error propagation of the software focusing process. The current focusing algorithm stems from a simplified Born approximation, referred to as the “Born-Doll approximation,” which linearly relates induction measurements to the formation conductivity distribution through the Born geometric factors. Based on this approximation, the focusing scheme linearly combines array measurements to achieve focusing through a set of focusing filters, which are designed on the basis of Born geometric factors. Consequently, the error due to the violation of the linearity assumption, or the nonlinearity effect, propagates and even amplifies through the focusing process. The Born approximation better represents an induction measurement owing to its two additive terms: the background response and the perturbation response. Starting from the Born approximation, a new focusing scheme is formulated and referred to as the “inhomogeneous background-based focusing” (IBF) method. By introducing an inhomogeneous background model, the IBF method splits an array measurement into two portions: the response of the background model and the residue, which is the difference between the array measurement and the response of the background model. The focusing result of the background model is directly calculated with the focusing target functions instead of through the focusing process. It is therefore free of any nonlinearity effect. Only the residue is processed with the conventional focusing procedure. The final focusing response is obtained by adding the two focusing results. Due to the relatively small amplitude of the residue, the nonlinearity effect introduced through the focusing process will be very small. Because the focusing result of the background model is ideal, the nonlinearity effect on the final focusing result is largely reduced. Thus, the IBF method improves the performance of the software focusing. In addition to the analysis of the current focusing scheme and description of the new method, this paper also presents examples with benchmark data to illustrate the value of the proposed method. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER GGG FLOW UNIT DETERMINATION IN A VUGGY DOLOMITE RESERVOIR, DAGGER DRAW FIELD, NEW MEXICO Neil F. Hurley, David Pantoja, and Robert A. Zimmermann Colorado School of Mines ABSTRACT Vugs are a key pore type in many carbonate reservoirs. Borehole-imaging logs are high-resolution images of the weilbore made by electrical or acoustic devices. Vuggy porosity has been measured for 845 ft (257 m) of digitally scanned core from 4 wells in Dagger Draw field, New Mexico. Core slabs were polished, coated with fluorescent ink, then photographed under ultraviolet light. Vuggy porosity was quantified from core photos using pixel-counting techniques at a rate of 120 samples/ft, matching the sample rate used by electrical borehole images. Pixel counts from borehole images were then calibrated to core results. Vuggy porosity pixel counts were resampled using a 1.5 ft boxcar filter, This filter was chosen because it represents the approximate vertical resolution of conventional density and neutron tools. Interestingly enough, filtered results generally resembled, but were less than or equal to porosity readings from conventional logs, especially the neutron log. We concluded that the neutron log is a good total porosity device in this reservoir. The difference between neutron porosity and filtered pixel-count vuggy porosity represents the intercrystalline, or “matrix” porosity. The purpose of this study is to scale-up our well-log results to flow units that can be used as input into reservoir simulators. In one well, the well-test permeability is more than two orders of magnitude higher than the geometric mean of whole-core permeabilities. Fractures are present, but not abundant. Interconnected vugs probably caused the high well-test permeability. Using linear interpolation, we resampled the neutron log to 120 samples/ft. In intervals where pixel-count vuggy porosity was higher than neutron porosity, we assigned well-test permeability. In intervals where pixel-count vuggy porosity was less than or equal to neutron porosity, we assigned core-analysis permeability. Using cumulative flow capacity (k*h) vs. storage capacity (phi*h) plots, we then subdivided the reservoir into flow units. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER HHH NEW BOREHOLE IMAGERY TECHNIQUES: AN AID FOR FAILURE MODES AND IN SITU STRESS ANALYSIS AND FOR MINIMIZING DRILLING INCIDENTS V. Maury ELF E.EY. A. Etchecopar Schlumberger P.A.Pezard CEREGE CNRS ABSTRACT Borehole imagery tools now provide an accurate vision of borehole geometry that allows to clearly differentiate stress induced deformation from drilling induced features. Once the stress induced deformations, ruptures, and damages have been identified, they can he analysed in a mechanical sense. They can be compared to rupture modes either proposed by theory or produced in the laboratory. Shear extension or true traction failures resulting in classical breakouts or tensile ruptures are commonly observed, Less common features such as shear along preexisting fractures, or even more surprising shapes probably due to particular flow and thermal regimes inside the wells are also observed. Examples of borehole imagery logs coming from field cases in very different settings (oil industry, scientific boreholes such as those drilled by ODP in the deep ocean) are presented. Predicted rupture modes with their corresponding borehole geometry and caving shapes are reviewed, according to rock lithology. The comparison between each example and predicted borehole rupture modes is discussed in conclusion. Interpretation of these observations, although more difficult, might in turn lead to - a better identification of the mechanisms responsible for the deformation, rupture or damage of the borehole wall - a determination of safe drilling trajectory, - a selection of remedies to minimize drilling problems the widespread technique of increasing mud weight to improve borehole stability may, in cases. appear verydetrimental rather than favourable, - an improved knowledge of the in situ state of stress, necessary for productivity evaluation, and geological basin modelling - an awareness of petrophysical properties changes due to the state of stress (and rupture) around the borehole as visible from high resolution borehole wall images. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER III INTEGRATED FRACTURE ANALYSIS USING BOREHOLE GEOPHYSICAL TECHNIQUES: RHOURDE EL BAGUIEL, ALGERIA B. B. Hornby Atlantic Richfield Company J. A. Lorsong SONARCO R. Wydrinski and A. R. Vittachi Atlantic Richfield Company ABSTRACT New borehole geophysical techniques to characterize conductive fractures were tested in Rhourde El Baguel. Fractures crossing the borehole were investigated using Ultrasonic Borehole Image (UBI) scans and full-waveform sonic data generated by a Dipole Shear Sonic Imager (DSI) tool. The full-waveform sonic data included acquisition of compressional waves, Stoneley waves and 4-component dipole shear waves. The UBI provided high-resolution information regarding the location and orientation of fractures penetrated by the borehole. Analysis of crosseddipole shear wave data revealed strong shear anisotropy consistent with locations of fractured intervals identified in core and UBI analysis. Also, the direction of the fast shear arrival closely matched the principal fracture direction identified from UBI analysis, suggesting that fracture direction may be determined using crossed-dipole sonic analysis. Reflected Stoneley-wave analysis was used to determine if individual fracture sets are conductive. Direct Stoneley waves excite fluid movement into conductive fractures crossing the borehole, generating reflected Stoneley waves. The amplitude of the reflected Stoneley wave depends on the conductivity of the fractures; processing yielded depth-continuous curves of Stoneley wave reflectivity. The processing was done over different frequency bands, with the low frequency band giving a significantly larger response, consistent with theoretical results for conductive fractures. Peaks in the low frequency curve coincident with fractures identified by UBI and core analysis confirmed locations of conductive fractures. In addition, interpreted locations of conductive fractures are consistent with locations of producing intervals determined by production log analysis. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER JJJ LOGGING-WHILE-DRILLING IMAGES FOR GEOMECHANICAL, GEOLOGICAL AND PETROPHYSICAL INTERPRETATIONS Tom Bratton, Ted Bornemann, Qiming Li, Dick Plumb, and John Rasmus Schlumberger Helle Krabbe Amerada Hess, Denmark ABSTRACT The cost of well construction can dramatically exceed budget if the drilling operations are plagued by wellbore instability problems. Excessive time is used to free stuck pipe or regain circulation. Subsequent well operations such as acquiring openhole logs or achieving zone isolation with a cement job are compromised, particularly in drilling extended-reach and horizontal wells. The development of a strength and stress profile for the well is the first step in understanding wellbore instability problems. These profiles are constructed using all suitable data, which include offset drilling experience, in-situ stress tests and strength measurements on recovered core. A wellbore stability forecast, guided by these strength and stress estimates, includes the identification of drilling hazards and the prediction of a suitable mud density. Real-time logging-while-drilling (LWD) data can be used to refine these profiles and help select the best remedial actions to optimize the drilling operations. Borehole images are essential for diagnosing the mechanism of wellbore failure and annular pressure while drilling data can help calibrate the strength and stress parameters. Although the majority of azimuthal images have been acquired to understand the geology and petrophysics of reservoirs, the images usually contain artifacts resulting from geomechanical processes. An analysis of these artifacts is important for understanding the geomechanics of the well and improving the geological and petrophysical interpretation. Time-lapse data are particularly important in monitoring dynamic processes such as formation failure and invasion. This paper shows several examples of how to use LWD data for a combined geological, petrophysical and geomechanical interpretation. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER KKK CONSTRUCTING CAPILLARY PRESSURE CURVES FROM NMR LOG DATA IN THE PRESENCE OF HYDROCARBONS Yakov Volokitin, Wim Looyestijn, Walter Slijkermail, and Jan Hofman Shell International E&P RTS ABSTRACT Measurements of capillary pressure curves are an essential element in rock characterisation as capillary behaviour is one of the main factors that govern the hydrocarbon distribution in a reservoir. Up to now such measurements were only available on core samples. Usually only a restricted number of core samples is present and subsequent averaging and interpolating is necessary to describe significantly large reservoir units. Advantage to obtain capillary pressure information from NMR log data, would be that it allows capillary pressure data to be acquired without coring, continuously and with sampling equal to that of the wireline logs. Methods have been proposed to convert NMR relaxation time distributions obtained on core plugs directly to capillary pressure curses (Loren et al., 1970, Marschall et al. 1995). Recently a number of papers have appeared where authors have applied this concept to formation evaluation purposes (Altunbay et al. 1998, Lowden et al., 1998, Hassoun et al., 1997, Hodgkins et al., 1999). A serious shortcoming of all methods proposed up to now is that the conversion is based on NMR T2 distributions obtained on fully water-bearing rock samples. The presence of hydrocarbons strongly dictates the shape of the T2 distributions and thus invalidates the predicted capillary pressure curve. In this paper we present an approach which can greatly alleviate this problem. This technique is based on reconstruction of the spectra at Sw=1 from spectra of bound water and model spectra of moveable water. The position of the moveable water peak is determined from a correlation between geometrical mean T2 of spectra at Sw=1 and irreducible water saturation. The so-obtained capillary pressure curves can be applied to saturation prediction in a reservoir. We show that the conversion of NMR data to saturation-height functions can be done applying the same value of conversion factor to all facies in the reservoir, rather than using individual relationships. In order to aid practical applications of the proposed method we have assessed the accuracy of saturation prediction with and without core calibration of the log data. To demonstrate the validity and possibilities of the new technique we also present a field case using real log data. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER LLL CALCULATION OF COMBINED T1 AND T2 SPECTRA FROM NMR LOGGING DATA S. Menger, MG. Prammer, and ED. Drack NUMAR Corporation, A Halliburton Company ABSTRACT Bulk and surface relaxation phenomena are commonly described in terms of longitudinal relaxation times (T1) and transverse relaxation times (T2). Of these, T1 information is not easily obtained by current nuclear magnetic resonance (NMR) logging tools for two reasons. First, T1 measurements are time intensive, resulting in low logging speeds. Second, tool motion destroys the depth coherence within a T1 measurement. Consequently, efforts have been concentrated on extracting as much information as possible from T2 spectra. This approach lacks resolution power for slow relaxation processes because long pulse-echo trains are required to follow slow T2 decay. Formation evaluation of carbonate reservoirs provides an example for the need for resolving slow relaxation processes. Both surface relaxation (which mostly affects the water phase) and bulk relaxation arise from weak molecular interactions causing slow relaxation. The newest generation of NMR logging tools can directly capture slow relaxation processes by observing the T1 polarization buildup. Multiple individual measurement volumes can be excited in rapid succession, with polarization sequences specifically designed to obtain T1 information. In this way, depth coherence of the measurement is maintained, and correct T1 spectra are obtained even when the tool is in motion. This paper shows how the appropriate acquisition sequences are implemented and how the resultant twodimensional data sets are processed to yield composite T1 /T2 relaxation information. We use forward modeling to show that improved resolving power is obtained with the combined T1 /T2 measurement. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER MMM NEW INTERPRETATION METHODOLOGY BASED ON FORWARD MODELS FOR MAGNETIC RESONANCE IN CARBONATES T. S. Ramakrishnan, E. J. Fordham. L. Venkataramanan, M. Flaum, and L. M. Schwartz Schlumberger ABSTRACT In an earlier paper (Ramakrishnan et al. 1998a) we showed that the conventional approach to the interpretation of nuclear magnetic resonance (NMR) measurements on water-saturated reservoir rocks breaks down in the case of grain supported carbonates in which the pore space consists of micro- (intragranular) and macroporosity (intergranular) in physical proximity. Our forward models. in agreement with both laboratory and log data, predict that the widely assumed relationship between the 1’2 and pore size distributions is not valid. This breakdown, due to diffusion between the micro- and macropores, limits our ability to predict permeability and movable fluid fractions from NMR measurements. In this paper our earlier work is extended in three distinct directions. First, our original forward model is extended to treat vugs, pore lining clay in sandstones, and rocks with a mixture of oil and water in their pores. Second, we report a new experimental finding relating to the temperature dependence of NMR in carbonates, a systematic shift of the decay spectrum to longer times as the temperature increases. Third. we show that the family of models we have developed can be used as the basis for a new interpretation algorithm in which the magnetization decay data are inverted directly for the rock’s geometrical parameters. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER NNN AN NMR HIGH-RESOLUTION PERMEABILITY INDICATOR A. Sezginer, C. Cao Mirth, N. Heaton, M. Herron, and R. Freedman Schlumberger G. Van Dort Amerada Hess UK ABSTRACT NMR permeability schemes in use today are derived indirectly from the T2 distribution, i.e. through the estimation of the log-mean T2 or the bound fluid volume, However, obtaining the T2 distribution requires stacking the data to improve the signal-to-noise ratio (SNR) before inversion, The drawback of stacking is either a slow logging speed or a reduced vertical resolution. A new NMR high-resolution permeability indicator is derived from the sum of echoes in a single phase-alternated pair (PAP) consisting of two Crr-Purcell-Meiboom -Gill (CPMG) sequences. The sum of all echo amplitudes in the echo train is proportional to the product of porosity and the average T 2. This, in turn, correlates to permeability. Because the sum of echoes has a high signal-to-noise ratio, it can be interpreted without stacking and., hence, with high vertical resolution. The vertical resolution achievable with the technique is equal to the antenna aperture plus the distance traveled during one CPMG sequence plus one polarization time, the sum of which is a few inches. The technique is verified in a test well where the formation consists of well-characterized blocks of different lithologies, porosities, permeabilities and thicknesses. The CMR Combinable Magnetic Resonance Tool, which has an antenna aperture of 6-in., was used in the tests. In a Gulf of Mexico turbidite example, the new NMR highresolution permeability indicator is able to identify several thin sand-shale laminations that are overlooked by conventional techniques, Other examples of the high-resolution indicator include a net-to-gross computation in laminated formations and a calibrated quantitative permeability computation. The new measurement is compared with standard NMR permeability estimators such as Timur-Coates permeability and also with core measurements. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER OOO INTEGRATION OF NMR AND CONVENTIONAL LOG DATA FOR IMPROVED PETROPHYSICAL EVALUATION OF SHALY SANDS Galy M. Ostroff, David S. Shorey and Daniel T. Georgi Baker Atlas, Houston, Texas, USA ABSTRACT The integration of NMR data with conventional log measurements can significantly improve petrophysical evaluation of shaly sands. Much more complete log-based reservoir descriptions are possible with the addition of NMR data, including the determination of several important reservoir properties that were previously quite elusive in the absence of core measurements. These NMR determined reservoir properties include; mineralogically independent total and effective porosities, irreducible water saturation, permeability, clay bound water volume, shale volume, shale distribution and Qv. Through the integration of NMR and resistivity data, we can improve in-situ determinations for formation resistivity factor and saturation exponent, resulting in far more robust water saturation solutions. Based on a direct comparison of resistivity-based water saturations and NMR-derived irreducible water saturations, the presence of moveable water in the reservoir can be ascertained. The end result is an improved determination of reservoir rock properties, hydrocarbon storage capacity and reservoir productivity. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER PPP HEAVY OIL VISCOSITY DETERMINATION USING NMR LOGS G. A. LaTorraca, S. W. Stonard, P. R. Webber, R.M. Carison, and K. J. Dunn Chevron Petroleum Technology Company, La Habra, CA ABSTRACT The apparent Hydrogen Indices (HIapp) of heavy (high density) oils as seen by NMR logging tools are <1, and dependent on the echo spacing (TE) value used in the measurement. The actual HI values of the samples we tested, however, were determined by geochemical means to be within 5% of 1. Transforms for estimating heavy oil viscosity for the MRIL and CMR logging tools were determined using laboratory NMR measurements (T2) made on oil samples with viscosity values ranging from <1000 cp to >100,000 cp at the probe temperature of 28°C, and API gravity values ranging from 8.7 to 14.5. For the echo spacings associated with each tool, the viscosity transforms were based on either the loss of signal (HIapp ) or from the log mean average T2 value of the oil. HIapp was related to viscosity (Ν) as n=a*exp(b* HIapp ) where a is inversely proportional to the square of TE and b is approximately constant. The T2 of the oil samples (1>1000 cp) can be related to the viscosity through an adaptation of the Vinegar equation (Kleinberg and Vinegar,1996). A larger prefactor is used (1.8 vs 1,2), and T 2 is replaced by T2 —10’3(TE-0.5) where T2 is in seconds and YE in ms. The larger prefactor may occur because the effects of any dissolved oxygen (paramagnetic) in the heavy oils would have negligible effect on the short T2 values measured. The TE terms are needed to account for under-sampling of the rapidly relaxing T2 signals from the heavy oils. We applied these transforms to NMR logs from a heavy oil reservoir where the oil viscosity was found to be 1,000 cp to 4,000 cp in one zone and 5,000 to 9000 cp in a deeper zone. High noise levels in the NMR logs, resulted in the large uncertainty in the viscosity estimates. th SPWLA 40 Annual Logging Symposium, May 30-June 3, 1999 PAPER QQQ APPLICATIONS OF FAST NMR LOGGING IN SHALY GAS-BEARING SANDS INVADED BY OIL FILTRATE Mario J.C. Petricola Schlumberger Overseas S.A. Xavier Toussaut Elf Petroleum Asia B.V. ABSTRACT Recent papers have described the use of NMR data acquired at slow speed and long wait time to enhance the accuracy of formation evaluation in shaly gas-bearing sands invaded by oil-based mud filtrate. Until now, the physics of the measurement have limited the use of NMR data obtained with fast passes and short wait time to determining bound fluid volume, and thus irreducible water saturation, when combined with other log data, such as density and neutron logs. However, recent NMR logs acquired in South-East Asia suggest that, against expectation and common wisdom, it may actually be possible to obtain a reliable porosity with such fast passes even in the very adverse conditions of oil filtrate and gas-filled pores. This paper first presents an example of NMR data used in the computation of irreducible water saturation, with comparison to core results and traditional resistivity interpretation. Such comparison should greatly enhance the confidence of the oil companies in the accuracy of booked reserves. Second, a method is described to effectively correct for the incomplete polarization of the gas phase and the oil filtrate phase in fast passes, equivalent to standard logging speed, with short wait time. It is shown that a reliable total porosity can be derived from the combination of NMR and density data. That porosity can, in turn, be used in the computation of the permeability from the Timur-Coates equation. TRANSACTIONS OFTHE? SPWLA FORTY-FIRST ANNUALLOGGlNGSYMPO.SIUM Sponsored by THE SOCEI’ Y OF PROFESSIONAL WELL LOG ANALYSTS, 8866 Gulf Freeway, Suite 320 Houston, Texas 77017 INC. Presented at THE INTER-CONTINENTAL Dallas, Texas June 4-June 7,200O HOTEL NOTICE TO EDITORS: Permission is hereby granted to publish elsewhere any of these transactions after June 7, 2000, provided that conspicuous acknowledgement is given to the original presentation of the paper and the authors of the paper have agreed to the republication. (The statements and opinions expressed in these transactions are those of the authors and should not be construed as an official action or opinion of the Society of Professional Well Log Analysts, Inc.) st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER A INCREASED OIL-IN-PLACE IN LOW-RESISTIVITY RESERVOIRS FROM MULTICOMPONENT INDUCTION LOG DATA B.F. Kriegshauser, O.N. Fanini, S. Forgang, R.A. Mollison, L. Yu Baker Atlas P.K. Gupta Roorkee University J.M.V. Koelman Petroleum Development Oman J. van Popta Shell EP Technology ABSTRACT Delineation of productive low-resistivity reservoirs is a challenge frequently encountered in hydrocarbon exploration. Proper identification and characterization of these reservoirs is essential to recover their reserves. An example of such a reservoir is a finely laminated sand/shale sequence of which the sand laminae contain hydrocarbons found, for example, in turbidite environments. Conventional induction logging tools with transmitter and receiver orientation parallel to the borehole axis provide a resistivity that is biased towards the low resistivity of the shales. A multicomponent transmitter-receiver configuration provides direct measurements to derive both horizontal and vertical resistivity that allows detection of the hydrocarbon bearing sand laminae. Under the sponsorship of Shell Technology BP, Baker Atlas developed and tested a new inductionlogging tool that utilizes this technology. The tool comprises three mutually orthogonal transmitter-receiver configurations yielding all necessary data to derive the horizontal and vertical resistivities of the formation. Prototype tools were run in several field tests, amongst others, in the Marmul Field in the Sultanate of Oman operated by Petroleum Development Oman. In the West Haima reservoir of this field, it is accepted from capillary pressure curves and production tests that the oil saturations evaluated from conventional resistivity tools are far too low. The presence of horizontal millimeter thin mica laminations coated with conductive pyrite is assumed to suppress the resistivity measured by conventional tools, despite the high net sand fraction of over 90%. Qualitatively, this model was confirmed by the high vertical resistivity derived from the multicomponent inductionlogging tool. The estimated oil saturations derived from the vertical resistivity are substantially higher than those based on conventional induction tool data, and more consistent with the higher oil saturation estimates based on capillary pressure curves. Interpretation results from the West Haima reservoir and from several other fields confirmed the capability of this new induction tool to detect and characterize laminated productive zones in low resistivity intervals. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER B 3D INDUCTION LOG MODELING AS A PRACTICAL AID TO HIGH ANGLE AND HORIZONTAL WELL INTERPRETATION Tom Barber, Tracy Broussard, Barbara Anderson and Bob Dennis Schlumberger ABSTRACT Induction logs have been used successfully for more than 40 years to extract resistivity and invasion profile information. However, these tools and most processing schemes-from charts to signal processing-are designed for vertical wells. At high angle, only inversion can separate out the conductivities of each bed. In horizontal wells, particularly at wireline time, making simplifying assumptions becomes much more difficult. Indeed, in some situations in horizontal wells, a tool may have more sensitivity to the resistivity in an adjacent bed or to invasion than to the true resistivity (Rt) in the bed it is logging. In these cases iterative forward modeling is often the only way to investigate tool sensitivity and to determine Rt. Previous work has shown it is possible to accurately model induction response in 3D formations. Recent advances in these 3D codes include building simple interfaces and, in particular, greatly improving speed. Run times of 10 to 20 seconds per logging station have been achieved on fast workstations. These advances now allow the use of 3D codes for practical log analysis in high-angle and horizontal wells, using the same “iterative interpretation” approach that was pioneered with ELMOD and similar 2D modeling codes. As a case study, we modeled multiarray induction response in horizontal and near-horizontal wells in a Middle East reservoir. The need is for an understanding of induction response for quick completion decisions. In the case of massive formations, the deep measurements can be used with some confidence in the center of the formation (invasion may influence only the shallower readings), with little effect of the adjacent beds on the deepest readings. However, in thin beds or those with complex geology leading to resistivity variations, the effect of adjacent beds combined with the invasion creates confusion in interpreting the logs. In this case, the sequence of the multiarray induction measurements with respect to depth of investigation does not make petrophysical sense and is considered to be a log anomaly that must be studied before a completion decision is taken. Current practice is to use lD forward modeling to illustrate the influence of the adjacent beds on the deepest measurements. The effects of invasion are ignored, and this simplistic approach to determining Rt is often incorrect. The 3D case study includes modeling the well path traversing facies and lithologies with the effects of invasion added to porosity and resistivity variations at a low angle. By carefully incorporating geological information, we refined the model to untangle confusing curve sequences and obtain more accurate values for Rt in individual facies. In addition, we extracted the geometrical structure of the formation layers and the invasion profile. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER C A NEW HIGH RESOLUTION ARRAY INDUCTION TOOL Randy Beste, T. Hagiwara, and George King Halliburton Energy Services, Houston, Texas, USA Robert Strickland Consultant, Austin, TX G. A. Merchant Multiphysics Numerical Research Group, Houston, Texas, USA ABSTRACT Array induction tools have become standard for wireline induction logging. These tools rely on measurements made between multiple transmitter-receiver pairs. In order to extend the radial depth of investigation of a measurement, it is necessary to increase the transmitter-receiver spacing. Multiple spacings allow profiling of radial variations in formation resistivity. However, as the depth of investigation is increased, progressively poorer vertical resolution is experienced. To compensate for this shortcoming, shallow-reading measurements have been combined with deeper-reading measurements to produce a deepreading log with good vertical resolution. Unfortunately, shallow-reading measurements are inherently more sensitive to borehole rugosity and other near-borehole effects; and the resulting noise often introduces artifacts into the processed, deep-log responses. A new, high-resolution symmetric array induction tool has been developed that provides deep- to shallowreading logs with identical, high vertical resolution. In contrast to previous tools, the new design permits separation of the vertical processing and the radial processing. As a result, each transmitter-receiver measurement can be filtered to a common vertical resolution prior to radial processing. Consequently, nearborehole effects are localized, in the shallow logs only; and these effects are significantly reduced in the deep logs. Tests of the new device have confirmed the responses predicted by modeling. The new tool provides resistivity logs at six radial depths of investigation (10-, 20-, 30-, 60-, 90-, and 120in.) with 1-, 2-, and 4-ft vertical resolution. The matched data resolution combined with multiple depths of investigation enables integration of resistivity data with nuclear magnetic resonance (NMR) and other logging data. This paper explains the principles underlying the design of the new tool and discusses the test results. Several field examples are also presented. In one example, a comparison with an earlier-generation highresolution induction device demonstrates the improvement in profiling the invaded zone. Another example shows that the new tool provides excellent vertical resolution of the deep resistivity log, even though the shallow measurements were compromised by a severe washout. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER D A NEW MULTICOMPONENT INDUCTION LOGGING TOOL TO RESOLVE ANISOTROPIC FORMATIONS B. Kriegshauser, O. Fanini, S. Forgang, G Itskovich, M, Rabinovich, L. Tabarovsky, L. Yu Baker Atlas M. Epov, Russian Academy of Science, P. Gupta Roorkee University J. v. d. Horst Shell Expro ABSTRACT Detection and correct evaluation of low resistivity pay zones using conventional induction logging tools is a major challenge in hydrocarbon exploration. Existing openhole induction logging tools are comprised of transmitter and receiver sensors, with their axes aligned parallel to the borehole axis. If the apparent formation dip is small, the induced currents flow mainly parallel to the bedding planes, thus measuring the horizontal resistivity of the formation, However, many geologic formations exhibit resistivity anisotropy, i.e., the resistivity varies with direction. For example, in thinly laminated sand/shale sequences, where the sand is hydrocarbon bearing, the vertical resistivity measured perpendicular to the bedding is larger than the horizontal resistivity. The low resistivity shales dominate the horizontal resistivity while the vertical resistivity is more sensitive to the more resistive sand layers. Existing induction tools cannot accurately detect and delineate this type of low-resistivity reservoirs and the measured resistivity will be biased towards the low resistivity shales. In exploration wells that are often drilled vertically, or close to vertical, hydrocarbon-bearing sand/shale formation are often overlooked. In collaboration with Shell Technology EP, Baker Atlas developed, built and field-tested a new multicomponent induction logging tool to resolve the formation parameters of electrically anisotropic reservoirs. The newly developed induction instrument comprises three mutually orthogonal transmitter-receiver configurations that acquire tensorial magnetic field data. The optimum tool design was based on an extensive resolution study. We analyzed more than 500 different benchmark models comprising l-D and 2D structures commonly encountered in hydrocarbon exploration. Part of the study aimed at gaining more insight into the underlying physics governing the complex responses of the new horizontal magnetic field sensors. The sensitivity of the data to formation parameters and the signal-to-noise ratio are used to evaluate our experimental design. We use the singular value decomposition of the sensitivity matrix to map uncertainties in the data into regions of uncertainties in parameter estimates, establishing a link between the uncertainty in the data and the confidence intervals of the interpreted parameter. To realistically quantify the resolution, we developed a comprehensive noise model, which combines thermal noise of the proposed acquisition scheme and systematic noise sources. We also analyzed the sensitivity of the new tool design with respect to environmental noise sources, e.g., borehole rugosity, eccentricity, “yo-yo” effects, tool bending, finite coil lengths, etc. Given the synthetic responses over the wide range of benchmark models and the sensitivities to the formation parameters, we could realistically compare and select an optimum tool configuration to resolve 2-3 m thick anisotropic beds in the presence of 5% data noise. The new tensor induction logging tool was deployed in various field environments and the data that were acquired and interpreted confirmed the predicted capabilities of the new tool. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER E A NEW METHOD TO CORRECT THE EFFECT OF SKIN-EFFECT IN INDUCTION LOGS Guoqiang Liu Department of Engineering Mechanics, Tsinghua University, China Wei Yang, Qining Feng, Guozhong Gao Rock physics Lab, Geoscience Department, University of Petroleum, Beijing, China ABSTRACT A new correcting method of induction logs is developed by introducing a widely used Lambert W function into the induction log measurements. Taking a more clear and convincing insight into the physical mechanisms, this new method can eliminate the skin effects and achieve much higher accuracy of data processing. This method removes the conventional limitation of ω σ<< ε for induction logs and therefore makes the accuracy of induction log measurements no longer dependent on the true formation conductivity. This method can obtain accurate formation conductivities in the full frequency band of induction logs. Theoretical analysis and numerical modeling demonstrate the advantages of this new method over traditional methods. The algorithm of this method is easy to be implemented and can also be realized with electronic hardware to produce accurate formation conductivity logs directly. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER F ROBUST COMPUTATION OF FOCUSED RESISTIVITY CURVES FROM HIGH-DEFINITION LATERAL LOG DATA Zeljko Jericevic, Antonio Fabris, Michael A. Frenkel, Zhiqiang Zhou Baker Atlas, Houston, Texas, USA ABSTRACT The High Definition Lateral Log (HDLL) instrument provides an army of high resolution, lateral-type measurements from which the apparent resistivity of a geological formation can be probed at a series of different radial depths. The instrument provides the most information about the formation when its data are processed by inversion technology. In order to provide results necessary for making operational decisions at the wellsite, we developed computational procedures for Software Focused Resistivity (SFR) curves that simulate Laterolog-7 type responses and deliver the results within one-half hour after the measurements. Currently, eight SFR curves for nominal depths of 10, 12, 14, 16, 20, 30, 40, and 50 inches radius of investigation are in use. Effect of noise, including uneven tool motion called “cable yo-yo’, on the HDLL instrument response have been studied. In particular, sensitivity of the SFR curves to such noise has been analyzed. The study was conducted using synthetic data as well as field data from different downhole conditions. The impact of each noise component on the SFR curve computation has been established. Subsequently, a correction procedure based on signal processing and linear algebra techniques has been developed. The developed methodology uses fast and robust digital signal and linear algebra procedures. The applied signal processing techniques are filtering, noise estimation, and Fourier phase shifting. Linear algebra procedures involve solving the focusing system of linear equations with imposed a priori constraints and eigenvalue shifting. The approach was tested on model and real data with various noise levels and different downhole conditions to estimate the sensitivity of the SFR curves with respect to resistivity contrast and noise level. Impact of each noise component (random noise, “cable yo-yo”, “stick-and-pull”) on data and SFR curve quality has been studied. The use of the developed methodology typically does not require user intervention. The strategy behind the development of this technology has emphasized robustness and computational efficiency as required for a wellsite deliverable product. The software provides an estimation of the invasion profile. The results can be presented both as eight distinct SFR curves and as a continuous resistivity image, further illustrating the flexibility of the HDLL technology that allows a single instrument to simultaneously probe the apparent formation resistivity at different radial depths. It also allows for additional SFR curves, if required, to be incorporated easily into the developed methodology, Presented case studies include comparison of HDLL SFR curves with other logging data to illustrate the versatility of the HDLL technology. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER G EFFECT OF PYRITES ON HDIL MEASUREMENTS L. Tabarovsky and D. Georgi Baker Atlas ABSTRACT The occurrence of pyrite is widespread and is known to lead to suppression of apparent formation resistivity, which in turn can lead to an underestimation of hydrocarbon reserves or to bypassed reserves. In this study, the influence of a dispersed pyrite on the High Definition Induction Logging (HDIL) measurements is investigated. The approach selected for the study is based on asymptotic estimation of inductive and galvanic effects. The low frequency is a key factor allowing us to simplify the theory. For the known limits of the pyrite conductivity (1.6-800 S/m) and for a typical grain size (0.1-1.0 mm), all appreciable effects are linear with respect to the frequencies used in HDIL (10-150 kHz). The effects of pyrite can be divided into inductive and galvanic contributions. Inductive effect is due to eddy currents excited in the pyrite grains by the direct magnetic field of a transmitter. At a low frequency, no inductive interaction exists between pyrite grains and no formation influence on the grains must be taken into account. The inductive effect is directly proportional to the pyrite conductivity, pyrite volumetric concentration, and the grain shape factor. The shape factor may be tabulated given the tool parameters and grain characteristics. Remarkably, the inductive effect does not depend on formation conductivity and allows for the correction of the effect prior to data inversion and interpretation. Galvanic effect is due to charges induced by the electric field on the grains’ surfaces. The effect appears as an equivalent increase of conductivity of the host formation. The galvanic effect is nonlinear with respect to the pyrite concentration and grain parameters. The proposed corrections for both effects include the following steps: 1) obtaining pyrite concentration and grain characteristics from cores; 2) correcting HDIL measurements for the inductive shift; 3) inverting corrected data; 4) scaling the resulting conductivity by a tabulated galvanic factor. For the realistic parameters of pyrite grains, the inductive effect is small and the inductive correction may be omitted. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER H USING ACOUSTIC ANISOTROPY T.J. Plona, M.R. Kane, B.K. Sinha, J. Walsh, O. Viloria Schlumberger ABSTRACT Sonic characterization of the anisotropic nature of a formation yields important new information for drilling and completion engineering as well as geophysics and petrophysics. Anisotropy analysis enables perforations to be oriented consistent with stress direction to minimize sanding problems and optimize production. Knowledge of the stress directions allows well placement to maximize oil recovery from hydraulic fracturing. Identification of anisotropy due to fractures leads to enhanced production. Cross-dipole sonic-logging tools detect shear-wave anisotropy in the formation. Traditional processing is performed in the time domain and yields information about the isotropic or anisotropic nature of the formation but does not identify the underlying cause of the anisotropy. However, with advanced frequencydomain processing of cross-dipole data, we can distinguish the type of anisotropy as intrinsic or stressinduced. We derive additional information from the dipole data that can supplement other logs. Numerous field examples in various rock types are shown highlighting the additional information derived from dipole dispersion curves from cross-dipole acquisitions. Dispersion analysis of these data enables identification of intrinsic anisotropy, where the dispersion curves are parallel (e.g., fractures) and stressinduced anisotropy with crossing dispersion curves. We show cases of both intrinsic and stress-induced anisotropy from the Middle East, Africa, Gulf of Mexico, California and Southeast Asia. We include examples from both horizontal and vertical wells and examples where fractured anisotropy (intrinsic) and stress-induced anisotropy occur in the same well and are clearly distinguishable. The dipole log identification of fractures is corroborated by borehole-image logs. The key to this new understanding is the combination of wide-band acquisition of cross-dipole sonic data, advanced frequency domain processing techniques, and good knowledge of acoustic-wave propagation in boreholes. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER I COMPARISON OF FEWD AND WIRELINE LOG RESPONSES: A CASE HISTORY FROM OFFSHORE SOUTH AFRICA J.C. Lean & R.T. Kelly PGS Reservoir (UK) Ltd. S.J. Lott Sperry Sun Drilling Services J.H. Jackson Mossgas (Pty) Ltd. ABSTRACT The authors are currently involved in providing services associated with the development of a number of gas fields offshore South Africa. Four horizontal wells have been drilled so far into the structures, which have been previously appraised by a number of vertical wells drilled in the 1980s and extensively cored and logged with conventional wireline tools. To give confidence that data from development wells, logged with formation evaluation while drilling (FEWD) tools, could be integrated with data from these earlier appraisal wells, both FEWD and wireline logs were acquired in a 45 degree pilot hole. On comparison of the logs acquired the major observations were that FEWD density and resistivity were respectively lower and higher than their wireline equivalents. The moderate quality gas reservoir was interpreted as exhibiting deep water-based mud filtrate invasion over the drilling-to-wireline logging lapsed time. The fluid density term within the porosity equation was modified to account for invasion over time and an FEWD effective fluid density was derived for use in porosity calculation for the horizontal hole wells. Invasion correction of the wireline resistivity was performed but was considered inadequate as resistivity was not increased sufficiently to match that seen by FEWD. Another approach was therefore required to validate the FEWD resistivity response. Porosity and water saturation analysis was performed using the FEWD data and then a porosity-dependent saturation-height function was used to generate water saturation independently. The comparison between FEWD derived water saturation and that predicted from the saturation height function was good and therefore the FEWD from the horizontal development wells could be integrated with confidence into the pre-existing field models. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER J A NOVEL APPLICATION OF NUCLEAR MAGNETIC RESONANCE AND FORMATION TESTER DATA FOR THE DETERMINATION OF GAS SATURATION IN PRETTY HILL SANDSTONE RESERVOIRS, ONSHORE OTWAY BASIN R. Ramamoorthy, Schlumberger(Malaysid) Sdn Bhd, Kuala Lumpar, Malaysia PJ. Boult Boral Energy Resources Ltd, Adelaide, Australia T. Neville Schlumberger Oilfield Australia, Pty, Ltd., Melbourne, Australia ABSTRACT Conventional log based interpretation of gas-bearing reservoirs in the Lower Cretaceous Pretty Hill sandstone, onshore Otway basin, has recently been greatly enhanced by the application of the nuclear magnetic resonance (NMR) log, calibrated to core capillary pressure data. The borehole NMR log also assists in fades differentiation of the reservoir section and in the selection of wireline pressure and formation fluid sampling points. Formation tester data are used to initialize the capillary pressures. The use of NMR to refine the selection of formation test depths significantly improved the efficiency of the survey. NMR interpretation parameters were refined by measuring the whole core at the wellsite directly using the borehole NMR tool in a novel application of the device. Ills provided both a validation of the downhole measurement and a means of differentiating the capillary-bound and free-fluid components in the NMR signal. This is the first time results of core measurements using the borehole NMR tool have actually been applied in the interpretation of the downhole log. This paper describes a novel approach to calibrate the borehole NMR data to capillary pressure information. The method facilitates the calibration when several core data points exist in a continuously logged interval. The results are validated through application in the specified well and comparison with core data and formation tester samples. The combined application of nuclear magnetic resonance technology and formation tester information, calibrated and validated by core data, resulted in a 30% increase in average gas saturation in the reservoir at the Redman field. Computed saturations were further validated by formation fluid samples obtained with the formation tester. The method described herein has been successfully implemented in other wells in the Otway Basin. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER K FORMATION EVALUATION IN SOUTHERN MEXICO’S LOW - POROSITY FRACTURED CARBONATE ROCKS USING IMAGING AND NMR TOOLS Ings. Jean-François Mengual, Laura Morales Recinos, Elizabeth Sanchez Ho SCHLUMBERGER Dr. Francisco Garcia Hernández PEMEX ABSTRACT The environmental settings of pelagic deposits in fields of the Cretaceous section of southern Mexico conflict with the tectonic features of this region. The sedimentary sequence has been deformed by compressive stresses and other tectonic effects (for example induced by salt intrusion) of varying velocity and intensity. Because the main producing reservoir is made up compacted carbonate rocks, the oil content and producibility are extremely influenced by the presence of fractures. Formation evaluation from conventional wireline logs is difficult because of the complex lithology (mixture of calcite and dolomite) and the rock texture (very low primary porosity, closed and open fractures, complex breccia texture and presence of vugs). Despite significant improvements in the estimation of porosity, grain density and true resistivity, conventional log evaluation of water saturation is insufficient to predict the producibility of a well. This paper will discuss the use of borehole resistivity images for fracture evaluation and recognition of rock textures and the use of nuclear magnetic resonance (NMR) logs for porosities (both effective and secondary) and permeability to complement conventional logging techniques and increase initial production flow rates. Hydrocarbons are produced through natural fractures caused by compressional stresses within the tight carbonate reservoir or the highly cracked and broken breccia. Productive zones identified with this new approach are fitting production tests, illustrating the fact that these carbonate sediments have high potential for hydrocarbon production only if the fracture and microfracture networks can be clearly recognized and identified within the thick carbonate sequence. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER L REAL-TIME FORMATION EVALUATION FOR OPTIMAL DECISION MAKING WHILE DRILLING: EXAMPLES FROM THE SOUTHERN NORTH SEA J.F. Bristow Schlumberger Data and Consulting Services, Gatwick, England ABSTRACT In recent years there has been a rapid growth in horizontal well completions driven by the need to reduce field development costs. Logging while drilling (LWD) technology and geosteering techniques have advanced to ensure high rates of success in reaching reservoir targets that are smaller and less clearly defined than those attempted previously. Three recent examples illustrate the benefits of these techniques where LWD data are acquired at the rig-site, transmitted real-time to the operator’s office and interpreted by a multidisciplinary asset team who update formation models to enable optimum geosteering decisions. Prior to drilling the horizontal wells, prejob forward modeling based on offset well data and structural information from the earth model is performed to predict the LWD log responses along the planned well trajectory. While drilling, the formation model is refined to minimize spatial uncertainties within the reservoir and to provide a predictive model of the formation relative to the wellpath. This is achieved by correlating the real-time LWD logs with forward-modeled log responses. The resulting correlations constrain the position of the bit in the formation, and so apparent formation dip can be computed. Synthetic LWD logs are predicted for the projected trajectory 150 ft ahead of the bit. Uncertainties in the formation structural model are further reduced by interpreting LWD azimuthal density images retrieved between bit runs. These are processed immediately on a workstation in the operator’s office and provide dip information to constrain the structural interpretation and lateral changes in stratigraphic thickness. The image data also provide facies information and in these wells help identify zones of higher permeability. Three case studies show how using geosteering based on predictive real-time modeling can help manage the risk associated with drilling horizontal wells by reducing positioning uncertainties. They also show how optimizing well placement improves well productivity. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER M ANALYSIS OF ENVIRONMENTAL CORRECTIONS FOR PROPAGATION RESISTIVITY TOOLS W. Hal Meyer Baker Hughes INTEQ ABSTRACT Measurement While Drilling (MWD) propagation resistivity tools require several environmental corrections to achieve accurate results. The environmental effects that require correction include borehole effects, invasion, dielectric effects, anisotropy, and effects of nearby formations. An effective strategy has been developed to address each correction even when more than one affects a given data set. This strategy starts with a borehole correction followed by deconvolution. Invasion, anisotropy, and dielectric effects are then evaluated simultaneously. These three effects are evaluated together because errors result if they are evaluated separately. Analyzing one effect alone is an attempt to explain all of the differences in the data curves with that one effect. This attempt will fail when other effects contribute to separation of the data curves. In addition, the improperly corrected data will not be easy to interpret. The best-fit solution to the data set must be found with a model that includes all three effects. This strategy has been tested using both computed data and field data. The results show that anisotropic effects often cause the largest residual error in highly dipping boreholes. In highly resistive pay zones, the dielectric effects cause a significant error, but these effects can usually be removed from the data during the processing. Residual errors due to side bed effects are usually small (that is, the deconvolution works well). When the relative dip angle is more than 70 degrees, however, the deconvolution does not work well and residual errors are large. Invasion effects can usually be eliminated from the data resulting in an accurate measurement of Rt, but the values of Rx0 and are often difficult to determine with the shallow or nonexistent invasion typical of MWD logging. After analysis, corrections are made for all effects except invasion. The data are then displayed at four different fixed depths of investigation for quantitative identification of any invaded zones. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER N GEOSTEERING EXAMPLES USING MODELED 2-MHZ LWD RESPONSE IN THE PRESENCE OF ANISOTROPY John Edwards Schlumberger ABSTRACT Resistivity modeling of logging-while-drilling (LWD) response from wireline logs can determine whether the well geosteering requirement is feasible. However, in the typical situation, the wireline logs are from vertical appraisal wells, and the geosteering wells are horizontal; if the formation is anisotropic, the prejob modeling will not account for the bed perpendicular resistivity. The presence of anisotropy can be unknown until the first LWD logs are run, when real-time model revision then becomes imperative to avoid compromising the well placement. The first example in the paper is geosteering for saturation in an oil rim. The reservoir development plan required the drain-holes to be 4 m -f-I- 1.5 m above OWC. Error models showed that because of the lateral displacement, this geometric tolerance was impossible using measurement-while-drilling surveys alone. Vertical logs showed the oil column was in transition, with capillary pressures decreasing the water saturation above the free-fluid level. An empirical algorithm was developed for this reservoir that related resistivity to the height above the OWC. Resistivity-derived height above contact was used to keep the drain-hole in the vertical window. After the first well, it was apparent that using the algorithm would require extracting the bed parallel resistivity from the LWD logs. This was done using the separation between phase and attenuation resistivity of the 2-MHz logs, while also accounting for the effect of adjacent beds in this dipping, thinly bedded clastic sequence. Resistivity anisotropy in this case was due to variations in grain size within the reservoir. The second example demonstrates geosteering around existing completions. The reservoir development required that new horizontal wells cross existing horizontal wells. Potential coning from fluid contacts required these crossings to be within a vertical separation of 2 m. Since this clearance is smaller than the closest approach possible using geometric collision avoidance, geosteering was used as the collision avoidance technique. The well crossings were made by geosteering in the formation layer either over or under the existing well and required a detailed resistivity model of the stratigraphic sequence, which was used to establish the relative positions of the existing and new well-bores. This model was made using wireline logs from vertical wells and LWD logs from landing sections. The anisotropy apparent on the 2MHz LWD logs was due to thin siderite layers. Both examples resulted in successful well placements, with production matching the reservoir simulation forecasts. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER O LWD SHEAR VELOCITY LOGGING IN SLOW FORMATIONS DESIGN DECISIONS AND CASE HISTORIES G. L. Varsamis, A. Arian, J. Blanch, J. Market L. T. Wisniewski SensorWise, Inc., Houston G. Althoff, C. Barnett Sperry-Sun Drilling Services, Halliburton Energy Services, Inc., Houston ABSTRACT As LWD formation evaluation applications evolve, the need for a full suite of logging sensors becomes evident. A key aspect of any complete logging suite is a full-wave sonic tool. Previously only wireline dipole sonic tools had the ability to measure formation shear wave velocity in slow formations, i.e., formations in which the formation shear velocity is lower than the mud acoustic velocity. Recently a new full wave LWD tool was introduced specifically designed to measure shear velocities in slow formations. The tool contains two opposing arrays of seven receivers each and two programmable multi-frequency transmitters. The transmitters can be energized at two separate frequencies, in-phase or out-of-phase, and either synchronously or with a time delay. The unique design of the tool allows for excitation of higher order borehole modes that facilitate the measurement of formation shear velocity in slow formations where monopole-type LWD tools cannot provide a shear wave measurement. This paper presents some key design choices associated with the ability of the tool to enhance the shear velocity measurements in slow formations and several case histories that include comparisons with dipole wireline logging tools. The case histories present data from the Gulf of Mexico and the North Sea and from both fast and slow formations. The following case histories are presented: (1) A case from a Gulf of Mexico well (with wireline comparison) that demonstrates shear velocity measurement in slow formations. (2) A case from a Gulf of Mexico well (with wireline comparison) that demonstrates the advantages of the tool in measuring newly drilled formations, versus the wireline tool that measured in a severely enlarged borehole. (3) A case from a North Sea well (with wireline comparison) that demonstrates the capabilities of the tool in a traditional fast formation environment (presence of refracted shear). In addition, two other interesting case histories are presented that are related to compressional velocity measurements at bed boundaries or in relatively thin beds. The first is a case from a Gulf of Mexico well that demonstrates the ability of the tool to measure compressional velocities in a sharp transition from about 170 µsec/ft down to 60 µsec/ft (drilling into a salt dome). The second is a case from a North Sea well in which the tool measured ∆t in several low-porosity zones down to a compressional slowness of 52 µsec/ft. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER P MUD VELOCITY CORRECTIONS FOR HIGH ACCURACY STANDOFF/CALIPER MEASUREMENTS Eric Brian Molz Baker-Hughes Inteq Ultrasonic measurement of the tool standoff and borehole diameter has many established uses in the well logging community: correcting neutron porosity and density logs, borehole imaging of rugosity and fracture detection, borehole shape measurements and cement volume calculations. Another possible application involves multistage drilling (Altoff, 1998, Bfrchak, 1993, Orbin, 1991). For all of these applications, accuracy in the ultrasonic measurement is imperative (usually to within ± 0.10” for caliper, and 0.05” for standoff). Larger errors can greatly effect the mud corrections in the standoff-based density log. The ultrasonic detection method must ensure that in lab conditions the ultrasonic standoff/caliper gauge is accurate to within these specifications. This can be achieved through proper a combination of ultrasonic transducer design and positioning and phase-sensitive averaged detection methods. However, such accuracy is meaningless if the large (-.20%) uncertainty in the sound velocity is not addressed. Our mud velocity correction technique was done using lab measurements of ultrasound velocity in drilling muds for various mud weights, temperature and pressure. A correction algorithm was then derived that reduces mud velocity error for known field conditions to, in principle, only a few percent. We present lab data demonstrating how the required accuracy is achieved when the fluid (water) sound speed is well known. We then present field examples of the logs where no velocity correction is made. In the case of heavy muds, a consistent offset as much as 0.3” in the caliper reading is observed above the expected borehole diameters. Wireline logs and known casing inner diameters verify these discrepancies. Once the correction algorithms are applied, excellent correlation (>0.10” error) between MWD and wireline is produced. In some cases, the MWD caliper is actually smaller than wireline caliper, showing regions where washout or damage has occurred. Still, problems remain with this post drilling correction, It is of no use to the standoff based binning as that data cannot be post-corrected. In a closed loop rotary drilling system, where the borehole diameter is expected to be in gauge nearly throughout, artificially large caliper values are, to say the least, a major concern to the driller. A system has been created so that, given the expected drilling parameters, the caliper and standoff values are adjusted for real-time accuracy. Log examples will demonstrate that mud velocity approximations can accurately adjust the MWI) caliper and standoff prior to drilling, thus producing highly accurate standoff-based binning of nuclear density data. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER Q NEW ELECTROMAGNETIC PULSE-EXCITED TOOLS FOR MWD/LWD MEASUREMENTS P.I.Dvoretzki GAZPROM,Inc, Russia. I.G.Iarmakhov, S.B.Popov. Inst. of Radioeng.& Electronics RAS, Russia. ABSTRACT The electromagnetic pulse excited (EPE) tool is a logging-while-drilling device designed to enhance both accuracy and depth of formation evaluation. New modifications of EPE tools with a short-pulse excitation of the radiating elements are developed for MWD/LWD applications in the following special cases: a) slot in a conducting cylindrical surface, b) dipole antenna on the mandrel of the tool. The first prototypes are built on a 6.75-in drill collar. The tool is 10 ft long and can be configured near the drill bit. Due to a wide-band excitation, the EPE tool performs azimuthally scanning measurements. Two- and three- dimensional (2D and 3D) mathematical models of the tools are presented. A computer program has been developed to calculate the time domain scattered fields for different parameters of rock formations, salty muds, and impulse types. Some effective new analytical approximations of the pulse responses are presented using Laplace transform of Sommerfeld type integrals in the 2D and 3D models. These asymptotic solutions are used to interpret and evaluate the accuracy of the direct numerical computations concerning the pulse propagation in borehole environment. The character of pulse dispersion at the point of observation, moments of arrivals of the direct, reflected, lateral waves, shape and polarity reversals of the transients permit to evaluate the constitutive parameters of borehole formation. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER R CORRECTING FOR WETTABILITY AND CAPILLARY PRESSURE EFFECTS ON FORMATION TESTER MEASUREMENTS H. Elshahawi, M. Samir, and K. Fathy Schlumberger Oil Field Services ABSTRACT Wireline formation testers such as the well-established Repeat Formation Tester (RFT) measure the pressure of the continuous phase present in the invaded region, typically the drilling fluid filtrate. Conventional interpretation techniques have long assumed this pressure identical to the pressure of the continuous phase in the virgin region of the formation such that a series of pressure measurements at different depths would be expected to consistently yield a pressure gradient corresponding to the density of the formation fluid. More recent work has shown that the measured tester pressure is not simply identical to formation pressure. Rather, it is different from the formation pressure by the amount of capillary pressure, which itself is strong a function of the wetting phase saturation. The effects of rock wettability and capillary pressure on wireline formation tester measurements are manifested as pressure gradient changes and/or fluid contact level changes on many logs, particularly those recorded with OBM as the drilling fluid. This paper explores the effects of wettability and capillarity on wireline formation measurements and investigates the possible techniques for estimation of and correction for these effects. The older techniques rely on the use of core capillary pressure data and log-computed flushed zone saturations, but they can often yield inconsistent results. Amongst the more innovative techniques suggested in this paper is the use of NMR measurements to estimate the magnitude of capillary pressure effects and hence correct the measured tester pressure. Also discussed is the use of the pump-out capability of the Modular Dynamics Tester (MDT) to estimate the capillary pressure effect by making measurements before and after the removal of drilling mud filtrate. This allows accurate correction for capillary pressure effects, or better still, the elimination of the need for the correction altogether. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER S NEW DUAL-PROBE WIRELINE FORMATION TESTING AND SAMPLING TOOL ENABLES REAL TIME PERMEABILITY AND ANISOTROPY MEASUREMENTS Mark A. Proett, SPWLA, and Wilson C. Chin, SPWLA, Halliburton Energy Services ABSTRACT Two new testing methods are used to determine horizontal permeability and anisotropy using a dual probe formation tester. Real-time interpretation during acquisition of pressure data is the focus of the new pressure testing techniques. A new multi-probe anisotropy model with storage is developed from first principles. Historically, pressure transient solutions are presented as Laplace transforms and approximated in the time domain using numerical techniques. In this paper, the Laplace transform is inverted to yield an exact analytical time domain solution making real-time parameter matching possible with higher accuracy. The second technique uses an oscillatory or harmonic source where the phase delay of the pressure pulse between the probes is used to determine permeability and anisotropy. This technique is particularly useful when the second monitoring probe signal is weak in high permeability zones or where there is large spacing between the probes. The pulse timing can usually be detected more accurately than its magnitude, thus extending the range of permeability and anisotropy measurements. Simulation results using a detailed finite element dual probe model are compared against the analytical methods to assess the effect of near wellbore parameters (i.e., mudcake effectiveness, probe and packer size). Field-test results are included to show the implementation of these new testing methods. In particular, the results indicate that monitoring permeability and anisotropy changes while sampling is helpful in identifying changes in the fluid type. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER T TECHNIQUES TO IMPROVE THE QUALITY OF WIRELINE OIL SAMPLES IN WELLS DRILLED WITH OIL BASE MUD Scott C. Wilmot ExxonMobil Exploration Company ABSTRACT State-of-the-art wireline formation test tools are routinely used in exploration wells in order to obtain hydrocarbon samples for PVT analysis. The use of oil base mud (OBM) complicates the sample acquisition process in that the OBM filtrate contaminates the reservoir hydrocarbons. In wells drilled with OBM, conventional monitoring of fluid resistivity change during the pump-out/cleanup phase of the wireline formation test tool sampling operation is inadequate to determine when the reservoir oil being extracted from the formation has reached a minimum acceptable level of contamination. Esso has been faced with this problem in Angola and has adopted a two-part strategy for the acquisition of oil samples. Part one involves tool specific techniques, such as careful monitoring of optical density curves during pump-out operations, innovative tool configurations, repeating phase separation tests downhole to monitor the change in fluid composition, and sampling above the reservoir oil bubble point. Part two is to use sample point guidelines such as sampling toward the top of long oil columns, sampling in thin beds, and sampling directly below impermeable bed boundaries. Implementation of this two-part wireline formation test tool sampling strategy has yielded more representative fluid samples, and has subsequently improved the reliability of reservoir fluid analyses. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER U DETERMINING FLUID VOLUME IN GAS AND/OR LIGHT OIL RESERVOIRS: USING A NEW TRIPLE-WAIT-TIME NMR LOGGING METHOD Bob L. Hou and Danny Miller Halliburton Energy Services, Houston, TX, USA ABSTRACT One challenge for determining fluid volumes from NMR logs nm in gas and /or light oil reservoirs is the “signal overlap” of the various oil, water, and gas phases. Current practice calls for a log analyst to use an echo-train difference obtained from dual-wait-time (DTW) acquisition in a magnetic field gradient to separate the signals from the individual fluid phases. However, the accuracy of the fluid volumes calculated by such a method frequently suffers from the difficulty of determining the correct values for the spin lattice relaxation time (T1) of the hydrocarbons. A triple-wait-time (TTW) method that overcame this shortcoming in the DTW approach was recently proposed, and a previous publication showed that the method performs well when used to determine fluid volumes in the laboratory. This paper describes the successful application of the TTW logging method in field operations with the MRIL-Prime logging tool. The method has been applied in an oil reservoir, and the results from the specific study are presented. The study starts by comparing the data acquired by the TTW method against those acquired by the DTW method. Then the fluid volumes resulting from each method are compared. The effects of noise, pulsing time, logging speed, data stacking, depth sampling, and magnetic field gradients are all considered in terms of their impact on the accuracy of the estimates of T1 and T2, and the fluid volumes from the TTW method. Various limitations of this method are also discussed. An outline for selecting the data acquisition parameters that provide optimal TTW data sets is presented to guide pre-job planning. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER V HIGH-RESOLUTION BOUND-FLUID, FREE—FLUID AND TOTAL POROSITY WITH FAST NMR LOGGING N. Heaton, C. Cao Minh R. Freedman and C. Flaum Schlumberger ABSTRACT One of the most important contributions of nuclear magnetic resonance (NMR) logs has been the provision of quantitative estimates of bound- and free-fluid volumes. which are fundamental for evaluating formation producibility. In laminated formations, it is important that measurements are sensitive to variations on a length scale comparable to or less than the lamination thickness. The vertical resolution of NMR log measurements is defined by the antenna length, the acquisition sequence and the logging speed. During depth logging, each individual measurement consists of a phasealternated pair (PAP) of Carr-Purcell-Meiboom-Gill (CPMG) sequences. Measurement schemes that combine overlapping measurements to form PAPs provide the best resolution. typically about 9 in., for a tool with a 6-in, antenna. However, because of the long polarization times that are sometimes required, the overlapping acquisition methods place limitations on the logging speed. For tools having a long prepolarization magnet, nonoverlapping PAP measurements may be acquired at logging speeds of up to 3600 ft/hr with a minor reduction in vertical resolution. In practice, however, the vertical resolution is often degraded by depth-stacking, which is necessary to improve the signal-to-noise ratio for inversion. High-resolution processing strategies and acquisition schemes have been developed, applicable to standard CPMG and Enhanced Precision Mode (EPM) data, that provide estimates of total porosity, free-fluid and bound-fluid volumes at the maximum tool resolution. The methods have been evaluated on synthetic data generated from known T: distributions with added random noise and on log data acquired in a test well comprising blocks of known lithology, thickness, porosity and permeability. Log examples presented here show how the processing is used to identify thin laminations and to determine quantitative estimates of bound-fluid and free-fluid volumes for the constituent beds. Results are compared with other highresolution porosity logs and extensive core measurements for the same well. Combination of the new processing with EPM acquisition and nonoverlapping measurement schemes, implemented on a tool with a short antenna length and long prepolarization magnet, provides high-resolution NMR logs acquired at fast logging speeds. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER W SUMMATION OF ECHOES: A SIMPLE, ROBUST MEANS FOR INTERPRETING NMR DATA Songhua Chen, Daniel Georgi, Christina Liu, and Holger F. Thern Baker Atlas, Houston, Texas, USA ABSTRACT Sum of echoes (SE) provides a simple but novel approach to process and interpret NMR log data. Summing the echoes or a subset of echoes within one single echo train has several unique benefits for NMR log quality control (QC) and interpretation. Firstly, the signal-to-noise ratio is much improved because incoherent noise is minimized while the coherent signal increases. Secondly, SE is sensitive to the product of <T2> and, and, thus, is more responsive to variations in either or both quantities. Thirdly, SE enhances the difference between long and short wait-time data (TWL and TWS, respectively) making it possible to identify small hydrocarbon signals and, because minimal vertical averaging is required, thin hydrocarbon reservoirs. Sum of echoes, SE, is a simple and beneficial means for quality controlling NMR logs. Good agreement of both porosity and T2 is expected for repeat logs. Thus SE is an ideal means for log verification of repeat passes and for cross-validation of multiple frequency data acquired with similar logging parameters. In addition, we developed a procedure to QC the individual quadrature channels of the echo data with the SE approach, allowing identification of anomalies in the raw data. Such anomalies are difficult to detect from the phase-alternate, averaged echo magnitudes. Moreover, the simplicity of the SE technique enables this QC procedure to be implemented in acquisition software and delivered real-time. A second, novel application of the sum of echoes is the T1 estimation of the slow-relaxing components using two or more echo trains with different wait times, e.g. TWL and TWS. These data can be acquired from the same or different logging passes. Previous approaches for estimating 2’1 from dual wait-time logs required simplified vertical averaging to reduce noise, which hindered the identification of thin hydrocarbon beds. In contrast, the SE approach achieves noise reduction by summing the echoes within each sample, without degrading the vertical resolution. Examples are presented to demonstrate the SE applications. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER X SIGNIFICANCE OF NMR T2 DISTRIBUTIONS FROM HYDRATED MONTMORILLONITES D.V. (Vivek) Chitale, John Gardner, and Richard Sigal Halliburton Energy Services ABSTRACT Petrophysical application of nuclear magnetic resonance (NMR) technology to achieve clay-typing in shaly sands requires that the NMR signature of the clay-bound water be clearly established. This paper presents NMR T2 distributions obtained from a series of experiments performed on montmorillonite samples containing either clay-bound water alone or both the clay-bound and inter-aggregate pore water. Calcium- and Sodium-montmorillonite samples, provided by the Clay Minerals Society, were hydrated using two separate methods. In one set of experiments, samples were hydrated by vapors under controlled humidity conditions to ensure that the entire water of hydration was adsorbed on the clay surface. In a second set of experiments, samples were hydrated by liquid water so that water in excess of clay-bound water occupied the inter-aggregate pores in the wet clay. NMR experiments were conducted on these hydrated montmorillonites using a 1 MHz spectrometer at an inter-echo time of 0.3 ms. The maximum clay-bound water, fully-recovered in the NMR T2 distribution obtained from hydrated montmorillonites, was observed to be 500 mg/g of dry clay; consistent with values reported in the clayscience literature. In hydrated samples at less than this critical saturation, the T2 distribution was uni-modal. However, at super-critical saturations the T2 distribution became bi-modal. As water saturation was increased above super-critical, the location of the faster T2 component remained fixed, while the slower T2 component shifted to later times. We propose that the fixed T2 component is characteristic of the claybound water, while the slower T2 mode represents discrete water phase in the inter-aggregate clay pores whose volume increases as the water content increases. This paper establishes the T2 characteristics of the clay-bound water in montmorillonites, which may provide a physical basis for applying NMR to clay typing in formation evaluation. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER Y TIME-DOMAIN CORRECTION OF MRIL LOGS CONTAMINATED BY BOREHOLE SIGNAL Holger F. Thern, Songhua Chen, Gary Ostroff Baker Atlas, Houston, Texas, USA Duncan Mardon, Scott Dodge EnonMobil Exploration Company, Houston, Texas, USA James A. Lorsong ARCO El-Djazair Company, Plano, Texas, USA ABSTRACT MRIL logs acquired in elliptical or washed-out boreholes can be adversely affected by borehole NMR signal contamination. Washout effect is observed if the sensitive volume of the tool intersects the borehole and is characterized by excessive porosity readings and a distorted T2 spectrum that is not representative of the formation. This paper outlines a procedure for correcting MRIL data that is partially contaminated by a mud signal from the borehole. The correction operates in the time-domain by subtracting a computed estimate of the borehole signal from the measured echo data (formation + borehole) to obtain the signal due to formation fluid(s) only. The amplitude of the modeled borehole signal is estimated by comparing the uncorrected MRIL porosity to independent log-derived porosity. T2 decay parameters of the mud signal are derived from laboratory measurements or MRIL data itself. The corrected echo data are processed by standard methods to obtain T2 spectra and petrophysical parameters therefrom that are representative of the formation. The proposed method has been validated by three case studies. In the first example, borehole-corrected MRTL irreducible water saturation is in excellent agreement with the results of a resistivity-based saturation model, The second example illustrates that the borehole signal can be effectively extracted from log data in low porosity shale sections. The third example shows that the correction can be successfully applied even with relatively poor definition of the borehole signal characteristics. A numerical modeling study illustrates sensitivities of the correction method to uncertainties in input parameters including reference porosity and mud T2 spectrum characteristics. The results illustrate the validity of the method, even at relatively high borehole signal levels, and lead to guidelines for proper application of the correction. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER Z A HIGH RESOLUTION NMR LOGGING TOOL: CONCEPT VALIDATION Joachim TABARY, Marc FLEURY Institut Français du Pétrole, Marcel LOCATELLI Laboratoire d’Electronique, de Technologies et Instrumentation, CEA/DTA Jean-Pierre MARTIN Geo Energy ABSTRACT The paper presents a pre-study of a new NMR logging tool having a vertical resolution of 2.5 cm, much shorter than existing devices (from 30 cm up to 100 cm). Such a tool is especially dedicated to shaly formation and thin beds where the estimation of porosity and saturation is very difficult with resistivity logs. On the technical side, several design difficulties such as very short echo spacing and signal to noise ratios are overcome or enhanced due to the small volume to be excited by the rf field. In a first part, we studied the effect of tool motion on the decay of magnetization. As intuitively obvious, a very small measurement volume is not compatible with a high speed of motion. Using 3D map of static and if magnetic fields, we determined the correction necessary to compensate for the attenuation of the long relaxation components. Three effects were taken into account: the variation of the effective measurement volume during a CPMG sequence, the pre-polarization and the additional dephasing of echoes due to the tool motion. As expected, long relaxation components are mainly affected by the variation of the effective measurement volume and can be fully corrected, while the magnet design provide a very efficient prepolarization. In a second part, we estimated the effect of tool motion on the petrophysical properties derived from T2 distributions calculated from the magnetization decay. We considered the difficult situation where the formation contains a large amount of clay (typically 1<T2<10 ms) and we calculated some parameters of interest: porosity, clay bound water and intergranular pore volume, main peak position. Different signal to noise ratios for a single CPMG sequence were taken into account, from 5 up to 20. Using a large number of different noise realizations, the errors on the above parameters are described statistically. The results indicate, for example, that the clay bound water volume can be found with a reasonable accuracy. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER AA INTERNAL FIELD GRADIENTS IN POROUS MEDIA Gigi Qian Zhang, George J. Hirasaki, and Waylon, V. House Rice University, Houston, TX ABSTRACT Strong internal field gradients were observed on North Burbank (N. B.) sandstones and chlorite/fluid slurries. Evaluation of the effect of internal field gradients on nuclear magnetic resonance (NMR) spin-spin (T2) relaxation is essential to formation evaluation. In this paper, systematic methods are developed to calculate magnetic fields and gradient distributions for three types of porous media. Magnetic fields of real systems like N. B. sandstone and chlorite slurry are simulated and gradient values from simulations are compared with experimental results. For pore lined with clay flakes, field gradients are concentrated around the sharp corners of the clay flakes no matter the orientation of these clay flakes. The radius of curvature of the paramagnetic particle determines the maximum value of the field gradients. Pore lined with clay flakes has the dimensional gradient scaled to the width of clay flake, whereas for cylindrical or spherical systems, it is scaled to the radius of cylinder or sphere. So thin chlorite clay flakes will have much stronger gradients than larger spherical siderite particles. Both N. B. sandstone and chlorite slurry are simulated as a square pore lined with rectangular chlorite clay flakes with the fraction of micropores matching with that of real systems. The field gradients in the micropores of N. B. sandstone and those in between the clay flakes of chlorite slurry are similar. The mean gradient value of the big pore in chlorite slurry is much higher than that of the macropore in N. B. sandstone, Both N. B. sandstone and chlorite slurry have much higher gradients than applied field gradients of logging tools. T, and T2 measurements at different echo spacings were performed on N. B. sandstone at various saturation conditions. Gradient values for the whole pore, micropore and macropore are determined from the slope of the first several data points on the plot of l/T2 vs. τ. Gradient values from simulations using a clay width of 0.2 µm are close to experimental results for the whole pore and micropore. For macropore, simulation matches the mean value of the experiments while individual experiment has larger variation. For chlorite/fluid slurries, simulation result with 0.2 Inn clay width matches well to the mean gradient value of the experiments. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER BB DISTINGUISHING RESERVOIR FLOW UNITS USING NMR HIGH-RESOLUTION PERMEABILITY COMPUTATION: THE OBIGBO NORTH-L0ST EXAMPLE Muyiwa Esho and Roya Simon Shell Petroleum Development Company of Nigeria, Limited- SPDC L.S.D. Onuigbo Schlumberger Oilfield Services, Nigeria ABSTRACT NMR permeability estimations are derived indirectly from T2 distributions, through the estimation of either the log-mean T2 or the bound fluid volume. NMR data is usually stacked (to improve the signal-to-noise ratio) prior to obtaining the T2 distribution. The drawback of stacking is either a reduced vertical resolution or the need for slower logging speeds. A new NMR high-resolution permeability indicator is computed from the sum of echoes in a single phase-alternated pair (PAP) consisting of two Carr-PurcellMeiboomGill (CPMG) sequences. The sum of all echo amplitudes in the echo train is proportional to the product of porosity and the average T2, which correlates to permeability. The sum of echoes has a high signal-to-noise ratio and so can be interpreted without stacking, with a resultant high vertical resolution. The vertical resolution achievable with this technique is equal to the antenna aperture plus the distance travelled during one CPMG sequence plus one polarisation time, the sum of which is a few inches. In a Niger Delta example, SPDC’s Obigbo North-l0ST, the new NMR high-resolution permeability indicator highlighted the absence of reservoir compartmentalisation and permeability barriers within a producing sand unit. There is good correlation between the permeability derived from this technique and permeability derived from core data in another well. The new measurement was compared with standard NMR permeability estimators such as Timur-Coates permeability and also with formation tester mobility measurements. A net-to-gross estimation in laminated formations is also possible with this computation. Reservoir simulation of the target sand using CMR high-resolution permeability shows very encouraging results. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER CC RECENT PROGRESS ON FORMATION RESISTIVITY MEASUREMENT THROUGH CASING P. Beguin, D. Benimeli, A. Boyd, I. Dubourg. A. Ferreira, A. McDougall, G. Rouault, and P. van der Wal Schlumberger ABSTRACT A new wireline logging tool has been developed for measuring formation resistivity in cased holes. The tool sends high current into the casing, which produces a small leakage current into the formation. An array of electrodes, put in contact with the casing, measures the leakage current, which is used to determine the resistivity of the formation behind the casing. The tool was run in producing, injecting and monitoring wells and the data were compared with openhole resistivity logs. The field logs demonstrate that the measurement is repeatable and directly comparable to the formation resistivity recorded at drilling time. The tool data clearly identified the depleted and unswept zones. The cased-hole resistivity logs have received wide acceptance because of the easily interpretable overlay of openhole and cased-hole resistivity curves. Applications of this new measurement include reservoir monitoring in low-porosity or low-salinity formations, formation evaluation when unstable well conditions have prevented the acquisition of openhole logs, and the identification of bypassed hydrocarbons. This paper will present the theory behind the measurement, modeling results of the new tool to better understand its performance in difficult environments, the quality of electrical contact with corroded casings, and the results of field logs in various types of wells. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER DD EARLY RESULTS OF THROUGH CASING RESISTIVITY FIELD TESTS H.M.Maurer, J. Hunziker Baker Atlas ABSTRACT Measuring formation resistivity behind metal casing allows us to apply well-established water saturation evaluation methods to a producing reservoir. This new technology can find widespread applications, especially in low porosity settings where pulsed neutron readings are problematic. However, building a device capable of measuring formation resistivity through metal casing constitutes several major technical and scientific challenges. In cooperation with a consortium of oil companies led by Gas Research Institute (GRI), Baker Atlas has developed a Through Casing Resistivity Tool. The instrument has completed several successful tests under field conditions. The recordings of through casing formation resistivity track the open hole logs very closely. Casing imperfections and collars are successfully compensated. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER EE A NOVEL AND ECONOMICAL PROCESSING TECHNIQUE USING CONVENTIONAL BOND LOGS AND ULTRASONIC TOOLS FOR ENHANCED CEMENT EVALUATION Gary Frisch, Larry Graham, and James Griffith Halliburton Energy Services ABSTRACT Previous conventional cement evaluation techniques that rely on combined data from a traditional acoustic cement bond logging (CBL) tool and modern ultrasonic tools can be problematic. These techniques utilize both the waveform and amplitude measurements from CBL and impedances from the ultrasonic tools to provide analysis of the pipe-to-cement and cement-to-formation bonding. A new interpretation technique has been developed that utilizes both CBL and ultrasonic data to evaluate conventional, complex, lightweight and foam cements. It is important to accurately evaluate the downhole placement and bonding characteristics of any type of cement to ensure zonal isolation of economic fluids from undesirable fluids. Inaccurate evaluation can lead to unnecessary and expensive remedial cementing operations. It is estimated that the industry spends about $200 million per year on remedial cementing. Of this amount, between $30 to $40 million per year is wasted because of misinterpretation of cement evaluation logs. Ultrasonic tools normally require an impedance contrast in the materials behind pipe to differentiate between cement and fluids. The impedance of foam or complex cements can be lower than that of water, drilling mud, or spacer fluid, and can even approach the impedance of free gas. Because of low acoustic impedance, the data and images may indicate fluid behind casing rather than cement even when zonal isolation is achieved. A new interpretation method broadens and refines on previously published methods to effectively evaluate cement with the common cement evaluation tools. The new process directly calculates the level of activity of both the CBL and ultrasonic data. The resulting variance image from the ultrasonic tools allows detection of minor changes in cement or fluid composition and aids in the interpretation of the pipe-tocement bond. Using this technique, several new ultrasonic cement-bond curves have been developed that, when used in conjunction with CBL amplitude data, refine the pipe-to-cement bond assessment The CBL variance will allow differences between free, partially bonded, and bonded pipe to be easily recognized. A revolutionary result is presented from the CBL data alone that highlights this technique in evaluating cement bond in multiple casing strings. Examples are presented showing that the new method is valid and effective in both time and cost Additionally, since the process does not require additional logging passes the expense for the improved evaluation is minimal. The process will work with several different tools from assorted service companies. In the examples, the interpretation is focused to answer the basic question, “Should remedial cementing be performed, or should the well be perforated for production?” st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER FF INITIAL FIELD APPLICATIONS OF A NEW 1.7-INCH PULSED NEUTRON INSTRUMENT W. A.Gilchrist, Jr., R. R. Pemper, D. Treka, E. Frost, Jr., and W. Wilson Baker Atlas, Houston, Texas, USA ABSTRACT We present field examples from our initial experiences with a new 1.7-in, multi-function pulsed neutron instrument. The new hardware system has operating modes making pulsed neutron capture. pulsed neutron spectrometry, pulsed neutron holdup, and neutron activation water flow measurements. The examples illustrate both the C/O and PNC operational modes, with an emphasis on acquisition data flow and quality assurance techniques. Examples also illustrate the interpretation process. Spectral calibration is a critical issue in pulsed neutron spectrometry logging. We describe the techniques used to produce properly calibrated energy spectra. Proper calibration provides high quality data and produces more accurate interpretation results. Log examples illustrate new interpretation techniques including our dynamic response generator for C/O log interpretation. This new method is based on our large modeling data base. The value of log data and interpreted results is strongly related to issues of calibration, data quality, and effective interpretation methods. We discuss each of these issues and present field examples as illustrations of how each is addressed for different operational modes of a new multi-function pulsed neutron instrument. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER GG IMPROVED DETERMINATION OF GAS HOLDUP USING OPTICAL FIBER SENSORS B. Théron, D. Vuhoang, F. Rezgui, G. Catala, D. MeKeon, L. Silipigno Schlumberger ABSTRACT A new production logging device has been field tested that uses innovative sensing technology to enable the direct detection and quantification of gas in multiphase flows. Four optical probes, deployed 90 degrees apart on the arms of a centralizer-like tool, measure the optical reflectance of the surrounding fluid. The probes are evenly spaced in the pipe cross section and their orientation in space is accurately known through use of an integrated relative-bearing sensor. In gas-liquid mixtures, the optical signal reflected by the probe is used to determine gas holdup and a gas bubble count, which is related to gas flow rate. In addition, the individual sensor measurements are used to build an image of the gas flow in the well. These images are particularly useful in deviated and horizontal wells for better understanding the multiphase flow patterns and interpreting their inherent phase segregation occurring at such deviations. The new tool has recently been successfully field tested in wells throughout the world and the tool’s capabilities are illustrated by a number of examples from both field and laboratory data sets. The new tool has been designed to detect the presence of gas, and hence its major application is to identify gas entries in oil/water wells or water/oil/condensate in gas wells. Because of its high sensitivity to minute amounts of gas, the tool can also be used to locate the bubble point when logging in the tubing. The introduction of optical sensing technology in this new tool represents an innovation in production logging. The provided data enable the direct detection and quantification of gas or liquid in multiphase mixtures, allowing the precise diagnosis of well problems and helping in design of production enhancement interventions. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER HH RESERVOIR FLUID STUDY BY NUCLEAR MAGNETIC RESONANCE Matthias Appel, J. Justin Freeman, Rod B. Perkins, Niels P. van Dijk Shell E&P Technology Applications Research, Houston TX, USA Shell Offshore, Inc., New Orleans, LA, USA ABSTRACT The Nuclear Magnetic Resonance (NMR) properties of an undersaturated oil have been measured over a range of temperatures and pressures from ambient to reservoir conditions. This work was performed to determine the extent to which the high solution gas-oil ratio of the reservoir fluid impacts the NMR response. Further, the pressure- and temperature dependence of the hydrogen index (HI) of the reservoir fluid was determined from laboratory measurements for application to fluid volume calculations from wireline NMR data. The measurements show that dissolved gas has a strong impact on the NMR response of reservoir fluids. The NMR relaxation times, T1 and T2, of the “live” oil are roughly ten times longer than those of gas-free “dead” oil under similar conditions and are also a strong function of both temperature and pressure. The hydrogen index of the sample decreased with increasing temperature. At a temperature of 1800 F, a reduction of about 20% was observed compared to its value at 76° F. For pressures above the bubble point of the sample, the hydrogen index is only a weak function of pressure. The experimental data are compared to recently developed models that use the stock tank PVT properties of oil and the composition of the gas phase to predict the hydrogen index of gas-oil mixtures. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER II FREQUENCY EFFECT ON RESISTIVITY INDEX CURVES USING A NEW METHOD. Marc Fleury, Fuhai Liu Institut Français du Pétrole ABSTRACT The frequency dispersion effect is known for more than two decades but there is little experimental evidence for oil/water saturated samples. We provide here laboratory data showing clearly the dispersion of resistivity index curves in the frequency range 100 mHz up to 10 MHz, due to interfacial polarization mechanisms and mainly governed by the pore structure, the surface properties and the fluid distribution (wettability). The method used (FRIM) is a true four electrode technique in the frequency range considered and therefore, the measurements are not influenced by electrode polarization or contact resistance at low water saturation. The measurements were performed on clean sandstones before and after aging with crude oil; on a clay rich sandstone and on a limestone. In general, the slope of the resistivity index curve is decreasing with increasing frequency and the saturation increases the dispersion mechanisms. But there is a large variety of effects depending on the saturation and saturation history. For example, an Archie type resistivity index curve measured on a sandstone at low frequency becomes strongly non-linear in log-log scale at high frequencies. Furthermore, a drainage-imbibition hysteresis can be observed at high frequency whereas totally absent at low frequency. A complex structure as encountered in carbonates yield also more sensitivity to frequency. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER JJ APPLICATION OF NUCLEAR SPECTROSCOPY LOGS TO THE DERIVATION OF FORMATION MATRIX DENSITY Susan L. Herron and Michael M. Herron Schlumberger Doll Research Ridgefield, CT ABSTRACT Formation matrix properties, such as matrix density, can be estimated from the elemental concentrations available from modern, openhole, nuclear spectroscopy logging techniques. Although this estimation is similar to that of mineral-based interpretation frequently practiced today, it can preempt the a priori selection of minerals by solving for matrix properties directly from the elements. This simple approach greatly enhances the ability to perform wellsite interpretations in both simple and complex formations. The interpretation for the matrix density is derived from a comprehensive database containing hundreds of core samples analyzed for both mineralogy and chemistry. The chemical analysis includes not only the major elements, but also the minor and trace elements that significantly influence wireline log responses. These data are used to forward model the matrix which is then solved as a linear combination of four elements (silicon, calcium, iron, sulfur) that are measured by prompt neutron capture spectroscopy. Comparisons are shown between measured and derived matrix density along with statistical measures of goodness of fit. Although in many cases the errors could be reduced by local optimization, the overall agreement is quite good. Although matrix density is empirically derived, the rationale is straightforward. For example, in sandstone, matrix density is approximately equal to that of quartz and feldspar, and it increases as the concentration of calcium- and iron-bearing minerals increases. Therefore calcium and iron heavily influence matrix density. The feldspar minerals are less dense than quartz and are not well sensed by the elements Si, Ca, Fe, and S. Therefore, separate algorithms are presented for non-arkosic, sub-arkosic, and arkosic environments. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER KK HBVI: AN NMR METHOD TO DETERMINE BVI AS A FUNCTION OF RESERVOIR CAPILLARITY Dave M. Marschall, NUMAR Halliburton Energy Services ABSTRACT Current methods used to determine Bulk Volume Irreducible (BVI) water using Nuclear Magnetic Resonance (NMR) log data do not adequately incorporate capillary pressure. The Cutoff-B VI method uses a single value of T2 as a discriminator between water that is bound by capillary forces and fluids that are free to move. The inherent assumption using this model is that a single pore size limits hydrocarbon accumulation. In other words, the smaller pores, below the selected or assumed cutoff pore size, remain at 100% saturation. The Spectral-B VI method eliminates this assumption and replaces it with a model that represents films of water, which are thicker in the small pores and continue to get thinner as the pores get larger. This method is more closely aligned with traditional capillary pressure theory. Yet, in either case, the capillary condition the model represents is often unknown or not considered. This paper introduces a new method, Height-based BVI (HBVI). Using spectral T2, and/or cutoff T2 models, the height-based method allows NMR data to be used to determine the capillary bound fluid saturations at multiple capillary conditions. Implementing this method permits several new or improved applications of NMR data, These include, but are not limited to, improved determinations of hydrocarbon pore volume, more accurate predictions of movable water, prediction and mapping of the reservoir’s free water level(s), pore size, capillary pressure curves, and the prediction of fracture fluid retention that may cause severe permeability reductions. When core material is available, the paper demonstrates how to derive formation-specific HBVI functions. To handle those situations where core data is not available, an alternative method to predict an HBVI function is proposed for sandstone reservoirs. This method recognizes that NMR is highly sensitive to the surface area of the pore system. Sandstones that are more quartz rich have a higher probability of exhibiting low pore surface areas causing them to exhibit weaker surface relaxation. The analyst simply needs to estimate quartz richness, via log analysis techniques or by direct laboratory measurement, and assign an HBVI function to be used. When reliable mineral data is available regarding clay content and minerals known to exhibit fast relaxation times, an improved estimation method is also presented. These simple techniques allow the analyst to determine BVI from NMR logs for a specific capillary pressure or multiple capillary conditions. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER LL COMAPRISON OF RESISTIVITY MEASUREMENT IN LAB WITH COMPUTER MODELING RESULTS AND THEIR COMBINED APPLICATION Jie Gao Di-jun Liu Qi-ning Feng Petrophysical Lab., Dept. of Geosciences, University of Petroleum, Beijing, P.R.C. ABSTRA CT For AC logging, the wave equation derived from Maxwell equations and its computer numerical methods are no problem in theory at all. However, the reservoir rocks are complex multipore media filled with different fluids, and its electrical characteristics cannot be fully outlined by the formation models used in conventional computer forward modeling. The laboratory measurement of the electrical parameters and computer modeling can verify each other, therefore, the comparison between them is present to re-interpret their results so as to obtain the true electrical parameters of formation. The conclusions apparently different from certain known theory are drawn in this paper, some of which have been confirmed by the measurement result in one model well. 1 The rock is one kind of complex multipore media, and its electrical behavior exhibits marked frequency dependence. The conventional modeling methods in which the rock with certain lithology is often treated as homogeneous media could cause large system error, therefore some special attentions should be paid to the study of electrical parameters (mainly the resistivity) in frequency domain. 2 The responses derived from conventional computer modeling are contradictory to the resistivity measurements in lab: the apparent real resistivity from computer modeling ascends as the frequency rises, and the conductivity measured in lab ascends as the frequency rises, Le., the resistivity descends. On the basis of the analysis of those data, we found that the electromagnetic wave propagation effect is only considered and the dispersion of resistivity is overlooked in computer modeling, while the rock resistivity frequency dispersion is mainly depicted in lab measurement. Thus it is naturally understandable that the simple Doll geometrical factor and Born geometrical factor based on homogeneous conductivity background are useful and effective in tool design and data processing 3 The phenomenon that no obvious dispersion of resistivity is found in in-situ LWD 2MHz data as that in lab measurements in some thesis of SPWLA could be interpreted as the different influences of dispersion and propagation effect on responses in this paper, At the frequency of 2MHz more or less, we found that the resistivity dispersion of sedimentary rocks is serious while the propagation effect is much obvious, and the influences of them are partially balanced, thus the resistivity dispersion phenomenon is not observed easily. Such interpretation may conform to the essence of the studied physical phenomenon. At last, the specific steps of the combination of resistivity measurement in lab with computer forward modeling are recommended to study the electrical parameters in frequency domain. And several examples that the responses were logged in one model well have been present to verify our suggestions. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER MM EVOLUTION OF PERMEABILITY-POROSITY TRENDS IN SANDSTONES Philip H. Nelson U. S. Geological Survey, Denver, Colorado ABSTRACT In many sandstone reservoirs, plots of core data show that the logarithm of permeability (k) is linearly proportional to porosity (φ)). Coherent trends usually are found if the data are derived from individual formations and/or depositional facies. I have compiled and replotted published permeability-porosity data, and extracted a trend line summarizing data for each formation. The trend lines represent the evolution of permeability and porosity: from the high k, high φ of synthetic sand packs and recently deposited sands, to the reduced permeability and porosity of consolidated sandstones, and finally to the low permeability, low porosity rocks typified as ‘tight gas sands’. Sandstones of different compositions follow different paths on a log (k)- φ plot. High-quartz sandstones (quartz arenites, with quartz>95%) with quartz cement form the low-porosity bound for all sandstones, preserving relatively high permeabilities at low porosity (<10%). Quartz arenites are typified by porosity values ranging from 4 to 10% at I millidarcy permeability and have the steepest slopes of all sandstones. that is, small porosity changes are associated with large permeability changes. The trend lines for sandstones with lower quartz content plot systematically at higher porosities for a given permeability value than quartz arenites. In artificial sand packs and unconsolidated natural sands, permeability is typically independent of porosity. Diagenesis (total of compaction, cementation, and dissolution) tends to reduce both porosity and permeability to a greater degree in finer grained sandstones than in coarser ones. As a result, most reservoir sandstones display a log( k)-linear φ correlation with positive slope between log(k) and φ . That is, relatively fine grained sandstones typically display lower porosity and permeability than relatively coarse grained sandstones in the same formation. Because the finer grained rocks lose both pore volume and pore size more rapidly than the next higher grain size, the range of permeability values, on a logarithmic scale, broadens as diagenesis progresses, The data fields of individual formations remain fairly coherent, although the slope, intercept, and degree of scatter of these log(k)- φ trends vary from formation to formation. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER NN EFFECT OF WETTABILITY AND WATER SATURATIONON THE DIELECTRIC CONSTANT OF HYDROCARBON ROCKS Ali A. Garrouch Petroleum Engineering Department, Kuwait University Safat, Kuwait ABSTRACT Complex impedance measurements have been performed on sandstone rock samples and bentonite-sand packs in the frequency range 10 Hz to 1 MHz. Effects of water saturation and wettability on the dielectric constant ate investigated both experimentally and theoretically using a generalized Maxwell-Wagner theory that accounts for double layer dielectric dispersion. The measurements appear to indicate that the dielectric constant varies linearly with water saturations above 50%. The rate of change of dielectric constant with saturation is found to be a function of frequency. As the frequency increases this rate of change decreases. The decrease in the slope of the dielectric constant-water saturation profile with frequency is not intuitively obvious, but has been proven theoretically in this work. The dielectric constant of water-wet samples was higher than that of the oil-wet samples at all water saturations. The wettability changes have been modeled by varying the amount of ionic surface charge of rocks. Simulations show that the effect of wettability changes on the dielectric constant is significant. These conclusions are consistent with our experimental results. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER OO ESTIMATION OF SHALE CATION EXCHANGE CAPACITY USING LOG DATA: APPLICATION TO DRILLING OPTIMIZATION Guncel Demircan, John R. Smith and Zaki Bassiouni Craft & Hawkins Department of Petroleum Engineering, Louisiana State University ABSTRACT Shale is the most common lithology encountered in drilling for oil and gas. Shale formations often cause bit performance and wellbore stability problems. This paper introduces a cation exchange capacity model for perfect shale based on Waxman-Smits’ shaly sand petrophysical model. The validity of the model is checked against measured cation exchange capacity. The determination of cation exchange capacity using this modified model and common log data has shown potential for identifying ineffective drilling in shale. Data collected under controlled conditions is needed, however, to optimize the proposed approach. Knowing the location and the characteristics of these shales allows for corrective actions either during drilling, if LWD data is used, or on future offset wells. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER PP A NEW TOOL FOR ELECTRO-FACIES ANALYSIS: MULTI-RESOLUTION GRAPH-BASED CLUSTERING Shin-Ju Ye Halliburton Energy Services Philippe Rabiller Elf Exploration Production ABSTRACT Log facies analysis is important for reservoir characterization, but is made particularly difficult by the problem of “dimensionality”: log space is not equivalent to geological space, and two points that are close to each other in log space may not always be similar geologically. A classic approach to facies analysis, automatic clustering, requires an estimate of the number of clusters, with the results being very sensitive to this parameter. If clustering is tightly constrained, with few clusters, the analyst may find that, because of the problem of “dimensionality”, the resulting clusters cannot easily be used for facies analysis. If log data is relatively unconstrained, the analyst is then faced with the daunting task of linking each cluster to a geological descriptor. Field experience shows that a two-step methodology provides a workable solution. First, one chooses a large number of clusters for automatic clustering. Second, one manually merges small clusters into electrofacies to which geological characteristics are assigned. Even with good visualization tools, performing this task manually in highdimensional (>3) space is still difficult, slow, somewhat subjective, and requires a skill or expertise that is not always readily available. This paper proposes a novel method for electrofacies analysis, Multi-Resolution Graph-based Clustering1 (MRGC), that solves the problem of dimensionality and derives valuable information about the geological facies from the structure of the data itself. MRGC offers all the advantages, while eliminating most of the drawbacks, of the two-step method. MRGC is a multi-dimensional dot-pattern-recognition method based on non-parametric K-nearest-neighbor and graph data representation. The underlying structure of the data is analyzed, and natural data groups are formed that may have very different densities, sizes, shapes, and relative separations. MRGC automatically determines the optimal number of clusters, yet allows the geologist to control the level of detail actually needed to define the electrofacies. This new electrofacies analysis tool has been tested under real-world conditions using conventional logs and NMR T2 distributions, and results from such studies are presented in the paper. In comparison with the existing two-step tool, MRGC has been found to make the work much faster and easier, and is both more direct and more intuitive. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER QQ A PROBABILISTIC APPROACH TO DETERMINE UNCERTAINTY IN CALCULATED WATER SATURATION M.C. Bowers and D.E. Fitz ExxonMobil Exploration Company ABSTRACT Uncertainty in calculated water saturation has a direct economic impact on both exploration and development projects, yet is rarely quantified by petrophysicists. Quantifying the sensitivity associated with each parameter in any water saturation model is required to determine the most cost-effective way to reduce the total uncertainty to an acceptable level and provides the framework for doing value of information calculations. An analytic method to estimate uncertainty in water saturation using the Dual Water model is presented. The method is based on the general formula for error propagation and implicit differentiation to calculate the standard deviation in water saturation from which the 80% confidence level (P10 and P~) is derived. Normal distributions are used to characterize the uncertainty for all log and model parameters. The approach is compared to Monte Carlo methods to assess the advantages and limitations of an analytic approach and the assumptions of normal distributions. An example is presented illustrating how this approach can be used to determine which types of petrophysical measurements provide the greatest reduction in uncertainty for a given cost. The method can be extended to any water saturation model. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER RR ENHANCED PRODUCTIVITY IN A HEAVY OIL FRACTURED RESERVOIR CASE HISTORY FROM THE NULLIPORE OF EGYPT Hani Elshahawi and Theodore Kilmentos Schlumberger Oil Field Services Taher Elzefzaf and Maher Emara The General Petroleum Company ABSTRACT The Nullipore reservoir in the fields of Egypt’s General Petroleum Company (GPC) is a heavy oil reefal reservoir located in the Gulf of Suez area. The Nullipore formation is characterized by a high degree of heterogeneity, which, in addition to the high oil viscosity, has made it difficult to locate and to drain oil effectively. This was further complicated by the lack of accurate seismic data. For these reasons, the reservoir was traditionally considered only as a secondary or backup prospect. Recently, an approach was initiated, based on the use of innovative logging techniques for better reservoir characterization as well as production data analysis for better understanding of reservoir dynamics. This has led to the realization that the reservoir’s productivity is primarily controlled by the flow contribution from natural fractures. New horizontal wells were hence planned and completed on the basis of the newly acquired information with excellent results. Moreover, older wells were recompleted using the data acquired from these horizontal wells, leading to large increases in production. This paper describes the methods used to locate the natural fractures that lead to the majority of production in the Nullipore reservoir. Central to these are the fracture detection logs and log-analysis methods employed. It presents a case study of one well and describes how new wells can be put in place to intercept these fractures and effectively drain the oil. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER SS REAL-TIME QUANTIFICATION OF OBM FILTRATE CONTAMINATION DURING OPENHOLE WIRELINE SAMPLING BY OPTICAL SPECTROSCOPY Oliver C. Mullins Schlumberger-Doll Research Ridgefleld, CT Jon Schroer Schlumberger Special SeMces Belle Chasse, LA Gary F. Beck Vastar Resources Inc. Houston, TX ABSTRACT Acquisition of representative hydrocarbon samples from downhole formations is extremely important to assess reservoir value. Accurate fluid analysis needs to be performed in order to properly estimate reserves and determine optimal production strategies. it is critical to identify the fluids coming from the formation during openhole sampling because fluid analysis quality is adversely affected by sample contamination. This is especially true for oil samples taken in oil-based or synthetic-based mud (OEM) wells, as estimation of OBM filtrate contamination can be difficult and qualitative. In this Paper, we describe a new technique utilizing visible-near-infrared (NIR) spectroscopy which enables quantitative estimations of OBM contamination as a function of sampling time. Significant improvement in sampling efficiency is made possible by analyzing NIR data; prediction of contamination levels can be made by extrapolation of sampling time. Sampling can be terminated early and moved to another depth if high contamination levels are predicted, resulting in significant rig time savings and improved sample quality. We describe methods to use NIR data to give quantitative OBM contamination as a function of sampling time. Several field examples are shown and the significant improvement in sampling efficiency is discussed. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER TT WELLOGML: A PROPOSED STANDARD FOR WEB-BASED EXCHANGE OF DIGITAL WELL LOG DATA Harry Schultz Oilware, Inc. Dan Schenck, Cary Purdy POSC Graeme Wallace Wallace Burnett International ABSTRACT WelILogML is a specification for encoding digital well log data for the Web, It is flexible, extensible and suitable for interaction with external software. WellogML files are easily rendered into high-quality graphics for a variety of media. WeIILogML is a well log specific version of XML, a standard defined by the World Wide Web consortium (W3C). The use of the Internet to exchange electronic business documents (eBusiness) is rowing rapidly. Organizations engaged in eBusiness are quickly converging on the use of XML as the best way to exchange information. The reason for this is that XML is a simple, easy to grasp method of encoding information in plain text. Well log service companies are starting to provide methods for the transfer of well log information using web-based tools. Government and other regulatory bodies are moving towards requiring well log data to be submitted electronically over the Internet. Major companies are distributing information and data via intranet systems. However, in the area of well log data, there is no recognized standard for transfer or exchange via the Internet. WellLogML takes advantage of many of the desirable attributes of the popular Log ASCII Standard (LAS), but provides additional benefits. Similar to LAS data, XML flies contain clear text formatted data so they are human-readable and understandable. However, WellLogML provides support for array curves and for array parameters, which is lacking in LAS. WellLogML also supports numeric and text-based channels, parameter information from multiple runs and both evenly and unevenly spaced indexes expressed in time or depth or any arbitrary indexing scheme. With WellLogML, or any XML file for that matter, there is less need for specific certification software. There are many tools available (free!) that are capable of certifying the structure of XML files. Currently, Internet Explorer 5.0 has built in functionality to structurally certify a WelllogML file. Because XML files separate data content from data display, there are also tools available to support the presentation of the XML data, even within your favorite Internet browser. This characteristic seems tailor made for well log data where petrophysicists can set up their preferred presentation formats for displaying the same data but in different ways. The power of the growing XML marketplace affords the oil and gas community the opportunity to leverage the XML organizations and software that are being created to tremendous advantage. By adopting this new technology, industry can lower the cost of maintaining and exchanging well log data and more easily exploit the Internet revolution to greatly improve the efficiency of doing business. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER UU NEW FACILITIES FOR CHARACTERIZING AND VALIDATING LOGGING INSTRUMENT PERFORMANCE Larry L. Gadeken, Dave Marschall, Gordon Moake, Tim Plasek, John Spallone, and Jerome A. Truax Halliburton Energy Services ABSTRACT New facilities for characterizing logging instruments have been constructed at Halliburton’s Sam Houston Parkway location in Houston. Rocks from former locations on the west side of Houston and in Fort Worth, Texas, were consolidated into two new laboratories. These installations provide distinct, but related, functions. The Tool Response and Characterization Laboratory (TRAC Lab) contains a variety of test formations that are used to determine the basic responses of neutron-porosity, pulsed-neutron, natural-gamma-ray, and gamma-density sondes for both wireline and logging-while-drilling tools. A wide range of borehole conditions and formation properties can be simulated. Two Sonde Acceptance Wells (SAWs) are being constructed with slabs of limestone, sandstone, and dolomite rocks. The SAWs will each contain over 50 vertical feet of rock and have predominantly eightinch-diameter boreholes. The rocks are stacked to simulate both infinitely thick and thin-bed formations. The formation pore space and borehole of one well is filled with fresh water, while the other is filled with 150,000-ppm brine. Ratholes extend 50 feet below the SAWs. The SAWs serve several purposes. Since the rock properties are known, the accuracy and precision of nuclear and magnetic resonance imaging (MRIL) measurements can be investigated under dynamic conditions. The vertical responses of the measurements can also be evaluated. These wells also provide a stable open-hole environment in which to verify that manufactured tools are performing within specifications. Besides nuclear and MRIL tools, acoustic and electromagnetic tools can be logged in the SAWs. Although the latter measurements may not accurately represent the actual rock properties, they will be repeatable. Therefore, complete tool strings can be compared to standard benchmarks. This paper presents a thorough description of the TRAC Lab and SAWs. Illustrations provide additional information regarding the layout and usage of these multi-purpose, multi-sonde facilities. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER VV HARNESSING ADVANCED IMAGE ANALYSIS TECHNOLOGY FOR QUANTITATIVE CORE AND BOREHOLE IMAGE INTERPRETATION Martin A. Kraaijveld Shell International Exploration and Production Willem J.M. Epping Shell U.K. Exploration and Production ABSTRACT Core and borehole images are generally interpreted in a qualitative and interactive manner, predominantly for geological applications. Quantitative use of image data for both geological and petrophysical applications is relatively rare, mainly due to the unavailability of adequate image analysis technology for petroleum engineers. Driven by the increase in computing power and the worldwide effort on image analysis research, sophisticated technology has been developed that allows quick and semi-automatic interpretation of large amounts of image data. This technology, if properly adapted and harnessed, can be of benefit to both petrophysicists and geologists. We have adapted, and where needed further developed, several techniques from the image analysis literature that support the quantitative interpretation of core and borehole image data The development was performed in co-operation with leading University groups. The technology has recently been made available to all Shell Operating Units to allow them to perform their own in-house quantitative analysis. The technology made available to date consists of four modules: 1. Application of new image processing techniques to automatically estimate dip and azimuth. 2. Facies classification based on the extraction of image textural attributes and statistical pattern recognition techniques. Successful applications of this module have been achieved for a variety of (complex) reservoir types. 3. Thin-bed analysis to extract properties such as bed orientation, bed boundaries, degree of anisotropy, and, after proper calibration, shale volume and net-to-gross. These properties feed into programs to invert resistivity logs and determine hydrocarbon saturation. 4. Image enhancement to remove noise and acquisition artefacts. This module is a preprocessing step yielding clean images for input into the previous interpretation modules. Ongoing developments are to establish links to feed image data directly into our static and dynamic reservoir modellng packages, and to perform fracture detection. To allow a quick dissemination of the new technology into our Operating Units, we have initiated a collaborative effort with Z&S / Baker-Atlas to integrate our software into their Recall / Review environment. Recall / Review is available in all Shell Operating Units as one of the platforms for interactive borehole image interpretation and serves as a corporate database for core and log data in others. Due to the modular architecture of the Toolbox, however, implementations on other platforms can relatively easy be developed. To further facilitate the uptake of the technology. dedicated training courses are provided on-site to Shell OUs. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER WW INTEGRATED BED BOUNDARY DETECTION FROM BOREHOLE MEASUREMENTS Raghu Chunduru, Elton Frost, Srinivasa Chakravarthy, Rainer Busch, Zhiyi Zhang, and Alberto Mezzatesta Baker Atlas, Houston, Texas, USA SUMMARY Delineation of consistent bed boundaries from various wireline logs, which have different horizontal and vertical resolutions, is one of the important and challenging aspects of petrophysical interpretation. In general, the characterization of bed boundaries is performed using inflection point, min-max, and neural network methods. Most of these algorithms work reasonably well for individual logs but are limited in use when integrating logs with varying resolution and in delineation of boundaries in transition zones. In this study, we developed a bed boundary algorithm and a layer integration algorithm to overcome some of the drawbacks associated with the existing methods. The proposed method facilitates integration of boundaries derived from instruments with varying resolutions, as well as surface seismic, geologic, and core data to provide a unified set of bed boundaries, that best represents subsurface lithology. The applicability of the proposed method is demonstrated using a field data set from Colorado. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER XX EVALUATION OF LAMINATED THIN BEDS IN FORMATIONS USING HIGH-RESOLUTION ACOUSTIC SLOWNESS LOGS T. Zhang, X. M. Tang, and D. Patterson Baker Atlas, Houston, Texas, U.S.A ABSTRACT This paper describes a signal processing technique that enhances the resolution of formation slowness from the waveform data of an array acoustic tool. The enhanced resolution allows us to detect formation features at a scale from the conventional 3.5 ft, the array aperture, to 0.5 ft, the inter-receiver spacing. This new technique has been applied to various monopole and dipole acoustic data sets logged in laminated formations. Details of formation features such as thin beds or laminations are increasingly resolved by varying the resolution scale from 3.5 to 0.5 ft. The validity of the enhanced resolution is verified using a resistivity image. The features of the 0.5-ft aperture slowness log curve correspond well to the locations of bedding and fracture tips seen from the resistivity image. The high-resolution (0.5-ft aperture) compressional and shear slowness logs can be used to enhance formation saturation evaluation, On a Vp/Vs (compressional to shear velocity) ratio crossplot, the high-resolution slowness data, compared with the conventional slowness data, show a wider spread over the range between dry and wet conditions. This allows a better determination of rocks of various saturation degrees and their locations. For hydrocarbon detection in a formation with thin beds, the high-resolution slowness logs provide a more accurate location of pay zones than the conventional slowness logs. The significant enhancement of the acoustic slowness estimation provides a powerful tool for detecting thin beds in laminated formations and evaluating their acoustic/pertrophysical properties. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER YY RECONCILIATION OF WAXMAN-SMITS AND JUHASZ ‘NORMALISED QV’ MODELS FROM A TENSOR PETROPHYSICAL MODEL APPROACH USING HELD DATA R.A. Mollison Baker Atlas T.V. Ragland Consultant J.H. Schon Joanneum Research, O.N. Fanini Baker Atlas J. van Popta Shell EP Technology ABSTRACT Shale distribution is critical to the correct reservoir characterization of low resistivity, low contrast laminated shaly sands. The tensor petrophysical model determines the laminar shale volume and laminar sand conductivity from vertical and horizontal conductivities derived from electrical anisotopy. The volume of dispersed shale and the total and effective porosities of the laminar sand fraction are determined using a Thomas-Stieber and Juhasz volumetric approach. Removal of laminar shale conductivity and porosity effects reduces the shaly sand problem to a single dispersed shaly sand model to which the WaxmanSmits or equivalent equations can be applied. Hill-Shirley-Klein and later Juhasz defined the cation exchange capacity per unit pore volume, Qv, in terms of effective and total porosity (i.e., bound water saturation, Swb) as it applies to the dispersed shale fraction of the sand as originally proposed by Waxman and Smite. Juhasz further demonstrated that Qv , as a function of Swb, is shale distribution dependent Laminar or bedded shale is a bulk volume conductor. ‘Exchangeable’ cations from these shale laminae are not available to the laminar sand fraction effective pore space and should not be included in the determination of Qv, or dispersed shale bound water conductivity, Cwb, in the sand. A comparison of Waxman-Smits and the Juhasz ‘Normalized Qv ’ (Qvn) dual water models, after removal of laminar shale effects, demonstrates that these models are equivalent if the correct Cwb for the dispersed shale component can be determined directly. Analysis of field data from a new multicomponent induction logging tool using the tensor petrophysical model results are compared to a traditional Juhasz Qvn, dual water model to demonstrate the improvement in reservoir characterization by examining the laminar sand porosity and water saturation utilizing macroscopic electrical anisotropy. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER ZZ WALSH-TRANSFORM-BASED DECONVOLUTION FILTER FOR ENHANCED VERTICAL RESOLUTION Teruhiko Hagiwara Halliburton Energy Services, Houston, Texas, USA ABSTRACT Logging tool responses have historically been plagued by shoulder-bed effects. Over the years, numerous corrections for these effects have been derived and applied with varying degrees of success. This paper presents a new, regularized, deconvolution filter, designed to reduce shoulder-bed effects. Application of the filter shows that it produces better results than the Van Cittert deconvolution filter, which is often used in log inversion. The new filter was created from the Walsh transform of the vertical response function of a candidate logging tool. Because of the characteristics of the Walsh function, construction of the filter is very simple. Regularization is an important feature; because, without regularization, the filter is not local and is severely affected by both boundary data and noise. The filter was initially developed for density logs, because these logs have relatively simple tool response functions. The paper discusses the new filter and examines the advantages and limitations of the filter under different conditions. To demonstrate the class of formation profiles to which the filter may be applied, the paper presents examples of the effect of filter length and the choice of the regularization parameter on the quality of the deconvolution. Because noise in the data can have a significant impact on the robustness of a deconvolution, this effect is also discussed. Synthetic log examples using logging data sampled at 1/2-ft spacing and with 2-ft vertical resolution show that, with the proper regularization parameter, some formation profiles may be recovered with a vertical resolution close to one foot. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER AAA INTERPRETATION OF INDUCTION LOGGING DATA IN HORIZONTAL WELLS M.Rabinovich, D.Beard, I.Geldmacher, L.Tabarovsky Baker Atlas, Houston, USA M.Fidan Phillips Petroleum, Norway ABSTRACT We have developed a new technique for interpreting induction logging data in horizontal wells. In addition to the accurate formation and invasion resistivity distribution near the borehole wall, the new inversion algorithm allows us to determine the distance to remote layers and their resistivities. The High Definition Induction Logging (HDIL) instrument collects data at multiple frequencies and various transmitter-receiver spacings. Focusing and inversion algorithms are designed for vertical and deviated wells to determine an invasion profile, to measure resistivity deep into the formation, and to provide high vertical resolution. In horizontal wells, the objectives are different. In addition to the resistivity distribution in the borehole vicinity, we wish to determine distances to remote cap rocks and water-bearing horizons. The vertical resolution (or, more accurately, the resolution along the borehole trajectory) is no longer important due to the relatively small lateral variation of the formation parameters. What becomes important is the depth of investigation. Low operating frequencies and long transmitter/receiver spacings allows the HDIL tool to provide reliable information from layers located up to 20 ft away from the instrument. The new inversion algorithm for interpreting induction logging data in horizontal wells consists of three components. First, we determine the parameters of the near zone formation using shallow and medium investigation measurements. At this stage, fast 2-fl inversion allows us to recover invasion and formation parameters without being affected by remote layers. Second, we correct the medium and deep measurements for the presence of the borehole and invasion using the results of the near zone interpretation. Third, we interpret the corrected medium and deep measurements using l-D layered inversion to characterize remote layers. We validate the new approach with a synthetic model consisting of a borehole, invaded formation, and a remote layer. All parameters of interest, such as formation and invasion resistivities (Rt and Rxo), invasion depth (Rxo), and the distance to the remote layer, are recovered with high accuracy (errors less than 10%). We also present a case study for a horizontal well in the North Sea. Successful completion of the well required distinguishing between low resistivity water-flooded zones with movable water and an underlying tight layer that exhibits low resistivity. The developed algorithms allow quantitative estimation of the distance to, and resistivity o~ the tight layer as well as the resistivity of the permeable formation. The distance to the tight layer correlates with information from seismic data. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER BBB OPTIMAL EVALUATION OF FORMATION RESISTIVITIES USING ARRAY INDUCTION AND ARRAY LATEROLOG TOOLS R. Griffiths, T. Barber. O. Faivre Schlumberger ABSTRACT One of the perennial formation evaluation questions is that of selecting the optimal technique for determining formation resistivities. Since the early days of well logging two techniques have been in common use, induction and laterolog. Each has strengths and limitations that tend to complement the other. We review the latest selection methods in light of recent advances in both hardware and processing. Developments in both induction and laterolog technology have advanced the state-of-the-art from relatively simple early tools to the sophisticated array measurements and associated post-acquisition processing available today. With these improvements the “overlap zone,” where environmental conditions allow valid measurements by both induction and laterolog tools, has increased. Comparison of the tool responses in the overlap region displays the complementary nature of the measurements. The primary selection criterion is based on borehole effects on the measurements. We have developed a Web-based job planner which compares the operating ranges of the array induction and array laterolog tools. With the improved information content of the array measurements we can better define borehole effects and thus either flag or more accurately correct the data. Development of environmental log quality control for the induction measurements helps identify when the tool is out of its range. The result is improved accuracy and confidence in the resulting formation resistivities. This paper reviews the techniques and the improved data quality they yield. Sensitivity to hitherto neglected environmental effects such as anisotropy and shale alteration have made comparison between induction and laterolog tools difficult in many situations. With the number of measurements made by array tools. one can now use the different sensitivities of laterolog and induction measurements to sort out these environmental effects. In addition, advanced processing such as maximum entropy and 2D inversion helps explain apparent differences in the formation resistivities derived from the two measurements. Several field log examples show both the practical application of the selection methodology and the effect of post-processing on formation resistivity evaluation. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER CCC A FAST TWO-STEP PROCEDURE FOR THE APPROXIMATE 3D MODELLING OF LWD INDUCTION TOOLS R.G. Hakvoort, M.R. Spalburg, Shell International Exploration and Production BV., The Netherlands ABSTRACT Proper evaluation of Measurement While Drilling (MV/fl) or Logging While Drilling (LWD) induction tool measurements requires the availability of a fast and accurate 3D forward model, incorporating the effects of the drilling process (borehole, mud filtrate and invasion) and the formation (layering, apparent dip). Presently available 3D forward modelling methods for LWD induction tools are based on a finite element or finite difference approach. Unfortunately, these methods are far too slow to be effectively applied in resistivity log inversion or modelling, particularly if an on-line analysis of the LWD induction logging data is required (FEWD or Formation Evaluation While Drilling). In this paper a new modelling procedure for LWD induction tools is introduced, the so-called “two-step’ procedure. This new procedure yields a speed improvement of about a factor 100, at the expense of some loss in accuracy. The procedure consists of two steps. In the first step the influence of borehole and invasion is computed with a lD induction code. In the second step the influence of shoulder beds and dip is computed with another 1D induction code. Simulations show the good performance (i.e. high speed and accuracy) of the new two-step procedure. In addition, a field example is shown of the applicability of the new and fast 3D LWD induction tool modelling method. In this field example the two-step method is combined with another powerful method, the turbo-boost inversion method. It is shown that much insight in the true formation resistivities is obtained by means of this titanic combination. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER DDD DEVIATED-WELL SOFTWARE FOCUSING OF MULTIARRAY INDUCTION MEASUREMENTS Jiaqi Xiao, Ingo Geldmacher, and Michael Rabinovich Baker Atlas, Houston, TX, USA ABSTRACT In highly deviated wells or when the relative dip angle between formation layering and wellbore axis is large, array induction measurements exhibit erratic spikes, misleading curve separations, and inaccurate resistivity values, preventing log analysts from accurately evaluating invasion and formation resistivities. To address these problems, various correction methods and inversion techniques have been developed. The correction methods, however, only yield satisfactory results when the relative dip angle is low to moderate, and inversion techniques are typically very time consuming. Conventionally, dipping bed effects are considered in terms of a charge effect and a volumetric effect. As even more complex earth models are considered, we find that formation anisotropy also exaggerates the dipping effect, manifested by misleading curve separations in the array instrument readings. Our newly developed deviated-well software focusing (DSF) method simultaneously accounts for all these dipping effects. The DSF method is derived from the Born approximation. The induction response is separated into two portions: a background response and a perturbation response. An inhomogeneous, anisotropic background formation model is used to calculate the background response, and the perturbation response is interpreted through a software focusing technique. The combination of the two solutions is the final result. After description of the theory and methodology, this paper presents synthetic and field examples. We show that our method significantly reduces spurious separation between shallow and deep reading curves and minimizes the confusion between apparent and real invasion. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER EEE FIELD TEST OF AN EXPERIMENTAL NMR LWD DEVICE M.G. Prammer, G. D. Goodman, S. K Menger, M. Morys and S. Zannoni NUMAR, Hailiburton Energy Services J. H. Dudley Sperry-Sun, Halliburton Energy Services ABSTRACT An experimental prototype of the Magnetic Resonance While Drilling tool (MRWD) underwent a successful field test campaign during the summer of 1999. The tool accumulated 130 hours of downhole use, including periods of harsh drilling conditions. The hardware was designed to be compatible with the MRIL wireline devices at the data processing and interpretation level. The testing program included the acquisition of conventional triple-combo and MRIL wireline logs. MRWD was run while drilling and after drilling, while stationary, and during sliding and tripping operations. The comparison of wireline MRIL and MRWD data shows that the MRWD provides source-free porosity, free-fluid volume and permeability data. The after-drilling data acquisition mode of MRWD replicates all the essential aspects of the MRIL wireline log. This paper examines the design considerations for acquiring magnetic resonance data in an LWD environment, describes the hardware used in the field test, and includes an analysis of the collected field test data. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER FFF GEOLOGICALLY-ENHANCED VERTICAL RESOLUTION OF NMR ECHO-TRAIN DATA Tarek Tutunji, T. Hagiwara, and Peter I. Day Halliburton Energy Services, Houston, TX ABSTRACT Petrophysical analysis of thinly laminated sequences with well logs is a challenging task. Because the vertical resolution of most logging tools may be very much coarser than the thickness of individual lamina, these tools usually record only averaged formation properties. Consequently, the log interpretation leads to a description of the formation that may seriously misrepresent the productive potential of the formation. This averaging is particularly problematic in thinly laminated sequences that consist of highly permeable, porous sand layers with less permeable silt or essentially impermeable shale layers. The properties of the sand-layers are masked by the silts and shales and thus are difficult to estimate from the log data. A new method has been developed for interpreting Nuclear Magnetic Resonance (NMR) echo-train data in thinly laminated sequences. To enhance the vertical resolution of the echo-train data, geological information with a higher vertical resolution is needed. Electrical micro-imaging devices can provided such high-resolution data, for example,. With the new method, echo-trains specific to particular lithology types can be determined, such that the typical T2 spectra of lithological laminae can be constructed and NMRbased permeability estimates subsequently obtained. These permeabilities are usually significantly larger than those obtained from “averaged” data. Besides NMR data, the method is applicable to multidimensional data from other logging tools; for example, the method can be applied to data from a pulsedneutron thermal-decay device. This paper describes the new method, examines its features through application to synthetic logs, and presents several field examples. In the field examples, the macroscopic correlation between the electrical micro-imaging resistivity data and the T2 bin distributions was used to derive the high-resolution geological information. From the lithology-specific information obtained for the laminae, echo-train data and permeability logs at l/2-ft, enhanced vertical resolutions were generated. These high resolution permeability logs predicted significantly enhanced permeabilities in the laminated sequences. The consistency of the algorithm was tested by comparing the reproduced log to the actual NMR log data that had been acquired at 4-ft resolution. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER GGG CARBONATE EVALUATION USING NMR TIME DOMAIN ANALYSIS Dan Buller Halliburton Energy Services ABSTRACT The log interpretation problems posed by the complex “texture” of most carbonates is a challenge faced by many petrophysicists on an almost daily basis. The variable “w” Archie solutions (m = n = w) using dielectric measurements seemed to show great promise in the 1980’s and early 1990’s. However, they have proven unreliable in productive gas reservoirs composed of mixed carbonate facies. This paper shows that a robust NMR-based technique can successfully address any of the shortcomings in applying the variable “w” interpretative technique in fresh filtrate invaded carbonates, or alternately, directly image gas and oil in non-invaded carbonates. The NMR-based technique uses enhanced T1 discrimination of fluids available from the latest generation field NMR tools to either solve the textural problem in the fresh filtrate case, or to obtain Sw, Sg, and So in the non-invaded case. NMR measurements consisting of concurrent 1-sec and 14-sec recovery time data are used to correct for hydrogen index and polarization effects in order to obtain a mineralogy-independent total porosity. The NMR measurements are also used to generate, through an automated NMR Time Domain Analysis (TDA), a direct determination of gas, oil, and water volumes present in the zone of investigation of the NMR log. Because this zone is generally defined as being in the flushed region surrounding the borehole, the NMR measurement also supplies an Sxo estimate that is independent of resistivity. This resistivity independent measurement of Sxo can be used to perform a backward calculation of “w” through the Archie equation. The calculation requires an appropriate value of fluid resistivity in the flushed zone and a resistivity measurement obtained from a shallow-reading resistivity tool with a zone of investigation comparable to the NMR log. Two carbonate examples are presented and discussed in-depth. The first is a gas-filled low-resistivity pay, Lower Cretaceous Rodessa limestone from Northern Louisiana; and the second is a gas- and oil-filled Jurassic Smackover dolomite from East Texas. TDA-based interpretations are demonstrated for both examples and are shown to produce results that are consistent with production. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000 PAPER HHH SATURATION ESTIMATIONS FROM MAGNETIC RESONANCE MEASUREMENTS IN CARBONATES Charles Flaum Schlumberger Doll Research, Connecticut Udit Guru, Supratik Banerjee Schlumberger, Reservoir Evaluation(Wireline), India ABSTRACT: The applications of magnetic resonance techniques for evaluation of rock parameters are well known. These parameters representative of the reservoir can be estimated reliably by logging the formation with shortened echo spacing, TE. The use of NMR as a direct indicator of the presence of hydrocarbon using techniques such as the differential spectrum method (DSM, Akkurt et al, 1995) has been quite effective. DSM performs successfully in gas or light oil. This is because of the appreciable T1 contrast between the brine and the hydrocarbon phase. The enhanced diffusion method (EDM, Prammer et al, 1995) exploits the diffusion contrast between the oil and the water, separating their NMR signals and using long T5. The long echo spacing emphasizes the diffusion contrast between water and oil in the formation. We have developed a quantitative time-domain technique to analyze the CMR enhanced diffusion data. This paper describes the results of an enhanced diffusion study carried out in a carbonate field in India. We recorded the magnetic resonance information in this well using both the diffusion and standard rock quality passes. The rock quality pass was used to estimate the rock quality measurements such as lithologyindependent porosity, permeability, and bound and free-fluid volumes. We estimated hydrocarbon saturation from the NMR using the passes from the long and the short echo spacing. The modeled results confirmed these estimates. We used a new generation wireline formation-testing tool to carry out formation pressure measurements based on the magnetic resonance results alone. The drawdown permeability from the wireline pressure measurements matched well with the permeability estimated from the magnetic resonance logs. We used the pump-out module of the formation-testing tool to collect formation samples and estimate the fluid contact, picking the pressure/sample points from the NMR data alone. Standard openhole logs recorded after we acquired the magnetic resonance and pressure data agreed well with the saturation estimated from the shallow-resistivity device. st SPWLA 41 Annual Logging Symposium, June 4-7, 2000