47th AIAA Aerospace Sciences Meeting Including The New Horizons Forum and Aerospace Exposition 5 - 8 January 2009, Orlando, Florida AIAA 2009-501 Fuel Composition and Coking Analysis of Endothermically Heated Hydrocarbon Fuels for Use in a Pulsed Detonation Engine Chris A. Stevens * and Paul I. King † Air Force Institute of Technology, Wright Patterson AFB, OH, 45433 Eric A. Nag ley ‡ Helicopter Anti-Submarine Squadron, Naval Air Station Jacksonville, FL 32212 and Fred R. Schauer § Air Force Research Laboratory, Propulsion Directorate Wright-Patterson AFB, OH 45433 Three key characteristics of JP-8 and JP-7 that have been heated endothermically are compared. Samples of both fuels were collected after being heated by the waste process heat of a Pulsed Detonation Engine (PDE). It was initially expected that JP-7 woul d have better heat absorption and carbon deposit formati on properties, and that JP-8 woul d have lower ignition ti mes due to i ts lower thermal stability. Samples of reacted fuel were collected from the endothermic cooling system of a PDE, and each of the three metrics was calculated. As expected JP-7 proved to have better heat abs orption, and lower presence of the precursors to carbon deposit; however, it was also discovered that JP-7 produced more of the gaseous products that reduce ignition ti mes in the engine. With this result, JP-7 outperforms JP-8 in all three metrics. Nomenclature a, b, c, d, e Cp h !"q T # $ = specific heat polynomial fit coefficients = specific heat (J/kg-K) = enthalpy (J/kg) = net heat addition = temperature at (K) = detonation cell size = mass fraction Subscripts in out = measured at heat exchanger entrance = measured at heat exchanger exit I. Introduction With the advent of pulsed detonation engines (PDE’s) and Supersonic Co mbusting ramjets (SCRAMjets) it has become important to understand the endothermic cooling properties of hydrocarbon fuels used to power these types of engines1 . JP-8 is the standard jet fuel used in all United States Air Force aircraft2 . Work has already been * Graduate Student Department of Aeronautics and Astronautics, 2950 Hobson Way, AIAA Student Member Professor, Depart ment of Aeronautics and Astronautics, 2950 Hobson Way, Senior AIAA Member ‡ LT, USN, Graduate Student, Depart ment of Aeronautics and Astronautics, 2950 Hobson Way, AIAA Member. § Senior Eng ineer, Head PDRF, AFRL/ RZTC, 1950 5th Street, Sen ior AIAA Member. † 1 American Institute of Aeronautics and Astronautics 092407 This material is declared a work of the U.S. Government and is not subject to copyright protection in the United States. completed showing that JP-8 can be effect ively used as an endothermic cooling fuel for PDE’s 3 . This paper presents a comparison between the performance of JP-8 and JP-7. The metrics used for comparison are ignition time, heat addition, and quantity of poly-aromatic co mpounds in the fuel. II. Background A PDE operates on the principle of fixed volu me detonation of fuel air mixtu re. Figure 1 outlines the PDE operating cycle for the research engine. 33 ms 33 ms 33 ms Fill Fire Purge 4 ms 15 ms 2.5 ms 1.5 ms Ignition Ignition Delay 10.3 ms Blow Down Transition Detonation Figure 1 – Research PDE operating cycle at 10 Hz pulse rate In the fill phase, a mixture of fuel and air is forced into an open ended tube. Then during the firing phase, the mixtu re is detonated. The resulting shockwave and hot, high pressure gas exits the open end of the tube creating thrust. In the purge phase, after the thrust tube blows down to ambient pressure, the spent mixture is purged with a blast of pure air before being refilled with another charge of fuel/air mixture. Because the detonation travels through a stagnant fuel mixture, much of the heat generated during detonation is transferred into the walls of the thrust tube. The purge air helps to carry away some of the heat, but it must be supplied by the engine and increases the span between thrust pulses. An alternative to purging the thrust tubes is to cool the walls of the thrust tubes using fuel. The process of detonation in a PDE has three stages4 . The times shown are representative of the experimental setup. The first stage is ignition of the mixture. The second stage is transition from deflagrat ion to detonation. The third is progression of the detonation. Ignition is long on the order of 15 ms. Transition is shorter at about 1.5 ms, and progression is very short, fro m one to two milliseconds. The obvious target for reducing the length of the fire phase is the ignition time. The length of the fire phase of the PDE depends on the ignition time of the fuel, the length of time and tube required to transition to detonation, and the blow down time5 . The ignit ion time and transition length are related to the chemical co mposition of the fuel used, and shorter times result in higher cycle frequencies for the engine and higher thrust. Elimination of the purge phase and reduction of the length of the fire phase both lead to increases in the performance of the PDE6 . Because the detonation process produces large amounts of heat, a regenerative cooling system is useful for two purposes. The first is cooling of the thrust tubes. For this purpose it is best to use a fuel with high heat capacity, resulting in smaller heat exchangers. The second is the addition of energy to the fuel prior to detonation7 . Liquid hydrocarbon fuels are formulated to have high energy densities; however, their comp lex mo lecular structures make them difficult to detonate8 . Figure 2 shows some representative fuels and their detonation initiation energies. 2 American Institute of Aeronautics and Astronautics 092407 Figure 2. Initiation energies of various fuels1 Liquid hydrocarbon fuels have initiation energies close to that of propane(cite), and fro m Fig. 2 that init iation energy is ~2x105 J. Hydrogen and acetylene have much lower in itiation energies, 5x103 J and 200 J respectively. A lo w initiat ion energy is desirable because the length of the spirals used to transition to detonation is reduced10 . However, both of these fuels are impract ical for use in aircraft fo r many reasons including fire safety, fuel tank weight, lo w energy density, and production costs 2 . When used as an endothermic coolant, a liquid hydrocarbon fuel such as JP-8 can be decomposed into products including hydrogen and acetylene7 . The presence of these species in the fuel mixtu re should reduce the initiation energy allowing a reduction in spiral length. The downside of endothermic cooling is the production of solid carbon particles in the fuel10 . These particles adhere to passage walls, and clog both filters and nozzles. In order to design an effective endothermic cooing system, the format ion of these particles needs to be kept to a minimu m. The characteristics of JP-7 and JP-8 make these two fuels ideal for the study of endothermic cooling. JP-8 is currently the standard jet fuel for the USAF2 . It has a high energy density and is distilled direct ly fro m crude oil. It is available in large quantities, and is the standard to which new fuels are co mpared. Developing the capability to operate a PDE on JP-8 precludes the need to operate the engine on more expensive, less available fuels. JP-7, however, was designed for applications where high thermal stability and low aromatic concentrations are important. Thermal stability is the resistance to chemical change brought on by heating. Originally developed for the SR-71 Blackbird 2 , it is blended from special stocks to ensure that it has a very low concentration of aromatics <3% and high thermal stability 2 . The low aro mat ic properties of JP-7 make it attractive, as aro matic co mpounds are precursors to coke format ion in endothermic reactions, and the high thermal stability makes it mo re effective as an endothermic coolant than JP-8. Knowing the behavior of these two fuels is important to deciding how specialized a fuel needs to be in order to be viable as an endothermic coolant for PDE’s. 3 American Institute of Aeronautics and Astronautics 092407 III. Experimental Setup In each experiment, fuel was heated above its known endothermic temperature and a samp le of the reacted fuel was collected before it passed on to the engine for detonation. The test rig used is identical to that used by Nagley 11 . A schematic of the test rig is shown in Fig. 3. Figure 3 - S chematic of test rig Neat fuel flowed fro m the pressurized reservoir to the shell passage of the thrust tube heat exchangers (Fig. 4) where it was heated and the endothermic react ion took place. After heating, the fuel was split into two streams. The proportion of the flow through each stream was governed by the flow nu mber of nozzles in the lines. Most of the fuel was then injected into the airstream feed ing the PDE. The smaller portion of the fuel was then cooled to room temperature in a cold water bath (Fig. 5). Cooling the fuel both halted the endothermic reaction, and protected the hardware down stream fro m damage. A remotely operated valve then directed the cool fuel either to the intake man ifold or to the liquid collection flask (Fig. 6). Any gas in the cooled fuel then passed through the gas sample cylinder (Fig. 7) and into the gas fuel collect ion bag (Fig. 8). The bag is tubular and unrolled fro m a spool increasing in volu me as more gas accumulated. The volu me o f gas collected was calcu lated by measuring the length that the bag unrolled. 4 American Institute of Aeronautics and Astronautics 092407 Figure 4 - Thrust tube heat exchangers Figure 5 - Cold water bath Figure 6 - Liquid sample flask 5 American Institute of Aeronautics and Astronautics 092407 Figure 7 - Gas sample cylinder Figure 8 - Gas collection bag At the beginning of a test the engine was operated with the sample collect ion valve in the bypass position until the fuel flow reached the desired temperature. Then the valve was activated drawing off a sample for co llect ion. After a few minutes of operation, the valve was returned to bypass and the engine was shut down. The gas sample cylinder was isolated to contain a portion of the gas sample for chemical analysis and then was replaced. Measurements of the liquid sample volu me and the length of the unrolled gas bag were taken. Then up to 120 mL of the liquid was bottled, and stored with the gas collection cylinder for chemical analysis. The endothermic cooling capacity of a fuel is best measured by the heat absorbed by the fuel as it is heated. This net heat addition is a bulk quantity that applies regardless of the level of reaction the fuel undergoes. Another applicable figure of merit is the liquid-to-gas conversion ratio, the ratio of mass that was converted to gaseous species by the endothermic reaction to the mass of neat fuel heated. The liquid-to-gas conversion ration was determined fro m measurements of the volume of gas collected, volume of liquid collected, the gas composition, the liquid density, and the flow rate of fuel into the heat exchangers. Nagley explains the process in full in h is thesis3 . The net heat addition (!q) was calculated by modeling JP-7 and JP-8 as a collection of 117 chemical species believed to be present in either the raw or reacted fuel. Part icular selections included long chain paraffins and olefins for raw fuel, light hydrocarbons for gaseous products, hydrogen, and poly-aromatic species involved in solid carbon formation. The heat addition was calculated using a control volume approach taking the pair of thrust tube heat exchangers as the control volu me and calculat ing the enthalpy of the fuel at the entrance and exit (Eq. 1). (1) The enthalpy at each station was calculated by assuming negligib le changes in potential and kinetic energy leaving only the effects of temperature fo r each chemical species (Eq. 2). (2) The mass fraction of each species ($) was measured by gas chromatograph (GC), gas chro matograph mass spectrometer (GCMS), or high precision liquid chro matography (HPLC) depending on the phase, and molecu lar weight. Gas samples were analyzed using GC, poly-aro mat ic co mpounds using HPLC and all others using GCM S. The mass fractions were then normalized and mult iplied by specific heat (Cp ) and temperature (T) resulting in the modeled enthalpy of the fluid. Specific heat functions for each species were calcu lated by fitting fourth order polynomials (Eq 3) to National Institute for Standards and Technology (NIST) and Thermodynamic Research Center (TRC) data tables. The fits were calculated using at least nine points between 273 K and 1000 K for a least squares fit of the data. (3) 6 American Institute of Aeronautics and Astronautics 092407 The resulting calculated heat addition is Joules per kg of fuel and is not dependant on fuel flow rate, heat exchanger geometry, or the metering of fuel between the intake and sampling rig. For both JP-8, and JP-7 there was some difficulty in startup of the engine as ignition time of the fuel was so long that it did not transition to detonation before the end of the thrust tubes. While the fuel did not detonate it did burn and began to heat the fuel in the thrust tube heat exchangers. The problem corrected itself as the temperature of the fuel rose into the endothermic range. Ignition times were calculated by measuring the time between the spark discharge at the head of the tube, and the beginning of pressure rise within the tube. Figure 4 shows a representative set of spark and pressure traces. The ignition time is the time between the drop of the spark trace when the energy discharges, and the time when the slope of the pressure increases beyond a preset threshold. Figure 9 - Representative spark and pressure traces Formation of deposits on the walls of fuel system is a major drawback of endothermic cooling systems. In an endothermic reaction such as the one that occurs in fuel, solid carbon forms via the assembly of poly aro matic hydrocarbons into graphite sheets. It is believed that by limit ing the presence of aromatic co mpounds in the raw fuel, solid carbon formation will be limited to that formed by pyrolysis alone. The concentrations of several polyaromat ic co mpounds were measured by High Performance Liqu id Chro matography (HPLC). The poly-aro matics are intermediate species in the solid carbon format ion process. Direct comparison of the concentrations in JP-8 and JP-7 samples with similar heat additions will indicate whether or not the low concentration of aromatic compounds in the JP-7 reduced the amount of solid carbon formed. Together, heat addition, ignition time, and poly-aro mat ic co mpound concentration provide a set of measurements that outline the benefits and drawbacks of JP-8 and JP-7 as working fluids in an endothermic cooling system for PDE’s. Heat addition demonstrates the fuels ability to cool much like heat capacity for a non-reacting fluid. Ignition time supplies a measure of the improvement in the fuel’s detonation properties, and poly-aromat ic concentration reveals the relative tendency of the fuel to form solid carbon particles. Fro m this knowledge, fuels can be selected for endothermic cooling systems that provide high heat absorption, short ignition times, and low tendency to form solid carbon particles during endothermic heating. 7 American Institute of Aeronautics and Astronautics 092407 IV. Results and Discussion Due to its high thermal stability, JP-7 is a better heat sink than JP-8. Figure 3 co mpares the heat addition for JP-7 and JP-8 as a function of the temperature of the fuel leaving the thrust tube heat exchangers. JP-7 consistently absorbed more heat than JP-8 at the same exit temperatures. Because of the limited volume of the pressurized fuel reservoir, the maximu m attainable temperature was limited to about 845 K for JP-7 and 940 K for JP-8. The resulting data set is more co mpact than desired, but a linear least squares fit of the data still shows higher heat additions for JP-7. 3000 Heat Addition (J/g) 2500 2000 JP-7 1500 JP-8 %&'()*"+,--7) 1000 %&'()*"+,--8) 500 0 800 820 840 860 880 900 920 940 960 Heat Exchanger Exit Teperature (K) Figure 10 - Heat sink capability While the heat exchanger exit temperature is easily measured, it fails to capture the effects of residence time. The mass flow rate of fuel through the heat exchanger is not constant due to the changing restrictions caused by coking of nozzles and filters. This becomes apparent when liquid-to-gas conversion is plotted as a function of heat exchanger exit temperature, and separately as a function of heat addition. Figures 6 and 7 show that the heat addition is a better independent variable for co mparisons of endothermic heating properties. Heat addition does not assume a constant fuel flow rate like analyses based on temperature alone. 8 American Institute of Aeronautics and Astronautics 092407 0.35 Liquid to Gas Conversion (kg gas/ kg fuel) 0.3 0.25 0.2 0.15 0.1 0.05 0 810 815 820 825 830 835 840 845 850 Heat Exchanger Exit Temp. (K) Figure 11 - Temperature as the independent variable (JP-7) 0.35 Liquid to Gas Conversion (kg gas/kg fuel) 0.3 0.25 0.2 0.15 0.1 0.05 0 800 1000 1200 1400 1600 Heat Addition (J/g) Figure 12 - Heat addition as the independent variable (JP-7) It is now appropriate to examine the liquid-to-gas conversion used by Nagley3 . By plotting his JP-8 conversion fractions against heat addition and superimposing the results from JP-7. It is clear that JP-7 produces more gaseous products than JP-8 (Fig 8). In fact, the more energy added to the fuel the greater the increase. Th is should lead to improved ignition times for JP-7 also due to the larger quantities of hydrogen, ethylene and ethane. The greater gas production was unexpected since JP-7 is formu lated for high thermal stability. A high thermal stability should mean a reduced level of reaction and low gas concentrations. This result strengthens the case for high thermal stability fuels. 9 American Institute of Aeronautics and Astronautics 092407 0.5 L to G conversion (kg gas/kg fuel) 0.45 0.4 0.35 0.3 0.25 JP-7 0.2 JP-8 0.15 0.1 0.05 0 0 500 1000 1500 2000 2500 3000 Heat Addition (J/g) Figure 13 - Gas production The larger quantities of gas produced by the JP-7 should reduce the ignition time of the fuel. Figure 9 shows that to be the case. Again high thermal stability is superior with lower ignition times by two to three milliseconds. Two milliseconds is an imp rovement of appro ximately 25% in the ignit ion times. At this point, two of the three metrics point toward high thermal stability as the better choice for fuels. Stable fuels have both better heat sink capacity and lower ignit ion times. 9 Ignition Time (ms) 8 7 6 JP-7 JP-8 5 4 3 800 820 840 860 880 900 920 Heat Exchanger Exit Temperature (K) Figure 14 - Ign ition time 10 American Institute of Aeronautics and Astronautics 092407 940 960 It was expected that the low levels of aromatics in JP-7 would reduce the amount of carbon formed by heating. Figure 10 shows that the total concentrations of poly-aro matics were consistently higher for JP-8. The reduced quantities of these precursor compounds resulted in fewer carbon particles forming in JP-7. The reduction was visually evident by the amount of carbon that collected in the fuel filter after each run. Careful formulat ion of fuel to reduce the presence of aromatic co mpounds in the raw fuel resulted in reduced format ion of solid carbon deposits. 0.06 0.05 Mass Fraction 0.04 0.03 JP-7 JP-8 0.02 0.01 0 0 500 1000 1500 2000 2500 3000 Heat Addition (J/g) Figure 15 - Poly-aromatic total mass fractions V. Conclusion At the beginning of testing, it was believed that there were qualit ies of both JP-7 and JP-8 that would give each fuel an advantage in some aspects of the endothermic heating. Due to high heat capacity, and thermal stability JP-7 would be a better heat sink, and that the low aro mat ic content would reduce carbon formation. However, the reduced levels of reaction would result in higher ignit ion times than JP-8. It became clear during the data analysis that JP-7 also produced more gaseous products than JP-8 including those that reduced ignition times. This unexpected result shows that JP-7 is an all around better fuel for an endothermic cooling system. In general, it means better cooling and reduced clogging. As applied to PDE’s it also means higher pulse frequency because of reduced ignition times. It is reco mmended that fuels selected for endothermic cooling systems in general, and especially those used with PDE’s, be supplied with fuels that have the high thermal stability, and low aro mat ic content. 11 American Institute of Aeronautics and Astronautics 092407 Acknowledgme nts Appreciation is expressed to the tech team at the PDE Research Facility who made this research possible especially to Dr. John Hoke and Curt Rice. In addition, Dr. Matt DeWitt and Linda Schafer (AFRL/ RZTG) were helpful in developing the fuel systems and conducting fuel analysis for experiments performed. Appreciation is also expressed to Dr. Robert Hancock (AFRL/RZTC) for his technical leadership. Funding was provided by the Air Force Research Laboratory, Propulsion Directorate and AFOSR. References 1. Huang, H., Sobel, D., and Spadaccini, L., “Endothermic Heat-Sink of Hydrocarbon Fuels for Scramjet Cooling,” AIAA 2002-3871, 38th AIAA/AMSE/SAE/ASEE Joint Propulsion Conference and Exhibit, Indianapolis IN, 7-10 July 2002. 2. Edwards, T., “Liquid Fuels and Propellants for Aerospace Propulsion: 1903-2003,” Journal of Propulsion and Power, Vol.19, No. 6, 2003, pp. 1098-1100 3. Nagley, E. A., “Fuel Compo sition Analysis of Endothermically Heated JP-8 F uel for Use in a Pulsed Detonation Engine,” M.S. Thesis Department of Aeronautics and Astronautics, Air force Institute of Technology, W right-Patterson Air Force Base, OH 2008 4. Schauer, F. R., Miser, C. L., Tucker, K. C., Bradley, R. P., and Hoke, J. L., “Detonation Initiation Hydrocarbon-Air Mixtures in a Pulsed Detonation Engine,” AIAA 2005-1343, 43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno NV, 10-13 January 2005. 5. Helfrich, T. M., King, P. I., Hoke, J. L., and Schauer F. R., “Effect of Supercritical Fuel Injection on the Cycle Performance of a Pulsed Detonation Engine,” AIAA-2006-5133, 42nd AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Sacramento CA, 9-12 July 2006. 6. Helfrich, T. M., “Cycle Performance of a Pulse Detonation Engine with Supercritical Fuel Injection,” M.S. Thesis, Department of Aeronautics and Astronautics, Air Force Institute of Technology, Wright-Patterson AFB OH, 2006. 7. Helfrich, T.M. Schauer, F. R., Bradley, R. P., and Hoke, J. L., “Evaluation of Catalytic and Thermal Cracking in a JP-8 Fueled Pulsed Detonation Engine,” AIAA 2007-235, 45th AIAA Aerospace Sciences Meeting and Exhibit, Reno NV, 8-11 January 2007. 8. Edwards, T., “’Kerosene’ Fuels for Aerospace Propulsion – Composition and Properties,” AIAA 2002-3874, 38th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Indianapolis IN, 7-10 July 2002. 9. Tucker, K. C., King, P. I., Bradley, R. P., and Schauer, F. R., “The Use of a Flash Vaporization System with Liquid Hydrocarbon Fuels in a Pulse Detonation Engine,” AIAA 2004-0868, 42nd AIAA Aerospace Sciences Meeting and Exhibit, Reno NV, 5-8 January 2004. 12 American Institute of Aeronautics and Astronautics 092407 10. Edwards, T., DeWitt, M.J., Shafer, L., Brooks, D., Huang, H., Bagley, S.P., Ona, J.O., and Wornat, M.J., “Fuel Composition Influence on Deposition in Endothermic Fuels,” AIAA 2006-7973, 14th AIAA/AHI Space Planes and Hypersonic Systems and Technologies Conference, Canberra, Australia, 6-9 November 2006. 11. Nagley, E., King, P., Schauer, F., DeWitt, M., and Hoke, J., “Fuel Composition Analysis of Endothermically Heated JP-8 Fuel for Use in a Pulsed Detonation Engine,” AIAA 2008-109, 46th AIAA Aerospace Sciences Meeting and Exhibit, Reno, Nevada, 7-10 January 2008. 13 American Institute of Aeronautics and Astronautics 092407