AFRL-PR-WP-TP-2002-201
`
`ON-LINE FUEL DEOXYGENATION
`FOR COKE SUPPRESSION
`ASME GT-2002-30071
`
`Louis J. Spadaccini and He Huang
`
`JUNE 2002
`
`Approved for public release; distribution is unlimited. I
`
`Copyright © 2002 by ASME
`
`This work is copyrighted. The United States has for itself and others acting on its behalf an
`unlimited, paid-up, nonexclusive, irrevocable worldwide license. Any other form of use is
`subject to copyright restrictions.
`
`PROPULSION DIRECTORATE
`AIR FORCE RESEARCH LABORATORY
`AIR FORCE MATERIEL COMMAND
`WRIGHT-PATTERSON AIR FORCE BASE, OH 45433-7251
`
`20020926 088
`
`GE-1012.001
`
`

`
`REPORT DOCUMENTATION PAGE
`
`Form Approved
`OMB No. 0704-0188
`
`The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, searching existing data sources,
`gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including
`suggestions for reducing this burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204,
`Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not
`display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.
`2. REPORT TYPE
`1. REPORT DATE (DD-MM-YY)
`Conference Paper
`June 2002
`4. TITLE AND SUBTITLE
`ON-LINE FUEL DEOXYGENATION FOR COKE SUPPRESSION
`ASMEGT-2002-30071
`
`3. DATES COVERED (From - To)
`
`6. AUTHOR(S)
`Louis J. Spadaccini and He Huang
`
`7. PERFORMING ORGANIZATION NAME(S) AND ADDRESSfES)
`United Technologies Research Center
`411 Silver Lane
`East Hartford, CT 06108
`
`SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
`Propulsion Directorate
`Air Force Research Laboratory
`Air Force Materiel Command
`Wright-Patterson Air Force Base, OH 45433-7251
`
`12. DISTRIBUTION/AVAILABILITY STATEMENT
`Approved for public release; distribution is unlimited.
`
`5a. CONTRACT NUMBER
`F33615-97-D-2784
`5b. GRANT NUMBER
`5c. PROGRAM ELEMENT NUMBER
`62203F
`5d. PROJECT NUMBER
`3048
`5e. TASK NUMBER
`05
`5f. WORK UNIT NUMBER
`EW
`8. PERFORMING ORGANIZATION
`REPORT NUMBER
`
`10. SPONSORING/MONITORING
`AGENCY ACRONYM(S)
`AFRL/PRTG
`11. SPONSORING/MONITORING
`AGENCY REPORT NUMBER(S)
`AFRL-PR-WP-TP-2002-201
`
`13. SUPPLEMENTARY NOTES
`This report contains copyrighted material.
`Proceedings of TURBO EXPO 2002; ASME Turbo Expo: Land, Sea & Air 2002; June 3-6,2002, Amsterdam, The Netherlands
`
`14. ABSTRACT (Maximum 200 Words)
`Fuel deoxygenation is being developed as a means for suppressing autoxidative coke formation in aircraft fuel systems, thereby
`increasing the exploitable cooling capacity of the fuel, enabling major increases in engine operating temperature and cycle
`efficiency. Reduced maintenance is an added benefit. A prototype membrane filter module for on-line removal of dissolved
`oxygen, which would otherwise react to form coke precursors, was constructed and successfully demonstrated. The fuel flows over
`the membrane, while oxygen diffuses through it at a rate that is proportional to the difference in oxygen partial pressures across the
`surface. Tests were conducted over a range of fuel flow rates (residence times) and temperatures. The filter was operated with air-
`saturated jet fuel for several hours at a steady-state condition, verifying the capability to remove essentially all of the dissolved
`oxygen (to <1 ppm) and proving the viability of the concept.
`
`15. SUBJECT TERMS
`thermal stability, coking, deoxygenation
`
`16. SECURITY CLASSIFICATION OF:
`
`a. REPORT
`Unclassified
`
`b. ABSTRACT
`Unclassified
`
`c. THIS PAGE
`Unclassified
`
`17. LIMITATION
`OF ABSTRACT:
`SAR
`
`18. NUMBER OF
`PAGES
`14
`
`19a. NAME OF RESPONSIBLE PERSON (Monitor)
`Tim Edwards
`19b. TELEPHONE NUMBER (Include Area Code)
`(937) 255-3524
`
`Standard Form 298 (Rev. 8-98)
`Prescribed by ANSI Std. Z39-18
`
`GE-1012.002
`
`

`
`Proceedings of TURBO EXPO 2002
`ASME Turbo Expo: Land, Sea & Air 2002
`June 3-6, 2002, Amsterdam, The Netherlands
`
`GT-2002-30071
`
`ON-LINE FUEL DEOXYGENATION FOR COKE SUPPRESSION
`
`Louis J. Spadaccini and He Huang
`United Technologies Research Center
`East Hartford, CT 06108
`
`ABSTRACT
`Fuel deoxygenation is being developed as a means for
`suppressing autoxidative coke formation in aircraft fuel sys-
`tems, thereby increasing the exploitable cooling capacity of the
`fuel, enabling major increases in engine operating temperature
`and cycle efficiency. Reduced maintenance is an added benefit.
`A prototype membrane filter module for on-line removal of
`dissolved oxygen, which would otherwise react to form coke
`precursors, was constructed and successfully demonstrated. The
`fuel flows over the membrane, while oxygen diffuses through it
`at a rate that is proportional to the difference in oxygen partial
`pressures across the surface. Tests were conducted over a range
`of fuel flow rates (residence times) and temperatures. The filter
`was operated with air-saturated jet fuel for several hours at a
`steady-state condition, verifying the capability to remove
`essentially all of the dissolved oxygen (to <1 ppm) and proving
`the viability of the concept.
`A convincing demonstration of coke suppression was
`performed when air-saturated (normal) and deoxygenated jet
`fuels were tested in a standard ASTM heated tube apparatus at
`wall temperatures as high as 850 F. With deoxygenated fuel,
`there was a dramatic reduction (more than an order of
`magnitude) in coke deposition relative to air-saturated Jet A,
`which will allow the maximum fuel temperature to be increased
`by more than 200 F, doubling the available heat sink. Moreover,
`deoxygenated Jet A was shown to perform as well as JP-7, the
`Air Force's highest thermal stability fuel. An analytical model
`foro xygen permeation through the membrane was formulated,
`and used in conjunction with the test data to estimate the filter
`size required for a practical (i.e., low-volume/high-flowrate)
`deoxygenator.
`
`INTRODUCTION
`Thermal management is a design driver for advanced gas
`turbine engines. Increasing cycle temperature increases engine
`performance and cycle efficiency. However, improved
`
`materials for enabling high-temperature operation are not
`available. Fuel is currently used as a heat sink for the engine oil
`system, but its cooling capacity is limited by a temperature
`constraint necessary to preclude the formation of coke deposits.
`Therefore, it is essential to develop a method to suppress coke
`formation and, thereby, significantly increase the available heat
`sink and permit extended utilization of the fuel for component
`and system cooling (e.g., turbine, compressor, and avionics) in
`advanced engines and sustained flight at high Mach number.
`Over the past 40 years there have been many attempts by
`researchers to suppress coking [1,2], but, with the exception of
`JP-8+100 additives [3] that allow operation up to 425 F, they
`have proven to be unsuccessful or impractical.
`The major factor contributing to coke deposition at
`temperatures up to approximately 700 F is oxygen that
`dissolves into the fuel when it comes in contact with air [4].
`When air-saturated fuel is heated above approximately 300 F,
`the dissolved oxygen reacts to form free-radical species (coke
`precursors) that initiate and propagate other autoxidation
`reactions, leading to deposit formation. These reaction paths
`become insignificant when the concentration of the dissolved
`oxygen is reduced from its ambient saturation level of 70 ppm
`to approximately 1 ppm (i.e., deoxygenated) [1]. At
`temperatures above approximately 900 F, the deposition
`mechanism is characterized by pyrolysis, wherein chemical
`bonds are broken and large alkanes are converted into smaller
`alkanes, alkenes, and some hydrogen [5]. Since pyrolysis
`requires much higher temperatures than autoxidation, reducing
`the oxygen content of fuel will allow it to be heated
`significantly before thermal decomposition begins, increasing
`the cooling capacity. Therefore, fuel deoxygenation should
`dramatically reduce autoxidative coke deposition, and make it
`possible to realize the thermal stability goals for JP-8+225/JP-
`900 [6],
`This study deals with the development of a small, practical
`membrane-based filter for on-line removal of dissolved oxygen
`
`Copyright © 2002 by ASME
`
`GE-1012.003
`
`

`
`Membrane Porous Support
`
`Jet Fuel
`
`7~> O
`u2 vacuum or N2 2 4
`
`Figure 1: Supported Oxygen -Permeable Polymer Membrane
`
`that would otherwise react and form the precursors for
`autoxidative coke. A permeable membrane is a selective barrier
`that permits the separation of certain species in a fluid by
`diffusion or sorption-diffusion mechanisms. In general, the
`membrane structure consists of an ultra-thin coating that has the
`requisite separation properties, and a micro-porous polymer
`support that provides strength. In the present representation (see
`Figure 1), the oxygen molecules in jet fuel dissolve into the
`membrane and then diffuse across it, driven by the difference of
`oxygen partial pressure (chemical potential or driving force),
`while the hydrocarbon molecules are unaffected and pass over
`it. Previous attempts at deoxygenating fuel have included
`sparging with nitrogen, and the use of molecular-sieve
`adsorbents and chemical reducing agents. These approaches
`have proven impractical for aircraft applications because they
`are costly, heavy and bulky, or even dangerous [7]. Some
`aircraft use an on-board inert gas generator system (OBIGGS)
`to reduce the oxygen concentration in the fuel tank below the
`flammability limit (~9 vol. %). However, it is unlikely that a
`similar system could be used to lower the oxygen concentration
`dissolved in the fuel to approximately 1 ppm.
`
`EXPERIMENTAL APPROACH
`An experimental apparatus for measuring the oxygen
`removal rate from jet fuel by membrane permeation has been
`developed and is illustrated schematically in Figure 2. The rig
`consists of a fuel aeration system, a fuel pump, a preheater, a
`membrane module, and an on-line oxygen sensor. Although the
`membrane is envisioned as being very thin to facilitate gas
`diffusion (e.g., micron thickness), for convenience, concept
`feasibility experiments were conducted using several closely
`sized commercially available thin-wall tubes made of the
`membrane material.
`
`In a fuel deoxygenation test, air-saturated jet fuel was
`metered through the membrane module at a predetermined and
`controllable pressure, temperature, and flowrate. After leaving
`the module, the fuel was cooled to ambient temperature (to
`eliminate measurement uncertainties) and the oxygen
`concentration determined on-line using a polarographic-type
`oxygen sensor. The membrane module consisted of a polymeric
`membrane tube (through which the fuel flowed) installed in a
`cylindrical shell that was either evacuated or purged with
`nitrogen, to provide the driving force (oxygen partial pressure
`difference) for oxygen removal. Tests were conducted in a
`single-pass flow arrangement over a range of fuel flow rates
`(residence times) and temperatures. The maximum fuel
`temperature was limited to 240 F to preclude oxygen depletion
`due to thermal reactions. This behavior was verified by
`substituting a stainless-steel tube for the membrane rube and
`demonstrating that at 240 F the oxygen concentration at the
`outlet was unchanged from the inlet (i.e., 100 % saturated). To
`verify the oxygen sensor accuracy/repeatability after each test,
`the fuel was re-aerated to demonstrate measurement of 100
`percent saturation.
`An initial series of tests was conducted to measure the
`permeability of a 0.036-in.- OD x 0.0015-in.-wall membrane
`tube to gaseous oxygen. The membrane is essentially
`impermeable to liquids and organic vapors, and the gas
`permeability increases as the size of the molecule decreases.
`The oxygen permeability of the membrane was determined by
`pressurizing the tube with oxygen, and measuring the amount
`of gas that diffused through the tube wall into the outer shell.
`Oxygen permeation rates were determined using bubble-type
`flowmeters for measuring very low flow rates. The pressure and
`temperature of the oxygen feed were varied, and the outer shell
`was maintained at atmospheric pressure. The feed and permeate
`sides of the module were purged with oxygen prior to each
`permeation test. The results are presented in Figure 3, and show
`that the oxygen permeability of the polymer membrane tube
`increases linearly with the oxygen pressure difference across it,
`and also increases significantly with increasing temperature.
`
`metering
`pump
`
`N2 or vacuum TC
`
`cooler
`
`sensor mass
`flow meter
`IIIU
`
`.-I
`
`Membrane Tube
`(0.036-in.OD x 0.0015-in.wall x 36-in.. typ.)
`
`aerator
`
`Figure 2: Membrane Test Apparatus
`
`20 30 40 50
`
`Oxygen pressure difference, psi
`
`Figure 3: Influence of Oxygen Temperature on Membrane
`Permeability
`
`Copyright © 2002 by ASME
`
`GE-1012.004
`
`

`
`0.036-in.-OD Memb
`T = 70F
`
`ane Tube
`
`E a -
`2 2
`
`^-^~~*"
`^^^-f—^
`
`X
`
`/ 0.001-in. wall
`
`/
`
`0.0015-in. wall
`
`——• 0.003-in. wall
`
`0 10 20 30 40 50 60 70
`
`Oxygen pressure difference, psi
`
`Figure 4: Influence of Membrane Thickness on Oxygen
`Permeability
`
`Another important factor that will influence the oxygen re-
`moval rate is the membrane thickness. Additional oxygen-gas
`permeability tests were conducted to evaluate the effect of
`membrane thickness using membrane tubing with wall
`thickness of 0.001-in. and 0.003-in. The results, shown in
`Figure 4, for ambient temperature (70 F) indicate that the
`oxygen gas diffusion through the membrane increases rapidly
`(nearly exponentially) with decreasing thickness, and
`emphasizes the desirability of the membrane being relatively
`thin.
`The influence of fuel temperature on the rate of oxygen
`removal is shown in Figure 5, for a 0.003-in.-thick membrane
`tube. At ambient temperature, the oxygen concentration in air-
`saturated JP-8 fuel was reduced by 70 percent (to
`approximately 20 ppm) in a single pass through the 3-ft-long,
`0.040-in.-OD polymeric membrane tube. At a fuel temperature
`of 240 F (a typical aircraft main fuel pump outlet temperature),
`more than 90 percent of the dissolved oxygen was removed for
`the same residence time (volume flowrate). That is, the oxygen
`100
`V
`90
`
`JP-8 fuel
`
`V--
`
`concentration was reduced from approximately 70 ppm to a
`level below 5 ppm in a residence time of 25 seconds. This rate
`of deoxygenation corresponds to a space velocity of 145/hr,
`where
`
`volume flowrate 7 ...
`reactor volume residence time' *■ '
`In an aircraft fuel system, higher space velocities will be
`required to minimize the filter volume. Reducing the membrane
`thickness and increasing the surface area will enhance oxygen
`diffusion and facilitate operation at much higher space velocity.
`Fuel heating enhances performance by increasing the
`permeability of the membrane, and by decreasing the solubility
`of oxygen in fuel, thereby increasing the driving force across
`the membrane. Increasing the pressure of the air-saturated fuel
`will not produce a significant benefit because it will not change
`the oxygen partial pressure (which is established in the tank) or
`significantly increase the permeability of the membrane.
`Diffusion of the oxygen molecules dissolved in liquid fuel
`to the membrane surface can be the rate-determining or rate-
`limiting step in the fuel deoxygenation process, depending on
`the oxygen diffusion rate through the membrane. When an
`effective and thin membrane is used, the overall deoxygenation
`rate may be controlled by the bulk diffusion of the oxygen
`molecules in fuel. This effect can be quantitatively described by
`the oxygen mass transfer coefficient,
` overall oxygen removal rate
`■'MT ~ oxygen diffusion rate through membrane
`where the oxygen diffusion through the membrane is
`determined from the gas permeability tests discussed above.
`Mass transfer coefficients for flow in a 0.034-in.-ID x 0.003-
`in.-wall membrane tube are shown in Figure 6, as a function of
`the fuel flowrate. The results indicate that the mass transfer
`coefficient increases as the fuel flowrate increases from 50 g/hr
`(Re = 16) to about 800 g/hr (Re = 260). Although the flow is
`still laminar at 800 g/hr, the bulk diffusion rate of the oxygen
`molecules in fuel is much faster than the rate through the
`membrane and, therefore, the mass transfer coefficient
`approaches 1.0, indicating that there was no limitation caused
`by mass transfer. Mass transfer in the fuel did limit the overall
`
`(2)
`
`80
`
`70
`
`60
`
`50
`
`40
`
`' 30
`
`20
`
`10
`
`-J0F
`
`240 F
`
`10 ppm
`
`Membrane Tube
`0.034-ln. ID x 0.003-in. wall
`
`10 15 20
`Residence time, s
`Figure 5: Influence of Fuel Temperature on Deoxygenation
`
`25 30
`
`400 600 800
`
`Fuel flowrate, g/hr
`
`Figure 6: Influence of Mass Diffusion in Fuel on Deoxygenation
`
`Copyright © 2002 by ASME
`
`GE-1012.005
`
`

`
`deoxygenation rate when the fuel flowrate was less than 800
`g/hr (i.e., Re < 260).
`
`CONCEPT DEMONSTRATION
`A prototype fuel deoxygenator module was constructed for
`measuring the oxygen removal rate from jet fuel, and tests were
`conducted over a range of fuel flow rates (residence times) and
`temperatures. The module consists of a spool of 0.040-in.-ID x
`0.003-in.-wall oxygen-permeable membrane tubing installed in
`a cylindrical container. The container is purged with a low
`flowrate of nitrogen to create an oxygen-free atmosphere
`around the tube. Fuel flows inside the tubing, while the
`dissolved oxygen diffuses through the wall at a rate that is
`proportional to the difference in the oxygen partial pressures
`(driving force) across the surface.
`Several tests were performed to demonstrate the capability
`to deoxygenate air-saturated JP-8 fuel to very low concentration
`levels (<1 ppm) in a single-pass on-line flow arrangement. The
`tests were conducted with aerated fuel at ambient temperature,
`a much more difficult condition than occurs in an aircraft fuel
`system. (In Figure 5, it was shown that the deoxygenation rate
`increases with increasing temperature.) Typical results, shown
`in Figure 7, verify the rapid removal of essentially all of the
`dissolved oxygen from the fuel, beginning shortly after
`initiating the nitrogen purge gas flow. The fuel deoxygenator
`module was operated at a steady-state condition for more than
`one hour before terminating the nitrogen bleed and initiating an
`air purge, thereby ending the deoxygenation process (as
`indicated by the return of the oxygen sensor to the air-saturated-
`fuel starting condition). This result clearly demonstrates the
`feasibility of the on-line deoxygenation concept. The fuel
`pressure loss through the membrane filter can be minimized in
`a practical design by interchanging the fuel and nitrogen flows,
`with the fuel being directed over and around a bundle of
`membrane tubes (enhancing mass transport) and oxygen
`Starting
`N2flow
`
`c
`■- 100
`
`B 90
`ra
`5 80
`
`§ 70
`C 60
`
`£ 50
`c
`g 40
`
`° 30
`
`> 20
`
`w 10
`Q
`
`' '
`JP-8
`1 Ifuel-'W-
`1 SV = 50 hr1
`
`purge gas
`(N2 or air)
`
`T
`
`Membrane
`module
`
`r^ggg.
`—==5^
`fuel ~*~
`
`f
`
`\
`
`20
`
`I
`
`Starting
`air flow
`\
`M
`80
`
`-
`
`100
`
`120
`
`40 60
`Time, min
`Figure 7: Demonstration of Single-Pass Fuel Deoxygenation
`
`permeating into the tubes. A candidate design is described
`below.
`To provide a simple demonstration of coke suppression by
`on-line fuel deoxygenation, a small-scale test apparatus was
`constructed and tests were performed with typical jet fuels.The
`fuel deoxygenator module was used in conjunction with an
`ALCOR Hot Liquid Process Simulator (HLPS) to carry out a
`series of standard ASTM Jet Fuel Thermal Oxidation Stability
`Tests (JFTOT Procedure, ASTM Standard Method D 3241).
`The JFTOT is a pass-fail tester for specification purposes, and
`not a precision instrument for rating small differences in carbon
`deposition among fuels. It is used here to reveal potentially
`large differences in coke formation resulting from fuel
`deoxygenation.
`Heated-tube JFTOT-Procedure tests were conducted with
`air-saturated fuel (baseline) and fuels deoxygenated on-line
`using the membrane module. The dissolved oxygen
`concentration was monitored continuously and observed to
`remain constant at either the saturation or the near zero level
`(shown in Figure 7). Because the tube is heated resistively (i.e.,
`by an electrical current), the axial temperature profile is very
`steep. Type 316 stainless steel tubes were substituted for
`conventional aluminum tubes to permit operation at high
`temperatures and quantification of the surface deposit by
`carbon burn off using a LECO RC-412 Carbon Determinator
`(precision of ± 2%). Evaluations were performed with Jet A,
`JP-8, and JP-7 in 5-hour tests at maximum tube temperatures of
`635-860 F and fuel flow rate of 3 ml/min, conditions that are
`much more severe than the normal JFTOT Procedure. These
`fuels and run conditions were used previously in the
`development program for JP-8+100, and an extensive database
`is available for comparison of results [8].
`The test conditions and results are summarized in Table 1.
`They were performed in the order shown (including carbon
`burn off), i.e., alternating between air-saturated and
`
`Table 1: Jet Fuel ThermalO xidation Tests
`
`Fuel
`
`Runninq Mode
`
`Peak Wall
`Temperature Surface Carbon
`uq/cm2
`F
`
`Pressure Drop
`across Filter
`mm-Hg/min.
`
`Jet-A-
`Jet-A-
`Jet-A-
`Jet-A*
`
`Jet-A"
`Jet-A"
`
`JP-8
`JP-8
`
`JP-7
`JP-7
`
`Air-saturated
`Deoxvaenated
`Deoxyqenated
`Deoxyqenated
`
`Air-saturated
`Deoxyqenated
`
`Air-saturated
`Deoxyqenated
`
`Air-saturated
`Air-saturated
`
`Air-saturated
`JP-8+100
`Deoxygenated
`JP-8+100
`91-POSF-2827
`"'96-POSF-3219
`
`635
`635
`700
`860
`
`635
`635
`
`635
`700
`
`700
`840
`
`635
`635
`
`103
`4
`3
`6
`
`61
`4
`
`92
`4
`
`5
`4
`
`25
`1
`
`240/105
`0/300
`1/300
`12/300
`
`240/160
`0/300
`
`1/300
`0/300
`
`0/300
`1/300
`
`0/300
`0/300
`
`Copyright © 2002 by ASME
`
`GE-1012.006
`
`

`
`Peak Wal! Temperature = 635 F
`
`| Air Saturated
`
`| Deoxyge rated
`
`Jet A
`91-POSF-2B27
`
`JP-8
`96 -P OS f- 3* 05
`
`Jet A
`9G-POSF-3213
`
`JP-8+100
`(9€-POSF-3219* BetZ 8Q462)
`
`Figure 8: Deoxygenation Suppresses Coke Formation,
`Independent of Fuel
`deoxygenated fuel. With deoxygenated fuel, there was a
`dramatic reduction (more than an order of magnitude) in coke
`deposition on the surface and in particulate suspended in the
`bulk flow. For the deoxygenated fuel tests, no increase in
`pressure loss was measured across an in-line test filter located
`at the exit of the heater tube. Surface deposition results
`presented in Figure 8 demonstrate that deoxygenation
`suppresses coke formation, independent of fuel type (cf, Jet A,
`JP-8, and JP-8+100) or batch (cf, 91-POSF-2827 and 96-
`POSF-3219). The data in Figure 9 indicate that the
`deoxygenator is effective in suppressing autothermal reactions
`up to 860 F peak wall temperatures, confirming the initial
`hypothesis. Furthermore, deoxygenated Jet A performed just as
`well as JP-7, a special high thermal stability fuel that is highly
`processed and much more costly (see Figure 10). Therefore, it
`can be inferred that deoxygenation could make possible
`increasing the maximum allowable temperature of Jet A to that
`of JP-7, or from 325 F to 550F , more than doubling the
`available heat sink. The membrane module was operated for
`approximately 40 hours with no change in performance.
`Clearly, additional testing in simulator rigs and engines is
`JetA(91-POSF-2827)
`
`120
`
`E
`
`c
`
`o .-
`o e"
`a.
`V o
`8 *«
`re
`3
`
`E3 Air Saturated
`
`■ Deoxygenated
`
`ssSüüsü
`
`700 F
`Peak Wall Temperature, F
`
`^^^H
`
`Figure 9: Deoxygenation is Effective up to 860 F PeakWall
`Temperature
`
`Figure 10: Deoxygenated Fuel Performs as well as JP-7
`
`required to validate the concept and establish the level of
`deoxygenation necessary for coke mitigation.
`
`MODELING
`Oxygen permeation rate through a membrane is governed
`by the fundamental rate equation [9]:
`
`(3)
`
`(4)
`
`dn0i (out) _ k(AP02 )A
`dt ~ 8
`where k is the oxygen permeability of the membrane, AP02 the
`oxygen partial pressure difference (driving force) across the
`membrane, A the membrane surface area, S the membrane
`thickness, and dno2 (out)/dt the oxygen permeation rate through
`the membrane. A mechanism-based kinetic model (function of
`time, temperature, pressure and concentration) for oxygen
`permeation through the membrane was derived from Eq. 3 and
`used in conjunction with the test data to estimate the filter size
`required for a practical (i.e., low-volume/high-fiowrate)
`deoxygenator.
`Oxygen removal efficiency (T\0R) is defined as:
`Cn
`n0
`c
`
`"2(0)
`where C0 is the initial concentration of oxygen dissolved in
`
`jet fuel at ambient conditions, and C^ is the final
`concentration of dissolved oxygen. The oxygen removal
`efficiency is primarily determined by the partial pressure of
`oxygen on the backside of the membrane (Pback), the membrane
`permeate diffusion coefficient (A), the membrane surface-
`area/volume ratio (r ), the oxygen mass transfer coefficient
`
`{fur), and the residence time (/)• The performance equation
`developed from Eq. 3 and 4 is given by:
`
`P„„
`1--
`
`uno)j
`
`•exp(-^fio + —
`
`"2(0)
`
`(5)
`
`Copyright © 2002 by ASME
`
`GE-1012.007
`
`

`
`In this equation, Pback = 0.21 • Pmim, or 0 for N2 purge, X
`= diffusion coefficient (i.e., permeability/thickness), and y =
`Po2ICo2 Im (for Jet A, JP-8, and JP-8+100 7 = 0.0127). The
`key kinetic parameter, the oxygen diffusion coefficient (A)
`through the membrane, is determined from the membrane
`permeability test data shown in Figure 3. The mass transfer
`coefficient of oxygen {fm) in liquid fuel is a constant for a
`particular operating condition and fuel flow configuration. As a
`demonstration of model validation, a comparative plot of test
`data against model predictions is illustrated in Figure 11, and
`shows very good agreement. Under these conditions, the fuel
`flow was highly turbulent. Therefore, the mass transfer
`coefficient was assumed to be 1. This assumption was validated
`experimentally. To develop a more comprehensive design tool,
`fundamental parameters for estimating the mass transfer
`coefficient of oxygen in fuel are required.
`100
`
`JP-8+100
`Membrane thickness = 0.001-in.
`r = 140 in."1
`T = 256 F
`Pback = 0.016 psia
`P~_ = 3.09 psia
`^(0)
`
`-
`
`-
`
`c 90
`.s
`5 80
`n
`VI
`S? 70
`
`I 60
`c
`S 50
`c
`o o
`c 40
`o
`a> >.
`S w
`"O ffi
`> 20
`to
`tfi
`5 10 Model
`
`O Data
`
`10
`
`1 I 1 . ..._J
`15 20 25
`Residence time, s
`
`40
`
`Figure 11: Kinetic Model Validation
`The membrane permeation model described by Equation 5
`was used as a conceptual design tool. Simulations were carried
`out to size an on-line fuel deoxygenator for the maximum
`(-30,000 lb/hr at takeoff) flowrate requirements of a typical
`commercial aircraft engine, and for different levels of fuel
`deoxygenation (i.e., 90 to 99%, or 7 to 0.7 ppm). A 50 percent
`fuel flow (open) area is assumed for the optimum combination
`of residence time and membrane surface area, corresponding to
`the minimum component volume. The results, presented in
`Figure 12, illustrate that a deoxygenator of 10-gallon volume
`would be adequate, and a smaller volume sufficient if the
`percentage oxygen removal requirement (efficiency) could be
`relaxed. Note that the filter size is very dependent on the level
`to which the fuel must be deoxygenated, since the driving force
`decreases exponentially as oxygen is removed from the fuel. A
`determination of the optimum level of fuel deoxygenation was
`beyond the scope of this study.A lternatively, the deoxygenator
`might more appropriately be sized for the cruise condition,
`
`30,000
`
`Filter size, gallon
`
`Figure 12: Fuel Deoxygenation Filter Sizing
`
`where engine temperatures are high,fu el flow rate is lower than
`during acceleration, and flight time is longest.
`
`SYSTEM DESIGN/DEVELOPMENT
`The membrane deoxygenator can be configured like a
`shell-and-tube heat exchanger to maximize the surface area of
`the membrane contacted by the fuel and minimize the volume
`of the component. As illustrated in Figure 13, the fuel will flow
`through the shell, over and around many micro-porous tubes
`that have been coated on the outside with an ultra-thin
`membrane layer. The inclusion of flow baffles in the shell
`enhances the diffusion of oxygen to the surface by forcing the
`fuel to follow a lengthy, tortuous path over the membrane.
`Dissolved oxygen will permeate into the tubes from the shell,
`driven by the difference in chemical potential across the
`membrane.
`Control of the oxygen partial pressure on the backside of
`the membrane can be accomplished by the use of a vacuum
`pump or by purging with a low flow rate of a gas that does not
`contain oxygen (a sweep gas). Because no more than 70-ppm
`oxygen is being removed from the fuel, only a small vacuum
`pump would be required (less than lA HP for a typical
`commercial aircraft). For operating safety, any fuel (liquid or
`vapor) that may leak through the membrane will be returned to
`the fuel storage tank. Alternatively, the sweep gas can be
`precharged and recirculated in a closed loop, and oxygen
`transferred from the fuel can be removed by an adsorbent that
`would be replaced periodically. Also, the pressure of the sweep
`
`fuel inlet
`
`sweep (purge) gas
`
`baffles
`
`~7 /
`
`membrane-coated
`porous tubes
`
`02 + sweep gas
`(return to tank)
`Figure 13: Fuel Deoxygenator Design Concept
`
`deoxygenated fuel
`
`Copyright © 2002 by ASME
`
`GE-1012.008
`
`

`
`gas could be controlled to minimize the stress on the
`membrane.
`The next step in the development process is to construct
`and demonstrate a reduced-scale prototype, and to validate the
`design and fabrication methodology through extended duration
`testing in fuel system simulators. A detailed system
`design/integration and the construction of a full-size component
`would follow, leading to a final phase to fabricate a full-scale
`component and demonstrate performance in an engine test.
`
`CONCLUSION
`The results demonstrate the feasibility of a novel on-line
`fuel deoxygenation concept, and the potential for improving the
`fuel thermal-oxidative stability and increasing the cooling
`capacity. The principal conclusions are:
`. On-line removal of dissolved oxygen from jet fuel using a
`membrane filter is a feasible method for significantly
`increasing the usable heat sink of the fuel. The results
`suggest that practical size deoxygenators may be designed
`for use in aircraft systems.
`• Fuel deoxygenation is very effective in suppressing
`autoxidative coke formation, making it possible to increase
`the maximum allowable temperature and more than double
`the available heat sink. The membrane filter is capable of
`deoxygenating the fuel to a level below that at which
`significant coking occurs, and should reduce maintenance
`in aircraft fuel systems.
`• The key variables controlling fuel deoxygenation are the
`difference in oxygen partial pressure across the membrane,
`the membrane thickness, and oxygen diffusion in the fuel.
`Thin membranes, operation at elevated temperature, and
`turbulent flows enhance performance.
`• Fuel heating increases the permeability of the membrane
`and the chemical potential across it.
`• The mechanism-based kinetic model for oxygen
`permeation through a membrane is useful for designing a
`practical deoxygenator component.
`
`ACKNOWLEDGMENTS
`This paper is based on research supported by the United
`Technologies Corporation, and work performed for the Air
`Force Research Laboratory, Propulsion Directorate under
`Contract F33615-97-D-2784, administered by Dr.T im Edwards
`of the Fuels Branch, Turbine Engine Division. The authors
`gratefully acknowledge the test support provided by Mr. David
`McHugh.
`
`REFERENCES
`[1] Hazlett, R. N., 1991, Thermal Oxidation Stability of
`Aviation Turbine Fuels, ASTM Monograph 1, American
`Society for Testing and Materials, Philadelphia, PA.
`[2] Spa