XXVI Reunión Nacional de Mecánica de Suelos e Ingeniería Geotécnica Sociedad Mexicana de Ingeniería Geotécnica, A.C. Noviembre 14 a 16, 2012 – Cancún, Quintana Roo Field Gmax relationships for Bay of Campeche clay Relaciones del Gmax de campo para la arcilla de la Bahía de Campeche Victor M. TABOADA1, Kuat C. GAN2, Diego CRUZ3, Procoro BARRERA3, Esteban ESPINOSA4 and Daulet CARRASCO4 1NGI, Inc., Houston, Texas, USA. Marine Geosciences, Inc., Houston, Texas, USA. 3Instituto Mexicano del Petróleo, México, D.F., México. 4Petróleos Mexicanos, Ciudad del Carmen, Campeche, México. 2Fugro-McClelland RESUMEN: Uno de los más importantes parámetros de suelo requeridos en el análisis de los pilotes de las plataformas marinas fijas sometidas a carga lateral sísmica es la velocidad de onda cortante (V s) del suelo. Debido a que se realizan mediciones in-situ de Vs en un número limitado de sitios de plataforma, existe la necesidad de desarrollar correlaciones para estimar Vs basándose en propiedades básicas del suelo. Para cubrir esta necesidad se estableció una base de datos con mediciones in-situ de Vs y propiedades básicas de la arcilla. Se recolectaron datos de 11 investigaciones geotécnica costa afuera realizadas para PEMEX para el diseño e instalación de plataformas marinas fijas en la Bahía de Campeche. La base de datos se diseño para desarrollar correlaciones empíricas entre el módulo de rigidez a pequeñas deformaciones (10-4%) in-situ (Gmax) y la resistencia al corte no drenada, esfuerzo vertical efectivo, contenido de agua, relación de vacíos y relación de preconsolidación, usando análisis de regresión simple y de regresión múltiple. Se recomiendan las correlaciones empíricas desarrolladas mediante análisis de regresión múltiple para proporcionar un medio de determinar la mejor estimación de Gmax o Vs de arcillas cuando no existen mediciones in-situ de Vs. ABSTRACT: One of the most important soil parameters required in the analyses of the piles of oil platforms subjected to lateral earthquake loading is the shear wave velocity (Vs) of the soil. Since in-situ measurements of Vs is part of the scope of work of a limited amount of platform sites, there is the need to develop site specific correlations to estimate Vs based on basic soil properties. To cover this need a database with in-situ measurements of Vs and basic soil properties of clay has been established. Data were collected from 11 offshore geotechnical investigations performed on behalf of PEMEX for the design and installation of fixed offshore platforms in the Bay of Campeche. The database was tailored to developed empirical correlations between the in-situ small-strain (10-4%) shear modulus (Gmax) and undrained shear strength, effective vertical stress, water content, void ratio and overconsolidation ratio, using simple regression analyses and multiple regression analyses. The empirical correlations developed using multiple regression analyses are recommended to provide a mean of determining the best estimate Gmax or Vs in clay when in-situ measurements of Vs are not available. 1 INTRODUCTION 1.1 Background The Bay of Campeche is located in the large bay comprising the southern portion of the Gulf of Mexico between the Yucatan Peninsula to the east, the Isthmus of Tehuantepec to the south, and the coast of Mexico at Veracruz to the west. The Bay of Campeche is limited to that section enclosed approximately by 91o W. Longitude on the east to 95o W. Longitude to the west and 20o N. Latitude on the north to the Mexican coast on the south, as indicated by the rectangular area in Figure 1. The Bay of Campeche covers an area of about 15,540 square kilometers, and is the largest oil field in Mexico which has prompted an extensive installation of shallow water fixed-pile platforms. The oil platforms installed in the Bay of Campeche must be designed against earthquake loading since the piles can be subjected to large seismic lateral loadings. The dynamic structural analyses of the oil platforms require acceleration time histories or acceleration spectra that already include the soil amplification of the earthquake motions. One of the most important input dynamic soil properties of the seismic response analyses performed to evaluate the soil amplification is the small-strain shear modulus Gmax. Although, there has been important advances in laboratory testing to obtain the small-strain shear modulus (less than 0.0001 percent) Gmax, the adverse effect of unavoidable sample disturbance and reconsolidation, as well as difficulty in quantifying the aging effect on Gmax, renders in-situ shear wave velocity measurements the currently SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. 2 Field Gmax relationships for Bay of Campeche clay preferred method for evaluating Gmax in-situ. Unfortunately, it is not always possible to make insitu measurements of Gmax or shear wave velocity, Vs, at all investigated locations, and the geotechnical engineer is left to use empirical correlations developed mainly for onshore soils with different geological settings than those of the Bay of Campeche clay. Thus, to take advantage of the abundant basic soil properties and in-situ measurements of shear wave velocity, correlations between Gmax and basic soil properties are needed for the Bay of Campeche clay soils. Regional correlation studies of in-situ measurements of shear wave velocity with other index test or engineering properties of the soils have been shown to be useful for estimating shear wave velocity profiles at sites lacking direct in-situ measurements of Vs. In an effort to establish a database for the evaluation and modeling of shear wave velocities of clay soil units in the Bay of Campeche, data have been collected from offshore soil investigation performed between 2004 and 2009. The in-situ shear wave velocity data were collected from measurements made with downhole P-S suspension seismic velocity logger. Data from a total of 22 boreholes at 11 sites in the Bay of Campeche are incorporated in this study. Figure 1. Location of the Bay of Campeche area. 1.2 Objectives and scope of work Collect, analyze and develop a high quality geotechnical engineering soil properties database using available site investigation data from the Bay of Campeche. Use the database to determine correlations between index and strength test parameters and in-situ small-strain shear modulus to aid geotechnical engineers in obtaining profiles of insitu Gmax specific for the Bay of Campeche clay. 2 GEOLOGY AND STRATIGRAPHY The Bay of Campeche is an isthmian embayment extending from the western edge of Campeche Bank to the offshore regions just east of Veracruz (~96 o W. Longitude). The Sierra Madre Oriental forms the south-southwestern border, and the associated coastal plain is similar to the Texas-Louisiana coast in the northern Gulf. The bottom topography is characterized by long ridges parallel to the exterior of the basin. Salt domes are prevalent in the region, and the upward migration of salt is theorized to be a cause of the complex bottom profiles. Similar to the northern Gulf, large quantities of oil are produced here, and thick terrigenous sediments predominate. The surface sediments of the Bay of Campeche are terrigenous; predominantly clays and silts, which are transported by abundant streams of the Grijalva, Usumacinta and San Pedro rivers and the mouth of the Laguna de Terminos. The terrigenous are transported and distributed on the seafloor accompanied with sand. From the mouth of the Grijalva River, where the movements of the sea are turbulent is the deposition of uniform fine sand. Offshore flow behaves like laminated and low speed, allowing the deposition of sediments from the size of the clay. In general, at the seafloor there is a layer of very soft to soft olive gray clay, with shell fragments with a thickness ranging from 5 to 25 m. However, there are areas with presence of granular soil; these areas almost always correspond to the mouths of the rivers of the area and surfaces where coral formations emerge or areas close to the coast where the waters are shallow. An example of this is the coast located in front of Ciudad del Carmen, where the seafloor sediments consist of a dense layer of sand about 10 m thick overlaying a firm to hard clay soil. The surficial clay layer reaches its largest thickness in the southwestern part where it reaches 25 m. The thickness of the clay layer reduces to the Northeast, allowing the underlying sand to appear superficially to the edge of the continental shelf. Another important feature of the stratigraphy of the Bay of Campeche is that below the sand layer that underlies the surface clay layer is an alternating sequence of clays and sands. The former vary from very firm to hard clays, and the latter are medium dense to very dense sands. SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. TABOADA V. et al. 3 BAY OF CAMPECHE DATABASE 3.1 Introduction The database contains 11 sites in the Bay of Campeche, Gulf of Mexico, Mexico where offshore geotechnical investigations were performed on behalf of PEMEX as part of geotechnical campaigns conducted between 2004 and 2010. The water depth in these sites varied between 14.0 m and 102.3 m. An overview of the sampling and in-situ testing performed at these sites is presented below. The soil conditions at the 11 sites were investigated by drilling and sampling to a termination depth of 121.9 m below the seafloor. In addition to drilling and sampling, piezocone penetration tests (PCPTs), remote vane tests and P-S suspension seismic velocity logging were performed. At each site, the soil investigation was completed in two combined borings. The first boring was drilled to allow sampling at close intervals of about 0.9 m from the mudline to 12.5-m penetration and about 1.5 m from 12.5- to 24.5-m penetration. A second boring (primary) was drilled to perform sampling, remote vane testing and piezocone penetration testing from 3.0-m penetration to the termination depth of the boring. Following completion of the drilling, sampling and in-situ testing activities, downhole P-S suspension seismic velocity logging was performed in the primary boring from 3.0 m to 121.9 m below the seafloor. In addition, nearseafloor in situ vane tests were performed in the immediate vicinity of the borings to a depth of 6.1 m below seafloor. The following three sections provide details related to in-situ measurements of shear wave velocity and a summary of the soil properties included in the Bay of Campeche database. 3.2 In-situ measurements of shear wave velocity The 11 sites in the Bay of Campeche contain in-situ measurements of shear wave velocity every 0.5 m over the depth interval of 3.0 to 121.9 m below seafloor using downhole P-S suspension seismic velocity logger. The digital P-S Suspension Probe, roughly 8.5 m in length, contains a seismic source and two receivers. The receivers, or geophones, are spaced 1.0 m apart from each other and are separated from the seismic source by a series of isolation tubes. The whole probe is suspended by an armored, 4conductor cable that serves both to support the probe and to convey data to and from the recording device and the laptop, on the surface. The probe is lowered into the borehole to a specified depth (a rotary encoder on the winch measures probe depth), where the source generates a pressure wave in the borehole fluid. The pressure wave is converted to seismic waves (P and S) at the borehole wall. Along the wall at each receiver location, the P and S waves are converted back to 3 pressure waves in the fluid and received by the geophones, which send the data to the recorder on the surface. The elapsed time between arrivals of the waves at the receivers is used to determine the average velocity of a 1-meter-high column of soil around the borehole. A schematic of the P-S Suspension Logging System is presented on Figure 2. Figure 2. P-S suspension logging system. 3.3 Soil properties included in the database The compiled database contains index and engineering properties obtained from classification tests, strength tests and consolidation tests as indicated below. The database includes index properties such as total unit weight, water content, specific gravity, void ratio and Atterberg limits. And engineering properties such as undrained shear strength derived from unconsolidated-undrained (UU) triaxial compression tests and in-situ remote vane tests, in-situ shear wave velocity, in-situ small-strain shear modulus, insitu effective vertical stress or effective overburden pressure estimated from the submerged unit weight, preconsolidation pressure (or maximum past pressures) interpreted from the consolidation test results obtained from one-dimensional incremental loading test or constant rate-of-strain consolidation test using the Casagrande (1936) method and the Work Per Unit Volume Method proposed by Becker et al. (1987), and overconsilidation ratio (OCR). SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. 4 Field Gmax relationships for Bay of Campeche clay 3.4 Summary of database soil properties Figure 3 to Figure 9 present summary plots for the combine data set from all sites in the database. Natural water content (W) range is from 20 to 85% with a majority of the values being between 40 to 60% (Figure 3). The liquid limit (W L) varies between 30 and 140% with most of the values being 80% (Figure 4). The plastic index (Ip) falls between 15 and 110% with most of the values being in the range of 50 to 60% (Figure 5). Figure 6 presents a Casagrande Plasticity Chart for the data from all the sites. The data show a wide range of values with most of the data plotting in a fairly tight band above the A-line and close to the U-line indicating predominantly high plasticity clay type of soils. The OCR varies between 1 and 7 with most of the values falling between 1 and 2.5 indicating that most of the soil samples are normally consolidated to lightly over consolidated (Figure 7). Figure 5. Histogram of plastic index of database content. Figure 6. Casagrande plasticity chart of database content. Figure 3. Histogram of water content of database content. The in-situ shear wave velocity (Vs) is in the range of 75 to 535 m/s with the majority of the values in the range of 200 to 350 m/s (Figure 9). Figure 4. Histogram of liquid limit of database content. The shear strength data is concentrated in two groups (Figure 8), one with shear strength from 5 to 40 kPa, with most of the values between 20 and 30 kPa, representing the soils near the seafloor, corresponding to the first layer that occurs in virtually all the Campeche Sound. In general, these soils are normally consolidated. The other Group with shear strength from 50 to 450 kPa, with concentration of values between 100 and 300kPa, occur in deeper strata. Figure 7. Histogram of overconsolidation ratio of database content. SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. TABOADA V. et al. Figure 8. Histogram of undrained shear strength of database content. 5 where s’o and Gmax are in kPa. Increasing the void ratio has two effects: (1) the density tot decreases, which should tend to increase Vs; and (2) the modulus Gmax decreases, which should tend to decrease Vs. Their data showed that the decrease in shear modulus outweighs the decrease in density and that for fixed s’o, Vs decreases when the void ratio gets larger. Hardin and Black (1968) found that equation (2) originally developed for sand, also gives reasonable results for Gmax of normally consolidated clays having void ratios up to 1.5. Hardin (1978) modified equation (2) to consider clays of high surface activity and high plasticity index having a void ratio greater than about 1 or 1.5, and also allowed for the stiffening effect of overconsolidation ratio (OCR). The equation proposed by Hardin is: G max 6,200 OCR k ( o' ) 0.5 0.3 0.7eo2 (3) where s’o and Gmax are in kPa. The effect of the plasticity is incorporated in equation (3) via the exponent k which is a function of the plasticity index, Ip (k = Ip0.72/50≤0.5). The final more general equation is provided below (Vucetic and Dobry, 1991): Figure 9. Histogram of in-situ shear wave velocity of database content. 4 SIMPLE REGRESSION ANALYSIS 4.1 Background The direct measurement of the in-situ shear wave velocity provides the most reliable and certain means of obtaining the shear modulus at low shear strains (less than 10-4%) Gmax. It can be calculated as a function of total mass density (tot) and shear wave velocity as: G max totVs2 (1) where rtot = gtot/g, gtot = total unit weight, and g = gravitational acceleration constant = 9.80 m/s2. Since the total mass density can be estimated with reasonable accuracy, the shear wave velocity or shear modulus at low shear strains can essentially be used interchangeably. Hardin and Richard (1963) using resonant column tests on sand illustrated that the main parameters controlling Gmax are void ratio (eo) and mean effective stress [s’o=1/3 (s’1+s’2+s’3)] and proposed the following expression for angular clean sands and crushed quartz silt: G max 3,200 2.97 eo 2 1 eo o' (2) G max 625Fe OCR k ( p a o' ) 0.5 (4) where pa is the atmospheric pressure in the same units as Gmax, and Fe is a void ratio function different for different type of soils. More recently, Jamiolkowski et al., (1991) have suggested Fe = 1/e1.3. The empirical equation (4) was developed based on laboratory data. It is therefore of interest to developed similar empirical correlations based on insitu measurements of Gmax (or Vs). Simple regression analyses are used in this section to provide as a first approximation expressions for Gmax in terms of one variable including void ratio, effective vertical stress and undrained shear strength. Expressions for Gmax in terms of several variables are presented in the next section. 4.2 Correlation with void ratio The void ratio of a soil is related to the water content by: eo Gs w 1 w tot 1 (5) where Gs = specific gravity of soil solids, gw = unit weight of the water, w = water content, and gtot = total unit weight. The void ratio at a given depth was computed using equation (5), the measured water content at the depth of interest, and interpolated values of Gs and gtot at the same depth, were SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. 6 Field Gmax relationships for Bay of Campeche clay estimated from developed profiles of Gs and gtot based on measurements made at selected depths. The effect of void ratio on Gmax is presented on Figure 10. It is observed that Gmax decreases as the void ratio increases. The relationship between void ratio and in-situ Gmax developed through least square regression analyses to the 243 data pairs (G max and eo) is: G max 1280 exp( 2.162eo ) (6) where Gmax is in MPa, and the coefficient of determination (r2) is 0.659. In the above equation exp(x) is the function ex, where e = 2.718. 4.3 Correlation with in-situ effective vertical stress The relationship between in-situ effective vertical stress (s’vo) and in-situ Gmax for all the sites is presented in Figure 11. The data follow a clear trend with the Gmax increasing as the effective vertical stress increases. The least square regression analyses to the 243 data pairs (Gmax and s’vo) gave the following relationship: G max 870( vo' ) 0.832 (7) where Gmax and s’vo are in kPa. The coefficient of determination is 0.859, indicating that 85.9% of the observed variability in Gmax is attributable to s’vo. Figure 11. In-situ Gmax related to effective vertical stress (s’vo). A total of 183 data pairs (su and Gmax) were used in the development of the correlation between s u and Gmax. A simple regression analysis of this existing data yielded the following relationship between su and in-situ Gmax. G max 1325S u0.881 (8) where su and Gmax are in kPa, and the coefficient of determination is 0.913. The Gmax versus su relationship is shown in Figure 12. Figure 10. In-situ Gmax as a function of void ratio (eo). 4.4 Correlation with undrained shear strength The values of undrained shear strength (s u) used to establish the correlation with Gmax have been determined by unconsolidated-undrained (UU) triaxial compression tests and in-situ vane tests. Figure 12. In-situ Gmax as a function of undrained shear strength (su). SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. TABOADA V. et al. 7 5 MULTIPLE REGRESSION ANALYSES Multiple regression analyses were also conducted in log-log format to provide power function expressions for Gmax in terms of several variables. A summary of the most prominent correlations derived using multiple regression analyses is presented in this section. The relationship with the highest correlation using three variables was: G max 942S u0.377 ( vo' w ) 0.428 (9) where Gmax, su and s’vo are in kPa, and water content (w) in decimal form. The coefficient of determination (r2) is 0.953, and a total of 183 datasets (Gmax, su, s’vo and w) were used in the multiple regression analyses. This equation is recommended to estimate the in-situ Gmax due to its high correlation and simplicity, and to the fact that when the overconsolidation ratio was further added no improvement in the correlation was attained. Figure 13 present a comparison between in-situ Gmax measured in the field with suspension logging and the expression given by equation (9). In general, all the predicted values are between the bands of ±50% of the measured Gmax, with most of them falling in the narrower bands of ±25%. The following expression was obtained when the overconsolidation ratio (OCR) was introduced: G max 815S 0.272 u ( vo' w ) 0.506 OCR 0.181 1183( vo' ) 0.745 OCR 0.360 eo0.613 The trend between measured Gmax and the expression given by equation (11) is illustrated by Figure 14. The graph shows most of the predicted values of Gmax are within ±50% of the measured Gmax. (10) where Gmax, su and s’vo are in kPa, and water content (w) in decimal form. The coefficient of determination is 0.959, and a total of 183 datasets (Gmax, su, s’vo, w and OCR) were used in the multiple regression analyses. As observed by comparing the multiple coefficient of determination between equations (9) and (10), 0.953 versus 0.959, respectively, no improvement or only marginal improvement is attained if stress history, in terms of OCR, is included as independent variable in the multiple regression analysis because the apparent effect of OCR has already been utilized to quantify the values of su, s’vo and w. The expression with the highest correlation derived without involving the undrained shear strength was: G max Figure 13. Comparison of measured and predicted Gmax as a function of undrained shear strength and effective vertical stress normalized with water content. (11) where both Gmax and s’vo are in kPa. The coefficient of determination is 0.933 and a total of 243 datasets were used in the analyses. The above expression is very similar to that obtained by Mayne and Rix (1993). Figure 14. Comparison of measure and predicted in-situ Gmax as a function of effective vertical stress (s’vo), overconsolidation ratio (OCR) and void ratio (eo). Multiple regression analyses were also conducted to provide power function expressions for G max in terms of cone net resistance obtained from piezocone penetration tests, and soil properties. The expressions are presented elsewhere (Taboada et al., 2013). SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C. 8 Field Gmax relationships for Bay of Campeche clay 6 APPLICATION OF EMPIRICAL CORRELATIONS A comparison between measured and predicted Vs profiles using equations (9) and (11) is presented in Figure 15 for one site included in the database of the Bay of Campeche. A good correlation between measured and predicted Vs in clay is observed despite the fact that at the site the clay strata are interbedded with relative thick layers of silty sand. The second relationship for field Gmax in terms of s’vo, stress history (OCR) and (eo) is consistent with the factors known to affect laboratory resonant column tests. 8 ACKNOWLEDGEMENTS The authors gratefully acknowledge the authorization provided by PEMEX to access the data collected in the Bay of Campeche and published the results of this research. Additionally, the support provided by NGI, Inc. to work on this investigation is greatly appreciated by Victor Taboada. 9 REFERENCES measured shear wave velocity profiles for a site with clay strata interbedded with four layers of dense to very dense silty sand. 7 CONCLUSIONS AND RECOMMENDATIONS A database of in-situ Vs measurements with P-S suspension seismic velocity logger and standard geotechnical engineering material properties for the Bay of Campeche clay has been established. The database allowed the development of several empirical correlations between field Gmax and basic soil properties. Two of them are recommended to be used to determine the best estimate field Gmax or Vs of the Bay of Campeche clay when in-situ measurements of Vs at the site are not available. The first equation presents the relationship between Gmax and undrained shear strength (su), effective vertical stress (s’vo) and water content (w). It is recommended for its simplicity, high coefficient of determination of 0.953, and because it could not be improved when introducing the overconsolidation ratio (OCR). Becker, D.E., Crooks, J.H.A., Been, K., and Jefferies, M.G. (1987), "Work as a criterion for determining in situ and yield stresses in clays," Canadian Geotechnical Journal, Vol. 24, pp. 549-564. Casagrande, A. (1936), "Determination of the preconsolidation load and its practical significance," Proceedings, First International Conference on Soil Mechanics and Foundation Engineering, Cambridge, Mass., Vol. 3, pp. 60-64. Hardin, B.O. (1978), “The nature of stress-strain behavior for soils,” Proc. ASCE Geotechnical Engineering Division, Speciality Conference on Earthquake Engineering and Soil Dynamics,” Vol. 1, pp. 3-90. Hardin, B.O. and Black, W. L. (1968), “Vibration modulus of normally consolidated clay,” Journal of the Soil Mechanics Foundation Division, American Society of Civil Engineering, 94 (SM2), pp. 353– 369. Hardin, B. O. and Richart, F. E. Jr. (1963), “Elastic wave velocities in granular soils,” Journal of the Soil Mechanics Foundation Division, American Society of Civil Engineering, 89 (SM1), pp.33–65. Jamiolkowski, M., Leroueil, S., and LoPresti, D.C.F. (1991), "Theme lecture: design parameters from theory to practice," Proceedings, Geo-Coast '91, Yokohama, Japan, pp. 1-41. Mayne, P.W. and Rix, G.J. (1993), "Gmax-qc relationships for clays," Geotechnical Testing Journal, ASTM, Vol. 16, No. 1, pp. 54-60. Taboada V.M., Gan K.C., Cruz D., Barrera P., Espinosa E. and Carrasco D. (2013), “Predictive equations of small-strain shear modulus for Bay of Campeche clay for seismic response analyses,” Proceeding of the Offshore Technology Conference 2013 (in preparation). Vucetic, M.V., and Dobry, R. (1991), “Effect of soil plasticity on cyclic response. Journal of Geotechnical Engineering, Vol. 117, pp. 89–107. SOCIEDAD MEXICANA DE INGENIERÍA GEOTÉCNICA A.C.