ad-a239 300
TRANSCRIPT
AD-A239 300
91-07268
Di tAIM NOTICE
THIS DOCUMENT IS BEST
QUALITY AVAILABLE. THE COPY
FURNISHED TO DTIC CONTAINED
A SIGNIFICANT NUMBER OF
PAGES WHICH DO NOT
REPRODUCE LEGIBLY.
IASA Contractor Report 187081
kDVANCED EXPANDER TEST BEDPROGRAM
PRELIMINARY DESIGN REVIEW REPORT
Pratt & WhimeGovernmnem Engines & Space PropulsionP.O. Box 109600West Palm Beach. Florida 33410-9600
May 1991 T :. '3
J i,
Prepared for:Lewis Research Center !Under Contract NAS3-25960 -'
NASANational Aeronautics andSpace Administration
CONTENTS
FO)REWORD . . . . . . . . . . . . . . . . . . . . . . . .. v
IINTRODUCTION .. . . . . . . . . . . . . . . . . . . . . . . I
11 SUMMARY...................................................... 3
A. Design Approach............................................... 3
B. Operating Cycles............................................... 4
C. Oxygen Turbopump............................................. 4
D. Hydrogen Turbopunip............................................ 5
E. Nozzle-Thrust Chamber Assembly..................................... 7
F. Control System................................................ 8
G. Integrated System............................................... 10
III ENGINE CYCLES AND OPERATION....................................11
IV COMPONENT DESIGN............................................. 29
A. Mechanical Design Requirements..................................... 29
B. Turbopump Overview............................................ 34
1. Turbopump Design Requirements................................. 34
2. Risk Reduction and Verification Plans.............................. 353. Turbopump Testing......................................... 35
C. Oxygen Pump................................................. 421. Design Features........................................... 42
2. Material Selection.......................................... 43
3. Liquid Oxygen Turbopump Operating Conditions....................... 43
4. Inducer/Impeller........................................... 44
5. Turbine Blisk and Shaft...................................... 45
6. Interpropellant Seal (IPS)/Vaporizer............................... 457. Bearings................................................ 45
8. Housing................................................ 46
9. Structural Analysis......................................... 4610. Thrust Balance............................................ 47
11. Instrumentation............................................ 47
iii
D. Hydrogen Turbopump ......................................... 761. Design Features ......................................... 762. Primary Turbopump. ...................................... 763. Secondary Turbopump ..................................... 79
E. Turbopump Hydrodynamics ..................................... 1031. Hydrodynaic Design Approach .............................. 1032. AETB Oxygen Turbopump ................................. 1033. AETB Hydrogen Turbopump ................................ 104
F. Turbine Aerodynamics ........................................ 1411. Turbine Aerodynamic Design Approach .......................... 1412. Oxygen Turbine Description ................................. 1413. Hydrogen Turbine Description ............................... 1414. Turbine Methodology and Verification ........................... 142
G . Bearings . . . . . .. .. ... ..... .. .. . .. . . . . . .. .. . . ... ... .. ... .. 1491. Design Conditions ...................................... 1492. Nomenclature and Background Information ........................ 1493. Design Description and Trade Studies ........................... 1504. Bearing Methodology and Verification ........................... 152
H. Combustion System ............................. ........... 173I. Injector/Igniter Assembly .................................. 1732. Combustion Chamber Assembly .............................. 1743. Exhaust Nozzle Assembly .................................. 174
I. Hydrogen M ixer ........................................... 186
J. Control System ............................................ 1881. Requirem ents ......................................... 1882. Electronic Controller ..................................... 192
3. Valves and Actuators ..................................... 1944. Sensors and Cables ...................................... 196
V SYSTEM MECHANICAL INTEGRATION ............................... 251
VI RELIABILITY AND SYSTEM SAFETY ................................ 266
VII APPENDIX A ................................................ 274
iv
FOREWORD
This technical report summarizes the results of the Advanced Expander Test Bed preliminary design aspresented at the Preliminary Design Review held at the NASA-Lewis Research Center (NASA-LeRC) on 29-31January 1991. The work was conducted by the Pratt & Whitney (P&W) Government Engines & Space Propulsiondivision of the United Technologies Corporation for NASA-LeRC under Contract NAS3-25960. Effort underthis contract started on 27 April 1990.
Mr. Wiliam K. Tabata is the NASA program manager, and Mr. James R. Brown is the P&W programmanager.
v
SECTION IINTRODUCTION
NASA mission studies have identified the need for one or more new space engines. The new propulsionsystems are to be oxygen/hydrogen expander cycle engines of 7,500 to 50,000 pounds thrust or more; and mustachieve high performance through efficient combustion, high combustion pressure, and high area ratio exhaustnozzle expansion. The engines will feature a wide degree of versatility in terms of throttleability, operationover a wide range of mixture ratios, autogenous pressurization, in-flight engine cooldown and propellant settling.Other engine requirements include: long life, man-rating, reusability, space-basing, and fault tolerant operation.
The Space Chemical Engine Technology (SCET) Program is charged with developing the technology basefor the design and development of these new space engines. The Advanced Expander Test Bed (AETB) willsupport this objective by providing a vehicle for the following:
" Validation of the high-pressure expander cycle concept
" Investigation of the system interactions, transients, dynamics, control functions, and health monitoringtechniques
" Verification of design and analysis codes to assure scalability and minimize the risk associated withspace engine development
• Investigation of throttling and high mixture ratio operation
* Testing of advanced, mission-focused components made available from other SCET contracts
• Evolution into NASA's Focused Test Bed Engine.
To satisfactorily perform these functions the AETB must challenge technology limits while providing ahigh degree of flexibility and rugged, reliable, low-maintenance operation. The AETB engine requirements aresummarized in Table I. A nominal operating thrust of 20,000 pounds has been selected.
Table 1. AETB Requirements
Propellants Oxygen/Hydrogen
Cycle Expander
Thrust >7500 lb (20,000 lb Selected)
Pressure Nominal 1200 psia
Mixture Ratio 6.0 ± 1.0 (Optional Operation at 12.0)
Throttling 201' Minimum (5% Desirable)
Propellant Inlet Conditions:Hydrogen 38 R. 70 psiaOxygen 163 R. 70 psia
Idle Modes Tankhead (Nonrotating Pumps)Pumped (Low-NPSH Pumping)
Life I(0 Starts2 Hours (5 Hours Desirable)
The AETB is being designed using the latest component technologies and design and analysis methods.Although similar to the SCE concepts, the AETB will differ in the following important areas:
* Current technology will be used, whereas the SCE could use technology developed over the next few
years.
* The AETB will be designed for sea level testing; therefore, will not require a high area ratio nozzle.
* Relatively high-pressure pump inlet conditions are supplied to simulate boost pump discharge pressures.
" Component designs will be flight-type, but not flight-weight.
" Components will be arranged to simulate expected flight engine line volumes, pressure drops, and otherfactors affecting engine response; however, accessibility and interchangeability will be emphasized,rather than working to specific envelope limitations.
* Extensive instrumentation will be provided for control and validation of engine operation. Limitedhealth monitoring diagnostic instrumentation will be available, however, provisions will be made forspecial instrumentation and evaluation of advanced diagnostic techniques.
The AETB design is based on current technology; however, there are some areas where the stringentrequirements of the AETB (such as adequate chamber pressure to realistically evaluate advanced systeminteractions) introduce some uncertainty into applications of this technology. The results of ongoing Pratt &Whitney (P&W) test programs will provide component and subcomponent verification prior to engine fabricationto minimize this risk. These Independent Research & Development (IR&D) programs are aimed at extending thehigh-pressure engine technology base to include space engine requirements. The subcomponent tests consist of:
Full-scale combustion tests with prototype hardware to measure total thrust chamber heat flux, heatflux profile and combustion efficiency
* High-speed cryogenic bearing tests to confirm bearing life
" Oxygen turbopump interpropellant seal tests and hydrogen turbopump brush seal tests to confirm sealdurability and leakage
" Turbine airflow testing to confirm turbine aerodynamics and predicted leakage losses.
The preliminary design of the AETB began on 27April 1990 and was completed in January 1991. Thepreliminary design review was held 29-31 January 1991 at the NASA Lewis Research Center (LeRC). Thisreport is a summary of the preliminary design and of the information presented at the review.
2
SECTION II
SUMMARY
A. Design Approach
The Advanced Expander Test Bed (AETB) operates on oxygen/hydrogen propellants and has a nominaloperating point of 20,000 pounds thrust, 1200 psia chamber pressure and a mixture ratio of 6.0. The AETBdesign approach is focused on achieving high chamber pressure with adequate cycle and component designmargins and on providing a high degree of flexibility. The flexibility will consist of: (1) the ability to operateover a wide range of conditions, (2) the ability to easily interchange components, and (3) a versatile controlsystem that can accommodate changes in operating conditions, incorporate additional engine diagnostics andaccommodate new components.
Five unique features of the design contribute to achieving the desired flexibility: (1) the split expander cycle,(2) a 25 percent cycle and component uprated design margin, (3) dual-orifice injection to facilitate throttlingand high mixture ratio operation, (4) a dual-shaft fuel pump for rotordynaic stability, and (5) use of a provenadvanced electronic brassboard controller design approach.
In the split expander cycle shown in Figure 1, a portion of the first-stage fuel pump discharge flow is routeddirectly to the injector. The remainder of the fuel passes through the second and third stages of the pump andis used to cool the thrust chamber assembly and drive the turbopumps. The two fuel streams are mixed prior toinjection. The split expander cycle reduces the energy needed to drive the fuel turbopump and allows a highercombustion chamber pressure to be achieved. A major advantage of the split expander cycle is that controllingthe flow split between the thrust chamber cooling flow and the bypass flow benefits engine throttling and highmixture ratio operation. At these conditions the fraction of fuel passing through the thrust chamber cooling jacketcan be increased, resulting in lower turbine inlet temperatures and lower thrust chamber wall temperatures. TheAETB split expander cycle has the further advantage that it can be operated as a full expander cycle.
Figure 1. Split Expander Cycle
3
B. Operating Cycles
The AETB is being designed for 1500 psia chamber pressure and 25,000 pounds vacuum thrust, 25 percentabove the normal operating level. This approach builds in component design margin, and adds flexibility foroperating at off-design conditions when testing non-Test Bed, focused technology components. In addition, highmixture ratio operation can be demonstrated and the AETB can be run as a full expander cycle. Table 2 listssome of the cycle key parameters at the design thrust level, the normal operating point, a five-percent throttledlevel, a full expander operating point, and the high mixture ratio operating point. The engine can also be runin a pumped idle mode.
Table 2. AETB Cycle Conditions
Normal Uprated Full HighOperating Design 5% Expander Mixture
Point Point Thrust Cycle Ratio
Thrust, lbf (Vacuum Equivalent) 20,000 25,000 1,000 16,400 17,000Chamber Pressure, psia 1,200 1,500 65 980 1,000Mixture Ratio 6.0 6.0 3.5 6.0 12.0Nozzle/Chamber Coolant Exit 957 1,020 750 1,000 805Temperature, *R
Fuel Pump Speed, RPM 87,700 99,200 18,900 90,000 79,000Fuel Pump First-Stage Discharge 1,640 1,920 103 1,840 1,490Pressure, psia
Fuel Pump Third-Stage Discharge 3,500 4,500 251 3,300 2,670Pressure, psia
Fuel Turbopump Horsepower 1,670 2,520 22 1,690 966Oxidizer Pump Speed 42,500 48,900 8,240 38,300 40,100Oxidizer Pump Discharge 1,900 2,360 154 1,630 1,500Pressure, psia
Oxidizer Turopump Horsepower 348 530 4 296 362
C. Oxygen Turbopump
The oxygen pump is a single-stage centrifugal pump powered by a single-stage, full-admission turbine. Thepump discharge pressure is 1900 psia and pump speed is 42,500 rpm at the normal operating point. Primarysubcomponents include a three-blade inducer designed for pump stability, a high-efficiency, single-stage pumpwith a shrouded impeller, an interpropellant seal package to separate the liquid oxygen and gaseous hydrogenturbine drive fluid, a single-stage, full-admission turbine, two 35-mm ball bearings and a 27-mm roller bearing.Attention is given to leakage control through careful seal placement, configuration selection and design. A crosssection of the turbopump showing the basic features is shown in Figure 2.
The oxygen turbopump hydrodynamic design emphasizes attainment of stable operation over a wide flowrange and achievement of cycle performance objectives. Moderate suction specific speed requirements have beenselected to avoid pump induced instabilities. The inlet-to-discharge diameter ratio of the impeller, a critical factorwith regard to instabilities, has also been limited within successfully demonstrated levels for the suction specificspeed selected. The impeller is shrouded and has a low discharge blade angle.
4
Figure 2. Oxygen Turbopump
The turbopump interpropellant seal package is based on use of knife-edge seals and gaseous helium to purgethe seal cavities. The helium purge is required only for ground testing. The seals ensure separation of the oxygenused as bearing coolant from hydrogen leakage through the thrust piston, and restricts propellant overboardleakage. Knife-edge seals work best when they are sealing a fluid in the gaseous state. A radially slotted rotatingslinger/vaporizer is located between the pump ball bearing and helium dam to vaporize any liquid oxygen leakage.
The oxygen turbopump turbine is a conventional single-stage, full-admission, turbine. The turbine ispredicted to attain a stage efficiency of 82 percent.
D. Hydrogen Turbopump
The main hydrogen pump is a dual-shaft, three-stage centrifugal pump with the first stage and inducerdriven by a single-stage, full-admission turbine and the second and third stages driven by a second one-stageturbine as shown in Figure 3. At the operating point, the fuel pump runs at 87,700 rpm to provide a dischargepressure of 3,500 psia.
Approximately 50 percent of the hydrogen exits the pump after the first stage and flows through a controlvalve, bypassing the coolant jacket, and flowing into a mixer downstream of the turbines. The remaining hydrogenproceeds through the final two stages of the high-pressure pump and is used to cool the thrust chamber assembly.
5
ALSECONDARY FUEL PUMPPRIMARY FUEL PUMP 1
Figure 3. Dual-Shaft Hydrogen Turbopump
Particular emphasis is placed on achieving stable pump operation over a wide range of throttle ratios andmixture ratios, as well as meeting cycle performance requirements. The pump inlet configuration shown wasselected primarily to provide adequate bearing support to meet rotordynamic requirements. The inlet struts alsoserve to minimize induced pre-swirl during throttling conditions to improve pump stability.
All three stages of the fuel turbopump employ shrouded impellers and low discharge blade angles tomaximize efficiency and provide steep head-flow characteristics for improved off-design stability. The secondand third-stage impellers are configured in an in-line arrangement to minimize leakage between stages.
The dual-shaft fuel pump is driven by two single-stage, counter-rotating turbines. Stage efficiencies are 82percent for the primary turbine and 84 percent for the secondary turbine. The fluid velocity leaving the first-segment turbine exits in the same direction as the counter-rotating second-segment turbine, thereby significantlyimproving the efficiency of the second stage by eliminating the second-vane gas turning losses found in typicalco-rotating, multi-stage turbines. Both stages are high efficiency, high reaction stages. The back-to-back, single-stage turbines eliminate the large overhung mass of a multi-stage turbine, a major driver in achieving subcriticalrotordynarnics.
The primary fuel turbopump rotor is supported by two 27-mm roller bearings. A second set of roller bearingssupports the rotor of the secondary fuel turbopump. Roller bearings provide the high radial stiffness, 3.0+ millionlb/in., necessary to achieve adequate critical speed margin. The fuel turbopump roller bearings operate at a DNof 2.37 million at the nominal operating speed of 87,700 rpm and at a DN value of 2.7 million at the designoperating speed of 100,000 rpm.
6
Nozzle-Thrust Chamber Assembly
The nozzle-thrust chamber assembly consists of: (1) a dual-orifice injector for high combustion efficiencyI wide range throttling, (2) a conventional torch igniter, (3) a milled channel copper thrust chamber capableacheving adequate cooling with 50 percent fuel flow, and (4) an extended length tubular 7.5:1 area ratiolevel nozzle extension capable of providing total heat transfer rates equal to a high altitude nozzle, Figure
The operational flexibility comes from use of the dual-orifice injection concept to provide good atomizationI flow stability over the range of combustion pressure and mixture ratio desired, and a thrust chamber coolantvpath geometry that provides adequate cooling with 50 percent fuel flow at the 1500 psia design point. Thepercent fuel cooling capability is an outgrowth of the split expander cycle requirements. but has the addednefit of allowing substantial overcooling when desired.
Propellant Combustion ChamberConical Exhaust Nozzle
'ure 4. Thrust Chamber Assembly
The combustion chamber has a contraction ratio of 3:1 (chamber area-to- throat area) and a combustiontflber length of 15 inches to provide an optimum trade-off between heat pickup to drive the cycle, coolantssure drop, and combustion efficiency. A thrust chamber length of 12.0 inches is predicted to be sufficient toieve over 99 percent combustion efficiency; however, the 15.0-inch length was selected to provide additionalirogen heating for cycle power.
The chamber cooling configuration consists of 120-milled passages sized to maintain a maximum walliperature of 1460 R without exceeding the cycle allow- able coolant pressure drop. At the normal operatingnt the maximum wall temperature is 1390 R. Closing the jacket bypass valve as the engine is throttleduces thrust chamber wall temperature at low thrust. Closing the jacket bypass valve also produces low walliperatures with high mixture ratio operation. At an oxidizer-to-fuel ratio of 12:1 and 1000 psia chamberssure, the predicted wall temperature is 1160 R. low enough to prevent oxidation of the copper walls.
7
A conical tubular exhaust nozzle is used to duplicate the cycle thermodynamics of a flight engine witha high area ratio cooled exhaust nozzle. Typical flight engine configurations studied in the past have 1000:1area ratio exhaust nozzles regeneratively cooled to an area ratio of approximately 200:1. The conical tubularnozzle is 16.25 inches long and provides expansion from the thrust chamber exit area ratio of 2:1 to an arearatio of 7.5:1. The nozzle design is single pass, parallel flow tubular construction. The nozzle divergence angleis 7.5 degrees off the centerline.
F. Control System
The AETB control schematic is shown in Figure 5. The primary control points are: (1) the fuel jacket bypassvalves (FJBV) for thrust and chamber coolant flow control. (2) the secondary oxidizer control valve (SOCV)for mixture ratio control and control of injector primary pressure drop, and (3) the main turbine bypass valve(MTBV) for additional thrust control.
L2INLE--T
WtY L02 INLET
MAL 7 VALW
- 'E
Figure 5. AETS Simplified Flow Schematic
The selection of the baseline control system for a throttleable split expander cycle with dual-orifice injectionis relatively straightforward. The jacket bypass valve is required to control the coolant jacket flow for throttledand high mixture ratio operation. The oxidizer secondary control valve is required to control the oxidizer flow
split during throttling. These two valves alone provide adequate control for mixture ratio variation between 5.0
8
and 7.0 and thrust control between 125 and 75 percent power. For thrust control below 75 percent, an additionalmeans of relieving turbine flow is required.
The control system is designed to allow control points and control function to be readily changed throughoutthe test program. Optional control features consist of: a second turbine bypass valve, a line between the twoturbines to the turbine bypass line, spool pieces that allow either one or two turbine bypass valves to be placedin any of four locations, and fuel and oxidizer recirculation valves that can be used if necessary in throttlingtests. These optional features provide enhanced flexibility to investigate various control modes.
Other control options may be desirable for oil-design operation. Control of the power split between the fueland oxidizer turbines appears to be necessary for high oxidizer-to-fuel operation ratios, i.e., 12.0. This controlis provided by using a second turbine bypass valve or by moving the turbine bypass valve so that it reducesthe fuel turbine-to-oxygen turbine power ratio.
Full expander cycle operation up to 750 psia can be achieved simply by closing the jacket bypass valve.Operation to 980 psia as a full expander cycle is possible with minor engine modifications. Tank head idleoperation is achieved by moving one of the turbine bypass valves to a location to cut off the flow through theturbines, and inserting a blank-off plate in the common turbine line.
Recirculation of fuel and oxidizer may be desirable as a means of investigating pump stability. The plannedapproach is to design the turbopumps so that recirculation is not required, but to provide recirculation valvesfor later investigations.
Tank head idle starting may be the normal operating mode for applications in space. While the AETB willhave the ability to start in tank head idle, less complicated starting is planned for most testing. Liquid hydrogenand liquid oxygen will be supplied to the AETB at pressures simulating operation with boost pumps. Cooldownvalves will be provided for pre-cooling of the lines and turbomachinery.
The AETB brassboard controller will be an electronic rack mounted system. The engine test schem,,:cshown in Figure 6 includes the AETB, the controller brassboard and a monitor system. The brassboard controllerfunctions as a full authority controller during pre-run checks, cooldown. start, throttling, steady-state operation andshutdown. The monitor system is used to simulate the vehicle interface, down-load programs to the brassboard.control execution, record data and analyze data.
PeripheralDevicesI
Monitor Daa Bus AL'8 Volvo Solenoid Outputs AdvancedSystem Engine -- Engine Sensors -- txpanderController TestBreadboardI Actuator Control Signals , Bed
I Solenoid Volvo Poitions EngineI 3000 ft I200 ft.
ITFigure 6. AETB Breadboard Control Schematic
9
A device termed "EMPRESS" (Experimental Multiprocessing Real Time Engine Simulation System) willbe used to facilitate software control development and system and engine checkout. EMPRESS has been usedextensively as a simulation tool for advanced gas turbine and National Aero-Space Plane (NASP) testing. In thesoftware test environment, EMPRESS is used in place of the engine. Tests that would normally require an engineare performed with EMPRESS, thereby reducing test requirements and reducing the risk associated with first-timeuse of new software prior to engine test. Engine anomalies experienced during an engine test can be simulatedwithout jeopardizing hardware, and in most cases, the anomalies can be diagnosed with electronic simulation.
G. Integrated System
The integrated AETB system is shown conceptually in Figure 7. The relatively small size is driven by theuse of a sea level conical nozzle rather than a high area ratio bell nozzle. The nozzle is designed to provide a heatflux equivalent to a high area ratio nozzle so that cycle energy requirements will be met while allowing sea leveltesting. The engine can be throttled to 15 percent thrust without nozzle separation. Testing below 15 percentwithout separation can be accomplished by removing the conical nozzle, or by testing in an altitude facility..
The integrated system is configured to provide a high degree of operational flexibility and ease of maintenancewhile meeting the test bed requirements of providing a close simulation of line lengths, pressure drops, criticalthermal masses, and valve responses.
Secondary Oxidizer Injector
Control Valve
I-- Thrust Chamber
Fuel Turbopump
OxidizerTurbopump
Main Turbine Fe ubnBypass Valve Exhaust Nozzle Shutoff Valve
Figure 7. AETB Assembly
10
SECTION IIIENGINE CYCLES AND OPERATION
The Advanced Expander Test Bed (AETB) engine is a hydrogcrvoxygen split expander cycle engine designedfor a thrust level of 25,000 pounds and a mixture ratio of 6.0. The engine has a normal operating thrust of20,000 pounds and is being designed to operate over a throttling range of 20 to I and at mixture ratios between5.0 and 7.0. With minor hardware relocations, the engine can also operate as a full expander cycle and at amixture ratio of 12.0. A simplified flow schematic of the test bed configured in the split expander mode ofoperation is shown in Figure 8. The control system, described in more detail in Section IV, allows closed-loopcontrol of chamber pressure and open-loop control of mixture ratio. The main turbine bypass valve (MTBV)is the primary control point for thrust control.
The secondary oxidizer control valve (SOCV) is used to control the oxidizer flow split, thereby maintainingadequate injector pressure loss during throttling, and is the primary mixture ratio control valve. The fuel jacketbypass valve (FJBV) controls the coolant jacket flow to maintain a turbine temperature below 1060 R duringthrottling. Selected key parameters are presented in Figure 9 for the normal operating level.
The hydrogen pump (primary and secondary) is a twin-shaft, three-stage, centrifugal pump with the firststage and inducer driven by a single-stage, full-admission turbine and the second and third stages also driven bya one-stage, full-admission turbine. At the normal power level and a mixture ratio of 6.0, the fuel pump operatesat 87,700 rpm to provide the required hydrogen pressure level of 3489 psia. Approximately 45 percent of thehydrogen exts the pump after the first stage and flows through a control valve (FJBV), bypassing the coolantjacket and flowing into a mixer downstream of the turbines. The remaining hydrogen proceeds through the finaltwo stages of the high-pressure pump and is used to cool the chamber and nozzle assemblies. From there, mostof the hydrogen coolant is used to drive the oxygen turbine and the back-to-back fuel turbines. The turbine drivefluid is routed through the oxygen turbine before the fuel turbine to minimize hydrogen turbine leakage losses.Over 16 percent of the hydrogen coolant flow at the operating point bypasses the turbines through the MTBVwhich is used for thrust control. The turbine flow then combines with the jacket bypass flow in the mixer andcontinues to the injector manifold and the combustor chamber.
Below a thrust level of 6,000 pounds, the power split between the turbines is no longer satisfied by theMTBV and SOCV alone. For adequate mixture ratio control below 6000 pounds thrust, the fuel pump must beloaded up via the fuel pump recirculation valve (FPRV), or the fuel turbine power reduced through a secondturbine bypass valve (FTBV), Figure 10.
The oxidizer pump is a single-stage centrifugal pump powered by a single-stage, full-admission turbine. Toprovide the required pump discharge pressure of 1898 psia, the oxidizer pump rotates at 42,500 rpm at the normaloperating level. From the pump discharge, the oxygen flow is divided into two separate streams which supplythe primary and the secondary manifolds of the dual element injector. As stated earlier, mixture ratio control isachieved with a control valve in the secondary oxidizer flow line. The two oxygen flow streams mix within theinjector element and are propelled into the main chamber where combustion with hydrogen takes place.
Operating the engine as a full expander cycle can be achieved simply by closing the FJBV and allowing theentire hydrogen flow to pass through the second and third stages of the fuel pump, the chamber and nozzle coolingpassages, and the turbines. The increased coolant flow to the chamber and nozzle assembly results in a largepressure drop and a relatively low chamber pressure. However, if the FJBV is relocated to the position shown inFigure 10 (CCBV), the coolant pressure drop penalty is avoided and the cycle still operates as a full expander.The achievable chamber pressure in this configuration is 981 psia. High mixture ratio (12.0) operation of theengine requires control of the power split between the fuel and oxidizer turbines. This control can be provided byusing a second fuel turbine bypass valve (FTBV) as shown in Figure 11. Chamber pressure of 1000 psia can be
I1
achieved with this configuration. Figure 12 outlines the AETB operating envelope while Table 3 lists several keyoperating parameters. Detailed engine cycle sheets are contained in Appendix A for selected operating points.
During the AETB preliminary design phase, both steady-state and transient models were developed using theP&W Rocket Engine Transient Simulation System (ROCETS) developed under NASA Contract NAS8-36944.The models are programed in Fortran 77 and structured in a modular building block configuration which allowseasy changeout of component characteristics to evaluate effects on the engine cycle. One feature of the modelsis an expanded range of component characteristics, allowing simulation of the test bed across the entire operatingrange, from 0 to 125 percent of normal operating level. Propellant properties are rr.++ from tables using thelatest sophisticated, high-speed map reader, which greatly enhances the efficiency of the program. The propellantproperty values are taken from the most recent NBS programs while the combustion properties are based on theNASA-LeRC Chemical Equilibrium computer program. Advanced simultaneous balancing techniques, developedfor gas turbine simulations, are used to improve precision and increase program efficiency. The transient modelcontains all the critical volume dynamics and rotor inertias necessary to simulate test bed ignition and accelerationto any rotating thrust level from 5 to 125 percent thrust.
The engine models were continually used during the preliminary design. The steady-state model can balanceto either chamber pressure and mixture ratio, emulating an engine trim test, or to control valve areas. The enginemodel was used to define the operating envelope discussed earlier, as well as define the effects of increasedsecondary flows on the engine cycle. The characteristics of component performance variations are also currentlybeing generated. The AETB dynamic model was used to define the engine cycle start and shutdown transientcharacteristics. The control system requirements have been generated and the engine system abort levels definedwith this model. During the final design phase, control valve sensitivity studies will be conducted and thecontrol logic for the real time model formulated. The steady-state and dynamic engine models are deliverableto NASA-LeRC at the Critical Design Review. Subsequently, the models will be updated and verified usingtest data. Figure 13 presents the information required to validate the engine models and the testing whichis scheduled to provide that data. Specific guidelines, as listed below, were followed to maintain structural.aerodynamic or stability design constraints:
• Fuel pump speeds kept below 100,000 rpm - maintains rotordynamic margins
* Maximum turbine inlet temperature set at 1060 R - protects turbine and valve seals
• Oxidizer injector Ap/P, maintained above four percent - ensures stable combustion throughoutthrottle range
* Valve flow turndown kept below 50:1 - manufacturing limit to provide producibility
* Valves sequenced to: - prevent pump stall during transients, - prevent reverse flow through FJBV,and - maintain the oxidizer-to-fuel ratio (O/F) within flammability limits during start.
The operational phases for the test bed were divided into five sections:
* Prestart - fuel and L0 2 lines purged out, - pump cooldown achieved (controller monitors)
• Start - timed valve movement, - L0 2 lead start using primary injector, - main chamber ignition(controller checks), - accel to mainstage using secondary LO, injector
• Mainstage - closed-loop P, control, - open-loop mixture ratio control
12
Shutdown ---abort scenario, - valves move to failsafe positions, - propellant injectors purged
immediately
* Post-Shutdown - fuel and L0 2 lines purged.
Control logic for each phase will bc defined using the engine models. The start and shutdown transientswere patterned after RLIO expander cycle engine transients with modifications to allow for the split expanderconfiguration and the dual orifice injector on the oxidizer side. The present start transient accelerates from staticconditions to 100 percent power (1200 psia chamber pressure) in four seconds, as shown in Figure 14. Theshutdown transient decelerates from 100 percent power and shuts down to 10,000 rpm primary fuel pump speedwithin two seconds, as shown in Figure 15.
The AETB start transient preliminary valve schedule is shown in Figure 16. The primary oxidizer shutoffvalve (POSV) opens first, at a rate of 330 percent per second, to provide L0 2 to the primary L0 2 injector for aL02 lead start. A helium supply of 0.01 Ibm per second is used to purge the primary L0 2 injector. The primaryL0 2 purge is fully on at the closed POSV position and linearly ramps shut at 50 percent POSV position. Thefuel shutoff valve (FSOV) begins to open at 0.5 second after the POSV at a rate of 200 percent per second. TheFSOV provides fuel to the fuel injector and is purged with 0.01 Ibm per second of helium. The purge flow to thefuel injector is fully on at the closed FSOV position and linearly ramps off at 50 percent FSOV position. Oncethe FSOV is fully open, the upstream fuel valves, the fuel jacket bypass valve (FJBV) and the fuel cool-downvalve (FCDV), dictate the amount of fuel flow into the fuel injector. Chamber ignition occurs at 0.6 second asshown in Figure 14. With chamber ignition, the pumps begin to accelerate and the L0 2 primary injector fillsat 1.45 seconds. The secondary oxidizer control valve (SOCV) begins to open once the primary L0 2 injectoris filled. The SOCV is held at the 15 percent position to limit chamber pressure thereby preventing combustionproducts from flowing back into the secondary LO2 injector.
A criterion for the FJBV during start is to prohibit reverse flow of the heated fuel mixer fluid to the secondaryfuel pump inlet, possibly causing secondary fuel pump cavitation. After the primary L0 2 injector has filled, theFJBV is opened to increase fuel flow to the chamber, thus increasing the available power. The FCDV allows apath for additional fuel to flow through the pump stage and dump overboard, preventing fuel pump stalls duringthe acceleration that occurs while the L0 2 injector is filling. The FCDV begins to close at 1.6 seconds, after theprimary oxidizer injector has filled. By closing the FCDV, more fuel is passed through the heat exchanger andturbines, and into the chamber, thereby further increasing available power to the turbines. With this increasedpower, the pumps accelerate further, filling the secondary L0 2 injector at 2.45 seconds. Since the primary fuelpump is designed for twice the flow of the secondary fuel pump, and reverse flow through the FJBV mustbe avoided, the FCDV must close at a slow rate, i.e., 90 percent per second, to maintain stable pump flowcharacteristics in the primary fuel pump. Meanwhile, the FJBV cannot open too quickly since this would resultin starving the secondary fuel pump of flow, or reversing flow through the FIBV. For primary fuel pump speedsbelow 12,000 rpm, reverse flow through the FJBV is a concern. For primary fuel pump speeds greater than12,000 rpm, starving the secondary fuel pump of flow is a concern. Figure 17 shows the stable operation of allpumps for the start transient. At 2.45 seconds, the SOCV can ramp to its 100 percent power position. The mainturbine bypass valve (MTBV) is used to trim the power. The MTBV is opened to its 100 percent position once theprimary fuel pump speed is within 10 percent of the 100 percent power speed. During mainstage operation, theMTBV maintains a constant chamber pressure level using a closed loop function tied to the brassboard controller.Mixture ratio control is open loop and is set by controlling the SOCV area. The turbine inlet temperature iscontrolled by setting the FJBV area. The valve areas and the chamber pressure level will be programmed intothe controller software prior to the test and will be determined using the engine steady-state model.
The abort shutdown should be the worst case scenario for the components. All valves begin to move at thesame time signal, in this case 0.1 second as shown in Figure 18. The POSV closes at a rate of 330 percent per
13
second and the SOCV closes at a rate of 250 percent per second. The oxidizer cool down valve (OCDV) opensfully at a rate of 330 percent per second. At 0.3 second, no more LO2 will be provided to the L02 injectors andthe helium purge is activated at 50 percent POSV and 50 percent SOCV positions, using the same ramp as thestart transient. The L0 2 injector purge clears the L0 2 injectors of any oxygen. Meanwhile, the FJBV closes ata rate of 330 percent per second and the FSOV closes at a rate of 250 percent per second. The FSOV closesafter the SOCV and POSV to provide a fuel-rich, cool shutdown. The fuel injector helium purge is activatedat 50 percent FSOV position, using the same ramp as the start transient. The fuel injector purge clears the fuelinjector of any fuel. The FJBV closes before the FSOV to prevent reverse flow through the FJBV. The MTBVis opened, at a rate of 3 10 percent per second, to assist in powering down by bypassing flow around the turbines.The FCDV opens at a rate of 200 percent per second to allow continued flow through the pumps while they arestill pumping. The FCDV also provides a bleed path for the turbine and heat exchanger fluids once the FSOVand the FJBV are closed. Figure 19 shows the stable pump operation during the shutdown transient. From Figure15, it can be seen that both the chamber pressure and turbine inlet temperature drop when the POSV and SOCVare closed, due to the sharp reduction of L0 2 flow to the chamber and the lower mixture ratio. At 0.5 secondthe chamber pressure drops further and turbine inlet temperature increases as a result of the reduced fuel flow.As the helium purges the injectors of fuel and LO, the chamber pressure drops to near static conditions. Theprimary fuel pump speed powers down to 10,000 rpm in two seconds, at which time there is no more poweravailable through the turbines, and the pumps spool down from this point.
14
I.
1-j-
15
tag
16~
IT
17
IrI.-
18
0
x 0
0i
OD 0 )
0.0
CC
CcD
0.
U),
20
qCU
19
ugc
ii'1%t0 r:-c
W) I cow MI O
(Cf t A
-R-
(A00. i c W.
,
CL 0.
IiOi
260CLa
CDC
_~ WaN r. Q
jyC4 4
40 i
CL C
IL E Em&
0 CLE a 2
2 -. Z Zo,
IL C CL "ft 21
0<Cflc~ z-o
CLa 8 w
LU <u
LL- E
CL u
CL c- a 0 ooC I- E 4.-0 )E E C) C~ - z rC= =U C C .
CL~ E M >%0E~~ U= 0)
U- U) . < )0.
0 O. Co
oc
0 c
N . 0 0 C. w E0-ooo E co
E S O 0 a)S> >.
CL M 22
.in
______ ____0=
EU
0.1
009 L 0001 009VISd - 3uflSS3d ?J3BV4VH3
23
X10 5 1ST FUEL PUMP SPEED - RPM1.5 1.0 0.5 0.0
E
I If
__ __ _ 03 -3uiIN3 oi
242
- - -3V AV
an
E.
La
St aOZL OL 0 09
S 3U)AV
25a
In
9- CN
w w
U It I I = I /
U. 0 L6I
06~
CLC-
* 0 OCLOOO O~C
- / *0 xa 10
0.l II
0. coo / OO o 000i Ic
22
X- -3V3AV
o *0
* - U
0(6
IU 102__
*0V
(.!C04
270
*mn 4
UI I / /
CL E E
OOLOI 009 0 OOLOO 00 /CL p~d rq RQ.d ic
E EIL m AWU, S.IS
/L IA
/0coo j/o 0I0* oop~sd - 088 2@Jd ~ sd 218888-d A
28j
SECTION IV
COMPONENT DESIGN
A. Mechanical Design Requirements
The AETB design requirements include a combination of contract requirements, P&W standard designpractice and guidelines established by the unique requirements of the AETB Program.
Table 4 delineates some of the more important of these requirements. A key item to note is the use of a25,000-pound (25K) thrust design point to ensure safe operating margins at the 20K thrust maximum plannedoperating point. Also, for life predictions, a cycle is conservatively defined to start at ambient temperatureand zero speed/pressure, go to design point speed/temperature/ pressure, and then return to ambient conditions.Another pertinent point is the desired 1000-cycle life goal for all components except the thrust chamber andnozzle. These two components have high thermally induced stresses and will be designed to meet the 100-cycleAETB requirement.
The structural requirements are based on experience and NASA's structural guideline handbook #505B. Table5 gives the key structural factors of safety. The approach to assuring adequate burst margins is predicated onboth jet and rocket engine experience. The burst factor definition for housings is relatively classical; see Table 6.However, the factors for disks are definitely experienced based. For example, for hollow bore disks, empiricallyobtained "Material Utilization Factors" (MUF) are applied. These factors are simply correction factors for thedifferent disk materials used in disk designs. For solid bore disks another experience-based approach is used,i.e., plastic growth limits at the web or rim (whichever is more limiting) arc applied to determine speed limits.Stress related burst factors, which are proportional to speed squared, are always used for consistency.
The approach to rotordynamic design is also based on jet and rocket engine experience. The fundamentalapproach in the AETB is subcritical operation with a 20 percent speed margin at the highest power point. Forthe AETB, the goal is 20 percent margin at the design point. Other features important to rotor stability includedouble pilots for impellers, inducers and turbine rotors, and two-plane balance for impellers and turbine rotors.
Other design requirements include using average dimensions for stress analysis except for:
* Minimum thickness on pressure vessels* Worst case dimensions for low cycle fatigue (LCF) limited areas* Minimum dimensions on tiebolts.
Materials property data are based on -3 sigma characterizations taken from approved sources such asMIL-HDBK-5, P&W material manual, and P&W Space Shuttle Main Engine (SSME) Alternate TurbopumpDevelopment (ATD) materials manual. Hydrogen and oxygen compatibility are also important in selecting thematerials.
All structural analysis is conducted using proven, state-of-the-art procedures. Many of the computer codesemployed arc off-the-shelf while others arc P&W developed proprietary codes. Table 7 depicts some of thekey structural codes, types, and uses.
29
I- LUZ Z Z cc0 c-0 0 T- - 1 w 0 w
CLz CLO jL cbo z _z 0 <
a a WLL < ixo CO - Z X 0
CO)O -JccLXC- a IC w 2 - LU 'Cz
00 0 iwXL.L
00 000to - 0 - o - -- -O
LUU
000 CL N
CC *LL 0 ~
UI- CC
Cf) -J L 0 L F.Jad10 0 C- OZ W. cc
a. 0-~~ 00 00 Q oO
- - - nin
to 0C x C LU 0JL
w 0
CO) 0. a -
a. 0
Cl ) i U L l) a
0 Z 0
30
Lu
*~4 U)O ) )
LU cm
co
imi
0~ +
o0 0J 1-- W- w
F-~ Lu c6 a-n.0 F LLWULU O)CCC cr .
Q0 0 J O0
<.0 a .o
31
w _j L
Cl wo J j
o .j<;CO OiZ~E FU M* u
00 0-E L0 M C
0 7 WLU COC, 0ox z z
C4 3co
o0 0 !W0 O
Z232
Ci)ca x
oS x x
~xx xx
(00
S.a.
I.- k. C) > ~
LO -OJ U LDU m IJ -
~WW0~~LL33
B. Turbopump Overview
1. Turbopump Design Requirements
The oxygen and hydrogen turbopumps are configured to meet the design requirements stated in Section I1,as well as specific requirements for the turbopumps. These requirements include:
" Meeting the requirements of the split expander cycle at the uprated 25,000-pound thrust design point(125 percent of normal operating thrust)
• Providing stable pump operation over the desired throttling range of 5 to 125 percent of normaloperating thrust
* Ensuring subcritical rotordynamics by maintaining an adequate margin above the uprated design speed,with a goal of 20 percent margin
• Providing the desired life of five hours and 100 starts without overhaul (required life is two hours).
Specific features included in the oxygen turbopump are as follows:
" A long-life, knife-edge interpropellant seal package
• Stability enhancing features for deep throttling and high mixture ratio operation
* Single-stage, full-admission turbine
* Two ball bearings for rotor transient thrust control and a roller bearing for rotor stiffness
* Ball and roller bearing speeds that are within previous experience
• Control of turbine blade tip clearances
* Proven materials: INCO 718 for liquid oxygen service, A-286 and Super A-286 for warm hydrogenservice
* Use of advanced design and analysis tools, including finite element analyses where required.
Figure 20 shows a cross-section of the oxygen turbopump.
The hydrogen turbopump also includes specific features to meet the engine requirements:
" Dual shaft fuel pump for subcritical rotor dynamics
" Two roller bearings on each shaft for rotor stiffness
" Materials and bearing speeds based on Space Shuttle Main Engine - Alternate Turbopump Development(SSME-ATD) experience
• Inlet vanes and interstage struts for enhanced stability
34
Figure 21 shows a cross section of the hydrogen turbopump.
Risk Reduction and Verification Plans
To meet the AETB engine requirements, three areas of component technology should be confirmed. Thesere impeller producibility, cryogenic brush seals, and high DN (diameter times speed) roller bearings. A briefescription of the technology verification plans is presented below.
Impeller Producibility - The need to produce a high-performance engine, with deep throttlingcapability, and within a reasonable time and cost, places conflicting demands on the hydrogen turbopumpimpellers. High efficiency and a high turndown ratio dictate high-speed operation together with featuressuch as integral shrouds, a large number of thin blades, a large sweep (wrap) angle, and a small dischargeangle. This complex geometry and the high operating speed raise stress leve., to the point where wroughtmaterial properties are required. However, this same geometry makes the part very expensive (both interms of cost and schedule), and quite difficult to produce with conventional fabrication methods. Tosatisfy these conflicting technical and programatic requirements, P&W is investigating non-conventionalfabrication methods. The principal method being pursued is to fabricate the impellers by machiningdiscrete sections, and then joining the sections using a diffusion b'inding technique. This effort isunderway, and the diffusion bonding trials are expected to be complete by mid-1991.
* Cryogenic Brush Seals - Because of the high turndown ratio required in throttleable engines, controlof pump internal flows at low power settings becomes critical. This limitation arises because at lowpower settings the internal labyrinth seal leakage can become an unacceptably large fraction of thetotal pump flow. Brush seals have inherently higher pressure drop than labyrinth seals, hence internalflows drop off less rapidly as the engine is throttled. Seal testing in a cryogenic hydrogen bearingrig is planned to confirm the seal design. Metallurgical laboratory evaluation of the properties of thestandard brush seal material in cryogenic hydrogen will also be carried out.
" High DN Roller Bearings - The use of negative internal clearance roller bearings in large turbopumpapplications, such as the XLR129 and the SSME-ATD programs has been demonstrated. An ongoingIR&D program will verify the scalability of the design methods to small, space-engine-size bearings.As part of this same program, ball bearings will be tested up to speeds commensurate with spaceengine applications.
3. Turbopump Testing
Both sets of oxygen and hydrogen turbopumps will be tested at P&W's E-8 high-pressure test facility prioro installation on the test bed. In addition to acceptance testing to verify that the pumps are ready for installation)n the test bed, design methodology verification testing will be carried out. The testing is summarized below.
" Oxygen Turbopump - The acceptance testing will consist of four tests of 20 to 50 seconds durationwith each pump. Figure 22 presents a test mauix for the acceptance testing. After the completion ofacceptance testing on the second turbopump, design methodology verification testing will be carriedout on this unit. Three tests are planned for this purpose; the testing will focus on investigatingany performance anomalies identified during acceptance testing. Figure 23 shows the instrumentationplanned for acceptance and verification testing.
" Hydrogen Turbopump - The primary and secondary segments of the first unit will be testedindependently. This will allow hydrodynamic and aerodynamic performance of the two segmentsto be evaluated without regard to possible interactions between them. Four speed line and cavitation
35
tests are planned for each segment. The second turbopump will be tested as a complete assembly. Anacceptance test program consisting of six speedline and cavitation tests are planned for this unit. Aseries of three design methodology verification tests is planned to be carried out after completion ofacceptance testing. The test matrices for the hydrogen turbopump are shown in Figure 24.
36
L
f-.Lz~
1~
Lz.
37
CL
LL
cI-
0
tLL.
38
P-4 0
04 ~ .t-4m
0'.
0
Iii
C.L. m VC4 #
@000 .
Cl a .- 4 cm . -
39O
z
0F Oa. LU
z' . L
C-40
Ii
1!1w
U
IL +2R
I-L_jIC
0+05I 0
ILL
4c 40
Aa
coz0
-a0 z
Z 40000 -- o- 0c
o 4= -4..40L . e .00 5
- 6000t-Cst-_o t- *@Q CO e)a
o LUz I--
0 ~LU
LU ac
z C.)LUC
0 0
CA
633 0LS . 0 3CL~ m
0. ills I pII
I- 00
z 0z
C4 ~ 0'C@C$ - so c ' -4. I..
a4
C. Oxygen Pump
1. Design Features
The AETB liquid oxygen turbopump shown in Figure 25 is a singlc-stage. high efficiency, centrifugal typepump. The impeller is driven by a single-stage, integrally bladed, full-admission, reaction turbine that is integralwith the rotor shaft. An interpropllant seal package (IPS), separates the turbine hydrogen gas from the oxygenat the pump end of the turbopump.
The hydrodynamic design provides stable pump operation over the entire 20 to I throttlc range of the testbed engine. An axial screw-type inducer is used to maintain impeller suction properties over this broad operatingrange. The inducer is close coupled to the impeller to minimize rotor length, loads and flow distortions. Theimpeller is a low exit angle design with an integral shroud for efficient operation at low pump speeds.
Rotor speed is approximately 49,000 rpm at the test bed design point of 25,000 pounds thrust. Pump flowis 45 lb/sec at a discharge pressure of 2350 psia. The pump discharges into a vaneless double discharge volute,which minimizes hydraulic side loads on the rotor,
The pump end of the rotor is supported by a single. liquid oxygen cooled, ball bearing. Eighty percent ofthe bearing cooling flow is recirculated to the impeller inlet. Sufficient prcssurc and temperature margins aremaintained with this flow to insure that pump cavitation does not occur. Downstream, a vaporizer is used toreduce the density of the oxygen entering the IPS thereby reducing the amount of oxidizer that is lost overboardthrough the IPS.
The IPS consists of five sets of labyrinth seals. A helium purge in the center of the seal system is used forground test to ensure separation of the oxidizer and fuel within the turbopump. Radial clearances have been setat 0.003 inch for the hydrogen side to provide the required sealing capability, and a slightly greater 0.005 inchfor the oxygen side of the pump to reduce the risk of rubbing in liquid oxygen.
The turbine disk and blades and the turbopump shaft are machined as one piece. This integral fabricationfeature results in a less complex design and provides greater rotor stiffness for increased critical speed margins.Outboard seal wings are used to prevent flowpath gas ingestion or recirculation. Radial location of the wings canbe changed if required during the design phase to make small adjustments to rotor thrust balance. The turbineis a single-stage. full-admission, reaction turbine. The reaction of the blades is being adjusted during the designphase to balance the major axial loads on the rotor. The current reaction is approximately 50 percent.
Due to the small size of the turbopump, seal clearances are very important to pump efficiency. Turbinetip clearances are particularly critical in meeting the performance requirements of the turbopump at the designpoint. Therefore, the turbine tip shroud will be thermally conditioned with liquid hydrogen to limit radial growthand help maintain design tip clearance.
The turbine inlet and exit volutes are a unique design. The radial inlet and axial discharge of the inletvolute (and the axial inlet and radial discharge of the exit volute) allow access from the side of the housingfor machining. Without this accessibility, the volutes would be very difficult to produce. The two volutes areconstant-area, full-admission configurations that are mirror images of each other and are clocked 180 degreesto reduce rotor side loads.
Rotor critical speeds dictate that the turbine end of the rotor be supported by a roller bearing. A ball bearingis also incorporated on the turbine end of the rotor to take out an), axial unbalance during transient or off-designoperation. Both bearings are cooled by hydrogen supplied from the third-stage fuel pump discharge.
42
The turbopump housings are a robust design, providing substantial stiffness for rotor support and sufficientroom for plumbing and instrumentation access. The minimal number of components al.o reduces the risk oftroublesome joint leakages.
2. Material Selection
Figure 26 shows the primary material selections for the major components in the LO, turbopump. Thematerial selections arc based on the fluid and thermal environments. A286 and Super A286 are used for most ofthe major housing components for strength and resistance to hydrogen embrittlcnicnt. The materials selected forL0 2 and GO2 service arc based on NASA material compatibility testing, SSME turbopump material specificationsand material strength requirements. Testing of other materials and combinations of materials is ongoing and willbe used as appropriate in the critical design phase.
Liquid oxygen bearing material selection (inner and outer races, rolling elements and cage materials) isbased on SSME-ATD and RLIO experience. The use of AISI 440C for the pump end bearing inner race. forL0 2 compatibility reasons, results in significant stress levels in the race at room temperature. These stressescan result in a stress corrosion cracking problem in the inner race in a very short time. However, a new heattreat cycle has been developed that dramatically increases inner race shelf life. This improvement allows theuse of 440C without a prohibitively low shelf life.
AISI 9310 is used for the races in the hydrogen cooled bearings at the turbine end since L0 2 compatibilityis not an issue with these bearings. This material allows greater margins for the radial fits between the bearingsand shaft.
The rotor is made of Super A286 to provide resistance to hot hydrogen as well as high strength. The inducer,the impeller and the vaporizer are made from INCO 718, chosen for its strength and the fact that it has a slightlybetter L0 2 compatibility rating than other high-strength nickel alloys.
At the time of PDR, NASA was planning L0 2 frictional heating tests to evaluate several other alloys.When the data are available from these tests, the L0 2 turbopump materials will be reevaluated to ensure thebest selections are made.
3. Liquid Oxygen Turbopump Operating Conditions
Figures 27, 28, and 29 show the L0 2 pump operating conditions at the design point (25,000 pounds thrust).the maximum operating point (20,000 pounds thrust) and the minimum required turn-down thrust (4000 pounds).The figures show pump and turbine inlet and exit flow conditions as well as shaft speed, torques, horsepowerand tip speeds. All three figures represent an O/F ratio of 6.0.
Figures 30 and 31 show the internal flows at the 25.000-pound and the 4000-pound thrust levels. Temperaturesand pressures arc shown for major cavities as well as the mass flow rates. These flows are based on 0.003-inchradial clearances on the hydrogen labyrinth seals and 0.005-inch radial clearances on the oxygen labyrinth seals.Windage heat-up has been accounted for and the vaporizer effectiveness is based on the E727 Vaporizer programdeveloped under the SSME-ATD program.
The rotor has been axially thrust balanced at the design point conditions. Figures 32 and 33 show the rotorthrust balance at the 25,000-pound and the 4000-pound power levels, respectively. The total imbalance is only40 pounds at the design point and increases to nearly 500 pounds at the 4000-pound thrust level. This loadis acceptable because it occurs at relatively low rotor speed. The objective is to minimize the axial loads athigher speeds. Analysis is continuing to determine what the axial imbalance is at other power levels. With thisadditional analysis, it will be possible to determine how much bias can be applied to the rotor to minimize thebearing axial loads at high speed.
43
4. Inducer/Impeller
The inducer, shown in Figure 34, is a thrcc-bladc design with moderate suction specific speed (N,,) forlow-speed performance. The impeller, Figure 35, is a shrouded design with a low discharge blade angle forimproved throttleability.
During preliminary design, work focused on defining an impeller configuration that was not onlyhydrodynamically sound and structurally acceptable but was also practical to produce. Manufacturing capabilityproved to be the most limiting requirement for the impellcr. Fortunately, the design did not have to becompromised for manufacturbility and all design hydrodynamic parameters fell well within design experience.
Early in the design phase, IN i00 material was thought to be necessary to achieve the required structuralmargins for LCF life. This material selection was a concern because IN 100 did not rate well in oxygen promotedcombustion tests. However, preliminary structural analysis shows that INCO 718 will achieve the requiredstructural margins and is currently the material of choice.
Subcritical rotordynamics is a primary design goal. The pump bounce mode is very dependent upon theinducer/impeller length and weight. The latest impeller length is 0.070 inch shorter than the original configuration.The resultant critical speed is 122 percent of the design point speed of 48,863 rpm which falls in what is consideredto be the low risk area of Figure 36 based on P&W experience.
Structural analysis of the inducer indicates that the LCF life exceeds the design goal of 1000 cycles. Theanalysis is based on a geometrically scaled inducer model from an existing design. The model was modified toreflect actual blade thicknesses, and stress concentration factors (K,) were conservatively established.
Blade vibrational analysis, Figure 37, indicates a 45 percent frequency margin for the blade's first bendingmode at 4E. This is conservative, because the analysis did not account for centrifugal stiffening. Figure 38shows the Campbell diagram for the inducer blade.
The impeller structural analysis was based on a two-dimensional (2D), finite element, body of revolutionmodel with general boundary conditions applied. Figure 39. The analysis indicates that the impeller hubconcentrated stresses are acceptable and result in an LCF life greater than 1000 cycles. The stresses, K1, andthe factors of safety are shown in Figure 40.
The impeller blade analysis is based on a modified impeller model. Although the model was geometricallyscaled by the tip radius, actual shroud and blade thicknesses were used. The hub was fixed radially. Differencesnot accounted for include blade wrap angle and the number of blades. Further details are shown in Figure 41.
The results of the impeller blade stress analyses indicate the design is structurally adequate. The principalstresses are 61 ksi and -69 ksi at the trailing edge and hub intert.,ce. The factors of safety are 2.56 on yieldand 3.19 on ultimate strength. The stress concentration factor for a 1,000 cycle life is 3.4.
Blade HCF capability was predicted using a modified Goodman diagram shown in Figure 42. The Goodmandiagram was debited by the maximum allowable Kt of 3.4. At the steady stress of 69 ksi, the maximum allowablevibratory stress is 12.6 ksi. Based on P&W experience, this is adequate margin to proceed with the design. Theactual vibratory stresses will be calculated during the final design phase of the program.
44
. Turblne 8lisk and Shaft
The design of the AETB oxygen turbopump features a single-stagc, full-admission, 50 percent reactionirbine with a 7.00-inch tip diameter. The preliminary turbine airfoil design is shown in Figure 43. Turbinefficiency is predicted to be 82 percent at the design point thrust of 25,000 pounds. The turbine disk is integralwth the rotor shaft to maintain rotor critical speed margin and IPS clearance control. The turbopump Campbelliagram, Figure 44, shows significant margins for the blade modes.
Turbine blisk analysis shows positive stress margins. A plastic analysis model (PWA deck 5138) was loadedwth cavity pressures, rotor speeds, rim loads, and temperature gradients. Bore stresses are 61 ksi radial resultingri a burst factor of 1.49. Due to time constraints, the disk LCF life has not been calculated, but, based on theDw stress levels, LCF life is not considered to be a problem.
Blade thermal mechanical fatigue is a concern as well as disk axial thermal gradients. For both problemshe internal conditioning of the pump prior to engine start, during operation, and at shutdown, must be studiedtnd may have to be controlled to meet life requirements.
The L0 2 turbine blades are similar in size and shape to the fuel pump turbine blades and turn at approximatelytalf the speed. Therefore the stresses are expected to be below the fuel pump blade stresses.
1. Interpropellant Seal (IPS)/Vaporizer
The amount of L0 2 leakage in the IPS is driven by the density of the oxygen entering the IPS. Reducing thelensity reduces the oxygen lost overboard. Therefore, a vaporizer, modeled after one that has been successfullylemonstrated in the SSME/ATD L0 2 turbopump vaporizer, is incorporated in the design. Although the vaporizerequires additional turbine power of 38 horsepower (hp), trade studies showed the additional power to beicceptable and the oxygen lost overboard was reduced by 90 percent.
Trade studies were conducted to optimize the number of knife edges, diameters, and clearances. The current:onfiguration is a blend of all the beneficial features that could be incorporated without compromising othermportant design features. For instance, the improvement gained from adding one more knife-edge was offset)y a decrease in rotor critical speed margin caused by the resultant increase in rotor length. Another examples that the decrease in seal diameter and rotor diameter at the same time would decrease rotor stiffness andlecrease the chances of maintaining tight seal clearances.
The IPS package consists of a helium dam with 11 knife-edges on the hydrogen side and 10 on the oxygende. Concern about rubbing in L0 2 led to limiting the radial clearances for the oxygen side of the IPS to
).005 inch. Leakage control requirements necessitated the use of 0.003-inch radial clearances on the hydrogenide of the IPS.
Additional benefit on the LO2 side was gained from the incorporation of a stationary vane system upstream)f the vaporizer. This vane counteracts the pumping action on the backside of the vaporizer and reduces theIownstream pressure. The lower pressure results in less leakage overboard.
'. Bearings
The first approach to the rotor support design consisted of two 24-mm ball bearings for axial load controlmd a single 27-mm roller bearing for radial stiffness and critical speed margin, Figure 45. Many bearing'onfigurations were evaluated. . As the LO2 turbopump design developed, the rotor size ;ncreased, as did theiearing loads. To maintain design parameters within current experience levels, the ball bearing size was increased
45
to 35 mm. This bearing design is very similar to a bearing used in the P&W RL-10 rocket engine. The RL-10test and operating experience adds significant credibility and confidence to the design.
Material selection for each bearing was based on its location. For bearings exposed to liquid oxygen, 440Csteel was chosen for the application based on experience and L0 2 compatibility tests. This choice creates adesign hardship with the bearing inner races. When the race is installed on the A286 shaft, the required fit foranti-rotation is so tight that the bearing race has a limited shelf life. However, material processing and designchanges have improved the life expectancy of the bearing inner race to acceptable levels. Bearing coolant flowsare provided through constant area orifices and are sufficient to achieve the desired bearing life of ivC hours.
8. Housing
The pump housing design features vaneless volutes. The pump discharge volute. shown in Figure 46, isdouble discharge for reduced radial loads. The turbine inlet and exit volutes. shown in Figure 47, are singleinlet and exit to provide high efficiency and low losses. A unique configuration was developed to allow theturbine volutes to be more easily produced. The two volutes are a semicircular design originating at a partingline j,, the turbine housings. The strategic location of the parting lines allows these volutes to be machinedwith ;'onventional techniques.
The pump discharge volute is a traditional configuration that will be produced in two halves and weldedtogether. The original concept called for the two halves to be separate pieces and be axially loaded by thehousings: however, analysis showed that the pressure and thermal loadings were too high to consider this aviable design.
The major structural housings, Figure 48, are structurally robust, reflecting the test rig approach to thedesign. The robustness of the housings adds radial and axial stiffness to the rotor, providing increased confidenceto critical speed predictions.
The thermal gradients in the housings are significant and preliminary analysis indicates some isolated high-stress areas. Minor configuration changes may be needed during the final design phase. To maximize turbineefficiency, a tip clearance control scheme. shown in Figure 49. has been added to provide thermal conditioningto achieve the required diameter for proper turbine tip clearances. Thermal mechanical fatigue (TMF) is also aconsideration in the turbine inlet duct. Super A286 material is used to enhance TMF life.
9. Structural Analysis
Preliminary structural analysis was completed for several of the AETB L0 2 turbopump components. Thecomponents analyzed include the inducer blade, the impeller blade and hub, the turbine disk, and the turbineinlet housing.
Structural analysis of the inducer blade included a 2D finite element plate model for blade stresses andvibratory responses. Results indicate that the blade aerodynamic design will meet all structural requirements.Hub analysis is pending.
Analysis of the impeller consisted of a finite element 2D body-of-revolution model for hub stresses and a2D plate model in space for blade stress estimates. All analyses of the impeller are favorable.
A 2D bending analysis of the turbine disk was completed. Initially, axial thermal gradients causedunacceptable axial deflections, indicating a need to change the internal flow scheme around the turbine. Thecurrent flow scheme eliminates the disk axial gradient, and analysis indicates acceptable stresses and deflections.A plastic/residual membrane stress analysis shows adequate burst margin for the disk, Figure 50.
46
A 2D boundary element analysis program (BEASY) was used to generate thermal gradients for tht turbinedlet housing based on predicted surface temperatures and film coefficients. A 2D finite element structural analysis,as then used to predict the thermal stresses and deflections as shown in Figure 51. The analysis pointed outne location that was overstressed due to the thermal gradient. A more detailed thermal model is currently beingDnstructed which will determine the validity of this preliminary analysis. Thermal conditioning of the housingsmy be needed to achieve the desired durability at all locations.
0. Thrust Balance
In an early version of the L0 2 pump design, rotor thrust balance was controlled through the use of a thrustalance piston. This thrust balance piston generated balance loads through the use of high-pressure hydrogen fromie third-stage fuel pump discharge. Subsequent internal flow and cycle analysis predicted that the flows required) make the thrust piston work would have a significant detrimental effect on cycle efficiencies. Therefore, themust balance piston was eliminated and the axial loads are transmitted through the ball bearings.
The axial loads on the L0 2 rotor were balanced at the 25K thrust level by adjusting seal diameters andlightly changing the turbine reaction. Thrust loads have been calculated at the 4K thrust level and are less than00 lbf. At the 4K thrust level, the rotor rpm and bearing cooling flow rates are such that the ball bearings areapable of operating with the 500 lbf axial load.
1. Instrumentation
A preliminary instrumentation plan has been formulated for the oxygen pump. The plan includesstrumentation for control, safety, and performance. Figure 52 is an instrumentation schematic for oxygen
ump acceptance testing.
47
-41
48
C140N
I0
0.)
co
CI.,
LUL
Coo
cc 0
04 0
co.
49
ui 4)
al 0w 0.
C )LU 4C
U)
1.-0a) < Q 0 4)
16 0 * CI W*
0 40 0
C0~oi~l M '4
0a I-
C.C
uCA M
0In
M 0
w 00m F- 0.-1 0 C"x p: 0u
U) uN
0 -j U)
cm 050
IL L.
V) 0CC*a.
Cy 4)cc0 W U)0 0C
a& rca ) ' IL 0
I. z . Z
o N
Lo L
og
0 e
0zLUim
w w cc
K coN~L cm o -
(AA
0 a..
cc
cc
51u
CL 4)
IW 0w 0p
4 C U)0E w
9W 4Z *a
2 CU. IV
V! LO e
r- C4 4 -t
CDI 1 =
III.
Co i)
x to 40 nCP-
0.
00
co t:M U)
a-
J5
anE
533
T..
C4-
4- 4
C
04 W
C-...
544
-(z z c
0u 02
coo
0 c0Ja,-
0 =;
o o
ac
0 U0-J-a be a) 4
0
LOS55
- Q)
z 0 to
0 0;0i o
(nI
0~ j
CC'
j 1uS1
-
56
z zz u
W cc~...m 1 WWW - cw
aIL U LW -J U) I.-. uU) 0 -R I < u a.
LU 0 Z c m :3Lu 2
cc_0M C
M0j C c Iz mo
> I--4c - LL. 1- z c
0-'
LuuCC z I-
4D
57
a -j
0I
z
LU
CL
LU 0-~ z
z4
a-
00
44 0
CL
i-ILLu
58.
coI
EX 0
(D 0.
00 u0
~ccow o
-4)
0))
00m0C0
M0 0
~ 10
a. C-
CDCC
590
S0
-Lo-LL
0ZLL.
Z -.
0 L
LUU
,
0
WIz cIUJ 1
60
02
c 0
CC
OocV.
zm o
ai. CL
w 40
0.............. ....... .............. .. ... ... ... .............. ...
9.3 0
*0 in CO) 04 V. 0o to on C.. C-4q* 0 ed 0Y - V- - V- T. Y-
ZH>I AON3fO3IUl
61
co0
zoC Q
F- w
<'~ ZI- z
3~cc
-i JoL(0
Z-I
62
7LW C CD
LL A A A
w
m C 0
I R -
LL~~ Cl occ
m Lt) 0 cc
LLI C4 61
C14 04 N~
N N N
ILI0CLO
U.
U)U) CC L Q
5.U a.0
C
U)
C4 L
0 04 Cf)
63
z
-j0L
0 D 0
00 II 0 LrJ.
Wi CJCD
I- co l)cl
cic z
LLI ZWaL
cc . L.
64X C IL I
a
C
CD
00
a _ l
U)U
0 8
U. -
CT b 0lU aCC L .C;~~b*
IL 0Ix C) C!Auim s 0
c~ a.,g
Oua cc
t 0iz0oC
LLL
65R
0
zF.- <0 > z
0 N ~0
001= Z
00C,,C
IL
~4-0
toW
66
49
o~rU)C4W
0 u
4.0
MiI.-
LAJ0 0 i Mi Mi M0 I0 CV
z* t
*A >
o Z(1
LAL.
too..
................ .............. .............. .... ....... ............... ...... ... . .. ..
XdH "*Ot LI~d MLSI.
............ ............. ..... ............. ... ...... .......67 ...
C0~4
L~i UJZ0- I-- (
iI ; I.-
0 z0x
0 z Z U) w
LU >w
RUUW
uzuZ Lu 00
Cl0 Z0~ L LL w
LU-c2 00.
M M L -
Cl)4LU qw C) E UU
0m 0.OZW X
mmmm
kz 0
Z (J)6 0
_
0u
IL -
z uJU'U L
LU z
69
wus
z w
z0 M -o
cf
ILl
-i L-
zM z
~UJILl
z-3i
0 ul> a
0
LL.
70
ID
a0
x O)CC m
I-
I.-,
LL.
........ .......
U)0
00
71
CC
Cr)
00
72
LLo
v Ir -H
A (z
LU 0W5 a 1-- 0
ot u LUJ
0 c0
00
CL
LU
0
Lu 0
(000 1
0.) 0J z-N Lu
ui40z ~ Co -i
12 0. UJ
73
CoCl)w
VI)
0
C)
w
.R"411
....... ~ .
, W2,
CL co
47
zz 5
0 9L zF wJ
-In
IL 0
4 F
00
w z
I lx
3 I-
CL
cc +
0. 0.
~~04 4 )a
I-. IL0
0C.CL
Z LL in c0 0.
>-70
D. Hydrogen Turbopump
1. Design Features
The hydrogen turbopump is a dual-shaft configuration, with counter-rotating primary and secondaryturbopump rotors mounted in a common housing. The secondary pump is sized for approximately 60 percentof the primary pump flow. Short, stiff rotors, with two roller bearings on each shaft arc used to provide thenecessary support for subcritical rotordynamics. In the primary pump, inlet struts upstream of the inducer areincorporatcd to minimize prcswirl, and interstage struts between the inducer and impeller are used to raise shutoffhead coefficient, thereby increasing stability. Integral impeller shrouds provide increased efficiency, and multipleblades are used to enhance throttled stability. The full admission turbines provide 2525 hp and feature integrallyshrouded rotors to further increase efficiency.
2. Primary Turbopump
The primary turbopump is being designed for operation at 98,240 rpm at the design point. The liquidhydrogen flow rate for this condition is 7.50 lb/sec at an inlet pressure and temperature of 70 psia and 38 R.It is discharged at 1917 psia and 68 R.
The turbine flow rate is 3.85 lb/sec at inlet conditions of 3451 psia and 896 R. It exits the primary turbineat 2507 psia and 815 R. Figure 53 provides more detailed information regarding primary turbopump operatingconditions.
The materials used in the primary turbopump have been selected based on the availability of characterizedmechanical properties, suitability for the expected operating conditions, and resistance to hydrogen embrittlement.Figure 54 shows the materials selected for the major components.
a. Inducer
The purpose of the inducer is to pressurize the inlet of the impeller to the level required to prevent impellercavitation. The inducer material is wrought A 110 ELI titanium. It has three unshrouded blades, each with awrap angle of 292 degrees. The maximum blade tip speed is 1065 fps.
Preliminary hydrodynamic design of the inducer blades is complete. The final design work will includeadditional analysis to finalize blade thickness distribution and blade-to-hub fillet radii. The preliminary hub stressanalysis indicates a burst factor of approximately 4.0, which is well in excess of the required 2.25. Figure 55presents more detailed information about the inducer.
b. First-Stage Impeller
The first-stage impeller will be fabricated from wrought Al 10 ELI titanium. Hydrodynamic considerationsrequire the incorporation of an integral shroud, a large blade wrap, a large number of blades, a small bladeexit angle and a small blade exit height. These features make it impractical to machine the impeller in onepiece. A manufacturing development program has been initiated to produce a monolithic structure from multipleconcentric rings which can then be machined by conventional methods.
The latest 3D NASTRAN stress analysis of the current impeller indicates its LCF life is near the goal of1000 cycles at the design point. The peak nominal local stress in the region of the blades is approximately 127ksi. This would translate into an allowable blade-to-hub fillet K, of 1.4. The impeller burst factor is estimatedto exceed the required value of 1.5 at the design point. More detailed design and structural information ispresented in Figures 56 and 57.
76
:. Primary Turbopump Housings
The various materials used for the turbopump housings have been selected on the basis of strength, stiffness,"ompatibility with the contained fluid, and ease of manufacture. The use of castings was ruled out early in the,rogram due to concerns about their timely availability, as well as the high cost of casting tooling in the context)f the small quantity of parts to be fabricated.
The pump inlet duct, which is not in the bearing load path, provides the attachment point for the hydrogennlet manifold. It will be fabricated from wrought aluminum.
The remaining cold pump-end housings, comprised of the inducer housing/bearing support, the interstagestrut housing and the volute collector, will be fabricated from wrought INCO 718 nickel alloy. The inlet struts andLhe interstage struts help pump stability during throttling operation, while the vaneless, double-discharge volutecollector helps pump stability and reduces bearing side loads. The volute collector is made in two mirror-imagehalves, fully machined except for the cutwaters and conical diffusers. The halves are radially electron-beamwelded, then the cutwaters and diffusers are machined. The thrust balance seal stators are machined fromwrought cobalt alloy to provide transient contact surfaces compatible with the thrust balance rotor (tungstencarbide coatings on the titanium impeller).
The turbine inlet housing will be fabricated from wrought A-286 steel, selected for its resistance to degradationfrom exposure to warm hydrogen as well as for its high strength. This housing also contains the discharge portionsof the double-conical pump diffusers and the single-tangential entry inlet volute for the full-admission turbine.The inlet volute insulates the inlet housing pressure vessel from large thermal loads.
The turbine inlet volute will be fabricated from A-286 steel alloy. The forward wall of the volute willinitially have an access hole to allow tool entry for machining the inner flow passage. A plug will then be weldedover the access hole. This inlet volute acts as a heat shield for the turbine inlet housing, since it is subjected tolarger thermal gradients and higher temperatures than the inlet housing.
The first-stage turbine stator will be fabricated from A-286 steel alloy to provide resistance to thermalshock and exposure to hot hydrogen. The stator has 14 integrally machined vanes, with a span and an axialchord length of 0.25 inch.
The turbine in'termediate housing is comprised of an outer support flange, the second-stage stator, and adiaphragm, all of wlhch will be fabricated from various grades of A-286 steel alloy. The stator has 8 integrallymachined vanes, each with a 0.25-inch span and 0.50 inch axial chord length.
The primary and secondary turbine blade tip shrouds are mounted in the intermediate housing. They aremachined from wrought A-286 steel alloy for resistance to thermal shock and are held in position by radial pins.The shroud diameters are controlled by admission of high-pressure hydrogen into the intermediate housing. Thiscontrol method is used to obtain the 0.003-inch radial turbine tip clearance required at the design point. Figures58 through 63 show more detailed information about the turbopump housings and the turbine airfoils.
d. Primary Turbine Blisk
The turbine blisk, with its integral shroud, will be fabricated from wrought A-286 steel alloy. This materialwas selected for its high strength, thermal shock durability, and resistance to degradation under hot hydrogenexposure. Preliminary analysis indicates a burst factor of approximately 1.82 at the design point, based on thecurrent bore-rim temperature gradient and the criterion of <0.5 percent residual growth at the disk rim. (Therequired burst factor is 1.5).
77
The preliminary bore/web/rim axial offsets have been determined using a shell analysis and an initial finiteelement analysis. The results of these analyses indicate an estimated LCF life that exceeds the required 100cycle life, and approaches the 1000 cycle goal at the design point.
These analyses will be updated when iterations are complete on the turbine reaction levels needed to optimizeaxial thrust balance. The temperature distribution associated with that turbine configuration will be used to updatethe analyses. Figure 64 shows more detailed turbine blisk design information.
e. Shaft/Bearlngs/Rotordynamlcs
The shaft is an integral part of the turbine blisk, thus it will also be machined from wrought A-286 steelalloy. The shaft supports and axially preloads the rotor stack, and transmits torque from the turbine to the pump.
At the estimated design point pump torque of 817 in.-lb, the shaft nortinal shear stress is 11 ksi; the shearyield strength is 96 ksi. The spline bearing stress is also 11 ksi, well within the design criterion of 20 ksi. Theshaft is also the tiebolt for the rotor stack. The tiebolt will apply approximately 20,000 pounds of axial preloadto the stack. The preload will prevent the rotor stack from becoming unseated during operation. The rotor usestwo roller bearings to provide sufficient rotor support stiffness for subcritical rotor dynamics. They provide aradial springrate of approximately 1.0 x 106 lb/in, at the pump end and 1.5 x 106 lb/in, at the turbine end. Thebearings are discussed in more detail in a subsequent section of this report. The latest rotordynamics analysisindicates that the primary turbopump lowest critical speed is 109,500 rpm, a margin of 9.5 percent at the designpoint. At nominal normal operating point the critical speed margin is 29 percent. Figures 65 through 67 presentmore detailed information about the shaft, bearings and rotordynamics.
f. Internal Flows
The internal flows are comprised of the bearing coolant flow, turbine conditioning fow, and thrust balanceflows. These flows have been minimized to achieve acceptable operation in the split expander cycle.
* Bearing cooling requires a flow rate of 0.20 lb/sec for each bearing at the design point.
" Turbine conditioning flows are required to control turbine blade tip clearance and housing gaps and tominimize axial temperature gradients from one disk face to the other.
- Turbine disk rim seals are used to reduce hot gas inflow from the main flowpath, therebyminimizing cooling flow requirements.- Brush seals are used to reduce the leakage flow from the bearing compartment to the disk rim.- Static seals are used upstream and downstream of the turbine tip seal stators to reduce leakageof conditioning fluid.
" Thrust balance flows are minimized by using only single-face thrust balance systems, and by closelycontrolling thrust balance component tolerances. The values of these secondary flows are shown inFigure 68.
g. Primary Turbopump Thrust Balance
The thrust balance system of the primary turbopump is comprised of a rotor (the front shroud of me impeller)and a stator (the inserts fastened to the housing) with axially variable orifices between them at the shroud ODand ID. These orifices vary as a function of rotor axial displacement from the neutral or null position (equaltravel possible in both forward and aft directions) in a way that causes the pressure on the impeller shroud tovary and oppose the displacement.
78
The rotating compone,. projected areas and internal cavity pressures of the pump and turbine, together withthe axial forces on the turbine blades are used to calculate the net shaft load in the null position. This net loadis then compared to the thrust balance system capability (change in axial force versus displacement from nullposition at maximum available displacement) to determine the thrust balance margin.
The primary turbopump rotor is balanced within 200 pounds at the design point with an 86 percent reactionturbine. The estimated thrust balance capability at design point is +3700 pounds. Work is in progress forcalculation of the net rotor force at the null position and thrust balance capability for other operating conditionsincluding 1. 4, 10, 15 and 20K thrust levels.
Transient rotor axial force unbalance is controlled by contact surfaces in the thrust balance system at the ODand ID of the impeller front shroud. The amount of axial travel possible is set by adjusting the thrust balanceseal stators to provide a space 0.016 inch wider than the thrust balance rotor. The design and materials selectionfor these contact surfaces is similar to P&W's SSME-ATD high-pressure fuel turbopump design, using cobaltalloy stators and tungsten carbide coatings on the titanium rotor. The durability of these components is beingcharacterized in the SSME-ATD program.
3. Secondary Turbopump
The secondary pump segment is similar to the primary pump segment in, many respects. Both have double-discharge volute collectors for the impellers and both use single tangential access turbine discharge and inletvolutes. The secondary pump uses identical roller bearings as the primary, and both the turbine blisk and theimpellers are integrally shrouded. Significant differences between the pumps include the following:
* The secondary segment uses two impellers and does not require an inducer. Its capacity is approximately60 percent of the primary pump.
* The secondary pump impellers have only two sets of blades versus the three sets featured on theprimary pump impeller. Both of the secondary pump impellers use the same blade geometry.
* The secondary pump impellers are smaller in diameter than the primary pump impellers, and operateat a lower tip speed with correspondingly lower stresses.
* The secondary turbopump's lowest critical speed occurs at 123,500 rpm, providing a margin of 24.5percent at the design point and 46 percent at the normal operating point.
Figures 69 through 71 provide more detailed information on the secondary turbopump.
The secondary turbopump is designed for operation at 99,220 rpm at the design point. Liquid hydrogenflows into the pump inlet from the primary pump discharge at a rate of 4.77 lb/sec, and at a pressure of 1912psia and temperature of 70 R. Flow is discharged from the third-stage impeller at 4503 psia and II I R.
The pumping elements require 1180 hp, which is supplied by the expansion of gaseous hydrogen through theturbine. The turbine gas comes from the primary turbine at a rate of 3.97 lbisec at conditions of 2507 psia and823 R. It exits the secondary turbine at 1843 psia and 745 R. Operating conditions and internal flow parametersfor the secondary pump are presented in Figures 72 and 73.
The materials used in the secondary turbopump have been selected based on the availability of characterizedmaterial properties, suitability for the expected operating conditions, and resistance to hydrogen embrittlement.Figure 74 shows the materials selected for the major components.
79
The thrust balance system of the secondary turbopump is comprised of a rotor (the rear face of the third-stageimpeller) and a stator (the inserts fastened to the housing) with axially variable orifices between them. Thrustbalance is achieved as orifice size varies with rotor axial displacement from the neutral or null position, therebycausing the pressure acting on the impeller to change so that it opposes the displacement.
The rotating component projected areas and internal cavity pressures of the pump and turbine, together withthe axial forces on the turbine blades, are used to calculate the net shaft load in the null position. This net loadis then compared to the thrust balance system capability to determine the thrust balance margin.
The secondary turbopump rotor is axially balanced within 25 pounds at the design point, with a 57 percentreaction turbine. The estimated thrust balance capability at the design point is + 2000 pounds.
Transient rotor axial force unbalances are controlled by contact surfaces in the thrust balance system atthe OD and near the ID of the third-stage impeller rear face. The amount of axial travel possible is set byadjusting the thrust balance seal stators to provide a space 0.016 inch wider than the thrust balance rotor. Thedesign and materials selection for these contact surfaces is similar to P&W's SSME-ATD high-pressure fuelturbopump design, using cobalt alloy stators and tungsten carbide coatings on the titanium rotor. Durability ofthese components is being characterized in the SSME-ATD program.
80
70
clI)
in.0
LUCL
0.iC
LC
0
(00
a: in
m N 0 c
co.
81
C t ',
Ul z
-cc~ III 0c
in-
I I- U
V)-
in 82
I <
z
-Ji
LL-
us.
0 0Icc
ocI
83x
ww
C0U.
Z0Q =
m L.
LL)(
wz
zc'o
I-I a *i
Z 0) o X LU8 0
z0
LZ
I
01I-,
-J-
CCI.- X
C) U70a
85
04
LL0
CY
-td>a
co 86
U. w
0.rI
00
w
0
87a
w
us w-J 0 ~ -J
0- >w to
I.- w
z4
u88
zz
x101
I- c-
W -J
COC
Cr 89
z. L
(0 mc
zd
a. .
90
co "Z -a z am co - 2~
C4 cJ m.C3 a
io
00
w I I
C-
C:
0 (
W) 0C:a
CC U0 3 0
C\I
0> 0004
-
OODN
010
91
CO6
a.-H
Lo 3
(0l 0 0...
a.~
0 >.to -
t2 0
Go 21.
co~ o0
W) C
22 S?2 tD 9cc ( U c(
(UnN~ OLL
*U)
7&* C D ) 0) C)to o
CLa
-< a:)Q).'C~
ca . < 4
COi0' - .- 0< cm~m C)C 0 E
>, C U- t; - a; .2.=.t =) - ( CL ca f .~
r- v (DCE (
2 EU) = E M
92
CLC
00
00I---
a0.
0
C4 -J r
Z UJ
V). IL
Cla I -I
CC ui Pl
0 ca Ir
IL x z
U. Z0a
IL WL*4 Lu 0 C
* . Cl 'Rz .z z 4 >: I
ui 93
z 13L ,LL
LL U)~
a(z
P- 'CC/Z DLUU ELU woLo a
z V-I"
LU LCLJW
0 z cc~0 - LUOo
ooo
-- coJ LL LL 0.( 0
o C
~ LUO-N0 O&0z 3.z
z <% <~
0 UV)
in OraIx c94
0
ai
CL
.2
00
0I0
00
IC IL
o 1; 0 0
AD~j3N3 NIVUIS 8iOLOj
95
ci CYC0
c;
Nc;
cf)
C40
t
U) LU caCL
Z a:,E Mv- 42 EU.
CY
t C11
cq CY
CY
LL
TAC x 0
96
a 0
- 0 w E
o~ 4)0.0*0 U) 0C0
zv 04)U C
4) CC C.o~ ~ *0F-
EE
CYU) B)u. --i
0 C*L
0))4if
0.)7
E -2
m 0 U,
0 0)
Q~ C.) Z )
03O
0 C0 C4.
0WO EEC.-
.0. (4a.--
S~~~ 12( SL m nC
tv U)(01)-.
> )4 Y*cn-
970
z 00
LL w ui-J LL UC )
w l WW..
z CL
0LU)
LU OW u . C
Z '2O *-
-10-
0 -W 0.-
1Uw w
zz0 Z P I--
M I-
CL -J LL rCl LO to V
I.-mz t"La
mo98
0 U
0-
0
0 0in0
00
0
3 -J
inL
<0 00
-k9J3N NIUS JOJ-8
99~
A
@59
0,3'0-0
0...
-CL.
ca'
LO.
10
810
c~cq
e~eq
E
CME
00
101
S0
.2 (J)
GoL
C-j-
CLI
T1LO
E. Turbopump Hydrodynamics
1. Hydrodynamic Design Approach
The AETB split expander cycle requires high turbopump efficiency through high pump design speed, stage-headrise and low internal leakage. The AETB is also required to demonstrate a 20 to I throttling ratio and highreliability. The combination of these three requirements has not been achieved in any previous rocket engineturbopump design. Also of concern are the performance effects associated with the small size of the pumpsrequiring tight seal and blade tip clearances, smoo)th surface finishes, and tight flowpath dimensional tolerances.High reliability will be achieved by designing pumps with adequate hydrodynamic margins for pressure rise,flow capacity, suction performance and stability.
The design methodology applied to AETB turbopump hydrodynamic designs is an iterative process whichbegins with the engine cycle analysis and proceeds through three design phases: conceptual, preliminary and final,as shown by the schematic in Figure 75. Each phase includes hydrodynamic analyses which are increasinglydetailed, proceed from one-dimensional meanline analyses to two- and three-dimensional flow field analyses.
Codes for the analysis of 3D viscous flow fields are now being developed for use in the final design phase.One code developed for this purpose by P&W is the NASTAR program. The ability of NASTAR to modelcomplex, 3D flow fields has been successfully demonstrated on the centrifugal impeller shown in Figure 76.The computational results compared very favorably with test measurements. The NASTAR program will beemployed to analyze the AETB first-stage fuel pump impeller flow passages at the 25K design point and atlow-power throttle points. Results of these analyses could yield insight into hydrodynamic effects that limitpump throttling range.
2. AETB Oxygen Turbopump
The AETB oxygen turbopump configuration, shown in Figure 77, is a single stage design with an axial flowinlet and inducer and a radial flow impeller. The pump has been designed at a rotational speed that results in amoderate suction requirement for the pump and thus high potential for stability and performance. The inducerserves to allow cavitation to occur in a controlled manner and to gradually collapse any vapor in the inducer bladepassages before the flow enters the impeller. The impeller has an integral shroud and features a tight clearance,four-tooth, stepped labyrinth seal for low leakage and high efficiency. The impeller has also been designed with alow discharge blade angle for high efficiency and throttleability. The impeller discharges into a double-dischargecollecter, which includes a vaneless diffuser, a double-tongue volute and twin conical discharge diffusers. As inthe case of the hydrogen pump stages, this configuration also minimizes radial loads and improves throttleability.
The geometric design parameters for the inducer and impeller are tabulated in Table 8. The oxygen pumpinducer features three low-camber blades with a tip solidity of 1.98. The impeller features a 6/6 bladingconfiguration for a total of 12 blades at the discharge. It has a discharge blade height of 0.160 inch and adischarge blade angle of 25 degrees for throttleability.
The hydrodynamic design parameters for the pump are tabulated in Table 9. The pump design speed is49,400 rpm. This speed resulted from the turbopump meanline analysis, which optimizes speed until specificspeed, suction specific speed, head coefficient and impeller diameter ratio are within the experience correlationlimits. Subsequent refinements to the engine cycle reduced the pump pressure rise requirements resulting in alower speed requirement of 47,665 rpm at the 25K thrust point. At this speed, the pump has 39 percent NPSHmargin when operating at the required suction specific speed of 22.560. A moderate design point stage-headcoefficient was selected to achieve a steep (negative slope) head-flow characteristic essential to high throttleability.Since the axial loads will be controlled by the ball bearings, the pump does not require a thrust balance piston
103
and therefore does not have the attendant leakage and efficiency penalty. The predicted design point efficiencyfor this pump is 73 percent.
The pump inducer has been designed for an inlet tip flow coefficient of 0.i 15 as shown in Figure 78. Basedon a correlation of rocket turbopump experience for inducer suction capability, the inducer has a design pointsuction capability parameter of 34,750. taking into account the effects of hub and blade leading edge blockage.For the AETB oxygen pump, the suction specific speed is 28,630. This defines the peak of the predicted suctioncapability curve at the design flow coefficient shown in Figure 79. The required operating points for the pumpillustrate that the pump is predicted to have adequate NPSH margin over the entire engine throttling range,including full-expander operation.
The pump design and oft-design headrisc requirements and efficiency predictions arc presented in Figure80. The IK thrust point (20:1 throttling) is the lowest flow coefficient (Q/N) requirement, which is 21.5 percer.tof the design point value.
The size of the oxygen pump in terms of specific diameter has been optimized in the same manner asdescribed in the hydrogen pump section. Figure 81 shows that for the design specific speed. the oxygen pumphas been sized to achieve a near optimum efficiency. This also yields a low stage-head coefficient which isdesirable for high throttleability.
Figure 82 compares the pump design point efficiency and stage-head coefficient to the correlations ofprevious rocket turbopump experience. The efficiency prediction is somewhat lower than the correlation due tothe additional leakage (lower volumetric efficiency) requirements of the bearing coolant flow, which is recirculatedto the impeller inlet, and inter-propellant seal package over-board leakage. The head coefficicnt is on the lowside of the experience band, but this was intentional in order to achieve high throttleability.
The inducer has been designed in accordance with the incidence and solidity criteria presented in Figure 83.The inducer has an inlet tip bladL angle of II degrees and a tip solidity of 1.98.
Figure 84 presents the impeller blade angle distribution at the hub, mean and tip streamlines. The impellerhas six blades at the inlet and six spliter blades at about 45-percent blade length to contain blade-to-bladeloadings within experience levels. Figure 85 presents the design point results of the quasi-3D streamline analysisand shows that the velocity distributions are smooth from the impeller hub to shroud.
The oxygen pump throttle characteristic is presented in Figure 86 and is similar to that presented for thehydrogen pump. At 20:1 throttle ratio (IK thrust). the pump is required to operate at 21.5 percent Q;N. which isclose to the points demonstrated by the RLIOA-3-1 LO, pump and 350K oxygen pump which also demonstratedhigh throttleability. Based on this experience, the L0 2 pump is expected to achieve its throttling goal of 20:1.
The AETB oxygen pump will be instrumented with sensors in strategic locations from inlet to discharge asshown in Figure 87 to verity the hydrodynamic design methodology. Number. type and location of the sensorshave been selected in the same manner as those for the hydrogen pump.
3. AETB Hydrogen Turbopump
The AETB primary hydrogen turbopump configuration. shown in Figure 88, includes several design featuresthat contribute to the achievement of either high performance or high throttleability. The pump features an axialflow inducer and radial flow impeller separated by an interstagc strut to enhance throttleability. This feature wasdemonstrated on the XLR129 fuel pump and showed a significant increase in the shutoff head coefficient, therebysteepening the head-flow characteristic. Inlet struts, a vaneless discharge volute, and a moderate impeller exit
104
blade angle are included in the design to further enhance throttleability. The inlet struts prevent inlet preswirlat off-design or throttle points. Preswirl, if not avoided, reduces the headrise capability of the pump, resultingin a less steep head-flow characteristic. The combination of vaneless diffuser and discharge collector avoids thestall susceptibility of incidence sensitive vaned or airfoil diffuser vane cascades. The moderate impeller bladeangle minimizes exit flow recirculations between the collector and impeller discharge at the throttle points, whichdelays the onset of hydrodynamic pumping instabilities, providing increased operating range.
An impeller shroud has been incorporated to minimize leakage and improve pump efficiency. The shroudface also serves as an integral thrust balance piston. Face seals at the ID and OD of the impeller act in analternating (open/closed) manner with axial shaft travel, as the controlling orifices between the shroud cavityand the sink (low pressure, impeller inlet) and source (high pressure, impeller exit), respectively, in responseto changes in the turbine axial load.
The second and third stages of the secondary hydrogen pump, shown in Figure 89, include several of thesame features as the primary pump first stage. Common features include double-discharge collectors and moderateimpeller exit blade angles for improved throttleability, and shrouded impellers for increased performance. Four-tooth labyrinth seals with tight clearances are included, along with an in-line impeller arrangement, for reducedleakages and improved efficiency. The stage inlets are of the double side-entry type scaled from the successfulXLR129 fuel pump. Like double discharge collectors, double entry inlets also help minimize radial loads. Theyalso minimize inlet preswirl, thereby improving pump throttleability.
The geometric design parameters for the first stage inducer and impeller are tabulated in Table 10. Thehydrogen pump inducer features three low camber blades with a moderate solidity of 1.88, as required to controlhydrodynamic loadings. The inducer is more than adequate for the required pump suction performance.
The impeller features a 6/6/12 blading configuration for a total of 24 blades at the discharge to reduce flowdeviation, or slip, and recirculation at off-design operating points. The discharge blade height is 0.100 inch,which is the desired minimum for producibility. The discharge blade angle is 40 degrees. This angle is higherthan the ideal angle for maximum throttleability, but was a compromise to keep impeller steady-state bladestresses within allowable limits and to ensure producibility.
The geometric design parameters for the second and third stage impellers are tabulated in Table 11. Theimpellers have identical flowpath and blading geometry and feature a 6/6 blading configuration for a total of 12blades at the discharge. Like the first stage, these impellers have a blade height of 0.100 inch and a 40-degreeexit blade angle for structural and producibility reasons.
The hydrodynamic design parameters for the first stage pump are tabulated in Table 12. The first-stagedesign speed is 100,000 rpm at the 25K thrust point. Higher speeds would be within hydrodynamic experience,however, this speed is the maximum allowable based on the rotordynamic critical speed margin and bearingdesign requirements. A stage specific speed of 682 derived from the selected speed, pressure rise, and flowrate indicates the pump maximum efficiency potential. After including the additional leakage due to the integralthrust piston, the stage efficiency is predicted to be 60 percent. With the tip speed of 1934 ft/sec set by structurallimits and headrise set by the engine cycle, a stage-head coefficient of 0.558 results. The tip speed, along withthe rotational speed, also sets the impeller tip diameter of 4.432 inches. With tip diameter, blade height and tipspeed determined, the impeller discharge flow coefficient of 0.125 is calculated.
The hydrodynamic design parameters for the second and third stage impellers are tabulated in Table 13.The design rotational speed of the secondary hydrogen pump was also set at 100,000 rpm based on criticalspeed and bearing considerations. The major hydrodynamic design parameters were determined in a similarmanner as the first-stage pump. A stage specific speed of 780 gives these stages a higher efficiency potential.
105
The second-stage efficiency of 73 percent is higher than the third stage's 65 percent, since it does not have theadded leakage of an integral thrust piston.
The first-stage pump inducer has been designed for an inlet tip flow coefficient that provides the pump withthe required suction capability including adequate margin. This selection is based on the correlation of suctionspecific speed versus flow coefficient for previous rocket pumps equipped with inducers, as shown in Figure 90.
The first-stage hydrogen pump design and off-design suction requirements have been analyzed and are allwell within the predicted suction capability of the pump, as Figure 91 shows. This provides ample NPSHmargin at all the required split-expander cycle operating points from 1K to 25K pounds thrust as well as thefull-expander cycle points.
The pump design and off-design headrise requirements and efficiency predictions are presented in Figure 92for all three pump stages. The plots show that the throttle requirements of the first stage are more severe thanthose of the second or third stages. The higher shutoff head coefficient of the first stage is a result of the stagingof the inducer and impeller at off-design conditions due to the presence of of the interstage strut, which reducesthe angular momentum generated by the inducer at the inlet to the impeller.
The sizes of the three pump stages, represented by specific diameter, have been optimized based on theirstage specific speeds and the empirical correlations of pump efficiency, Figure 93. Since the second and thirdstages have higher specific speeds than the first stage, they inherently have a higher efficiency potential as thetrend of efficiency islands indicates. The stage specific speed and diameter together explicitly determine thestage-head coefficient.
Figure 94a shows the hydrogen pump design point efficiency predictions compared to previous experienceas correlated with specific speed. The stage-head coefficients of all three stages are within experience levels, asshown in the specific speed correlation, Figure 94b.
The incidence of the pump inducer at the tip of the leading edge has been selected based on a correlationof incidence at maximum demonstrated suction specific speed versus inlet tip blade angle. The inlet tip relativeflow angle, was determined from a streamline analysis of the inlet and inducer. This analysis indicates the angleis 5.0 degrees at the 25K design point. An inlet tip blade angle of 7.5 degrees, gives an incidence of 2.5 degrees,within previous experience as shown on Figure 95a.
The inducer solidity chosen for the hydrogen pump inducer is within the previous rocket turbopumpexperience correlation of suction specific speed versus tip solidity as shown in Figure 95b. This ensures thatthe pump will have adequate suction capability.
Figure 96 presents the impeller blade angle distribution at the hub, mean, and tip stream surfaces for thepump impellers, with the locations of the splitter blade leading edges indicated. The chordwise and spanwisedistribution of these blade angles, along with splitter locations, is the result of an iterative design process betweenthe geometry and quasi-3D streamline flow codes to optimize the internal hydrodynamic loadings. Results of thestreamline analyses, which include the effects of incidence, deviation (or slip) and hydraulic losses, are presentedin Figures 97 and 98 in the form of relative velocity versus percent blade length. Flowpath area distributionsand blade contours have been optimized to achieve smooth meanline velocities and diffusion rates.
The first-stage pump inlet is an annular duct which serves to deliver the flow over the No. 1 bearingcompartment to the inducer inlet. The axial inlet features two non-turning airfoil strut rows. The first strut hasfour equally spaced airfoils and provides structural support for the inlet nose cone and inner flowpath wall. Thesecond strut has eight equally spaced airfoils and provides support and stiffness for the bearing. An axisymmetric
106
streamline analysis of the inlet flowpath including strut blockage has been performed. Results of the preliminaryinlet design are presented in Figure 99. Hub and tip streamline absolute velocity distributions indicated anexcessive wall diffusion loading along the inner flowpath wall of the second strut row leading into the inducerleading edge hub. During the final design phase this loading will be reduced to an acceptable level by increasingthe flange inlet area and refining the flowpath area and curvature distributions. The increased inlet area will alsominimize the radial flow profile stemming from the inlet flowpath curvature entering the inducer and therebyenhance its suction performance.
The first-stage pump throttle characteristics are presented in Figure 100. Pump throttling is represented bya curve of percent of design flow-to-speed ratio (percent Q/N) as a function of engine vacuum thrust. Enginethrottle ratios of 20, 10, and 5:1 have been noted on the x-axis. At 20:1 (K thrust, 6.0 0/F ratio), the AETB firststage pump is required to operate at 22.7 percent Q/N. The RLIOA-3-1 fuel pump, which featured a first-stageimpeller with a 50-degree exit blade angle and a second-stage impeller with a 90-degree exit blade angle,demonstrated deep engine throttling with the pump operating down to 24 percent Q/N. Based on this experience,the first-stage fuel pump, with a 40-degree blade angle, is expected to achieve its throttling goal of 20:1.
The AETB hydrogen pump stages will be instrumented with sensors in strategic locations from inlet todischarge as shown in Figure 101 to verify the hydrodynamic design methodology. Static pressure taps alongthe impeller shrouds and backfaces will provide data for the calculation of pump axial loads. Inlet and dischargepressure and temperature sensors will provide data for the calculation of stage-headrise and efficiency. Dynamicwall static pressures will provide high response data to measure potential pump-induced hydraulic oscillationsover the full range of operation.
107
t',
cc P
50c I C
LO1
JL 0
iuXI
A~ 11o)C~
0
0olot E IIfItjdlIj~J1wi
§1 ~08 0 0
0coE
-
6 0
C0c
.i20a. cc C cx
ID aR. 0
0 CC 0E) 6
LL
2U).
C,,109
V
E~o
C
00 -
Sm-
CL'- -
,0'0 I-C
00) 0 C z
U) 0~g 0'aOM-~- P*' 4)CD
0G >..>WI-x
LLn 00L-LE
110
N - ;y o wi
*0))
.0 0 _ FMI__ %4.
S Co 0) p) o __ 0 0
CL (D
E 2
EE- E E~
~~~~ O~D DN la OQ AN
U -aziw
0 0 0.70 co mq CM 00)m
C 0 0i O q (6m q 1 0 cv) IT 0
IW ~ ~ ~ ~ ~ 0 (D-1 '4WI 0011s
IE C Cd
E0 E
.0 0 L
CL 0 026 Of) __) n
0 CL 0 V 0.00 -L U) I.
M 0 *n(n-L.
0)C 0 CLC 0 0. 0
E 0-0 cn cno w 0 0 3;=0.
0 00 r ..- z 4 EL. QX.u
ffU0. wZ 0. 0M' U)c0 C/)) co 0) 0 cmcm
0. 0 _ = 0 (. 0acc -0: co cnJ 0 n U) ) 0)) w
000
~I) (Z .~D CO (0CD
ElL
0 04
am .(00 I C
112
C4
0
x-2 0
CL 0
x .0
Liq
Z' t t *0900 * 1goio
,Irgg~rV~c O'r~dSOIJ09dSNOI01 0
_ _ _ _ 0 113
z
I.-
N*euUI1 - - "%J%
azzz zzzzz IU
IL zzzzzzzz Ia.
0aa - ___4
w ONN0,0X1 t 00Socco 0U
o~o~agoooI
a 9iw0
114
10'0,N
'0
.0
LA. U U. 16 0
00000 o o0000 0
00 0!.- ! 0
.-
00
*
.0
10
ao r% to L It n N0 0 0 * * * '0
LM331A33 OR io
115.
00.
co
C)
Sl
0Lt
116)
0
0 0CD C!
0
G M
Ci z0- w
EE 0 .0 W, CI)
< ~ ~ Q. .
:3 0
-~ ~~~ 0 o o o
CL a:
0 117
CL'
Y0.
0 0 0C0 0
co CL
0. 0
IL) 0
E~ CLL
0. 0. "t
CL
EEa..
L I-
118
A-o p
I
044-
00
119
:3 CD
E c C
0 CI ' 0
j V
0"D -O
oo C 0
0 0 0 0 Cl
0~0 .~ ~
0802 O.,LL
4)g5
0.2in 4
Cu .0OD <
0 0 N
E . -b I - -0 0 0 00M '1 0 0 =3hi .. Fm In0 2
*~~c CC CY
a: >
120
ILLJ 0
0 0
Lii 0
-J-
00
CL~
E x 0's i
Inn)
OZL 0L 0909 o 0 z0
sap(NO) 0
0L9-2
12
ILI
IL13 I r
-P
-A
FO
Ii or__cc
12
age.
>
son~
_ 0 r-
1-66
E 06
0. Ir
0.0123
00
E
V S
0 U)
1244
toJ
tv)Ul)CV)CV) (D 0 c) cr) ~r) 0 0 +
4))
4)0 )
4 - *' 5 ) ) -1.. 4L)L 4 <4 0 L ).10 4 4 )< )..
-* 4 4)
E .2-~c 0 E C)
p P2 .0.
to
S00 0 c Z oN La 00 cN
Z-I
125
a. 0 0 +
a.a
0~ ~:0 0
CC
w. w 0 z
LU0 a 0 N.j0 E -r i -
0Q.0 (U C
E ~ .-126
0 0o 0i La) ~ cr. d0 C) e)f1lc 0: 0 L 0
f- f CIF MW Y CY(0 00 (D
(.4
E E
E ~L *
CL 0)0
w a L C U w oEof 'a an S 0
0)- 0 0. )a)~~~4 000 - 4a
~ 0 0(.CO 0 o5nn a. 0 . 0
0O 0 . C)) cnV
u) EW 0 _n 0. m.. 0 CL CM
9LF.. .2 o 0 . CD~
-7E - 0 >- 0 0 -0 0 w(U:
4.. 4.(U (l(
a. o. v . .= C
CL if0zz Z
1271
(D 0 C~l-l (Vr )
C6 m6
TD f'. o 0DU)V y6(ai0 In 0 y)
CLL
CL 4)
Inn
4) 1L 0
-o 0 C m0(D0
E wo E
us z :
0
128
0
44
Ob .0
cp c0
0
o
I q
9.'L. . O- . .. .9-0 *0' Vo,... '.0
SOIX r.
Ssffrlrdardsalliar a 0J
129in
-i
LL.
w.
z I
a%~lf:-k a
wzzzzzzzzzww
LL
00000S0000 00
00"0aawooq z
flooDs pools 0009 00991 099 09900@ON OfldS 31.I13dS Noils
130
*L ULL:
- CLL LL
0)) o 0 0S~ CD (0 (D00 0
r_ ( 0 6 .0 D (0 0 -0) iz j - -
4) LO * 0 0 Le 6e e elCJC C*J -~ -~ ~ (D q CO CVJ.
00))
0
'2'0
c
0) 0) 0 i 6 c i 6 c
00
Vw 00
131~
V)
E0LLJU
C114
LL--
.... ... .. ... .... .. .. ..... .. .. ... .. .... .. ... ...Lj .... ... .... .. . .... .........
LL'U
C l ... ..... .................
2F~
132s
00LL 00
iat) CY
4) 'C -.
Lh U Q to
G 1
00. 0 v
0 lN 0
- C-00 S
Ed _&IX 40.x
M u0I-Gil
C14 co
04
0§
a. M (-
LU 0
00
133
LA--
0
- NO CC
CL m-
-~ 0 N0 x0c
cc~a 0)-
* F a-,EUE
LL-u
o0
o 00 C4 0
e-a
CLC
0 x%Ln cc
co CL
E 134
m0
*00
0 :0
*Y m0. .o cc
(n -. - 00
V) Cm v0o c
c 0 0
00(I)s
C *N3
((x0
0 135
to
a) 0
0'
cimcj 0~ -JE aa)d
Lo WU 0a
ca a) * Y00
to C~J
MU c
0 . .
co .2 COC4
(U 0 0
o 0o
0 0~M I
00 0
c'Jo
0 C3O 0Eca 0 ro cm' an co I.
cm
Cr 0.. (0=I Lcoa
Z o 0 0 0 0
LL >>
136
GO00.0
00
S.00
-o 0
Eo
c~co
co -0
o o .2V
CL C
Go * coQ
oo to0
coC 0 0I 0D
aw. . 0 0mc o It C
.2.
0 -0
E >2 *e * .* 4 0 0
0 1 Q
>- 0
*m Go 0 V C
a:a>
13
I job
4
1384
LI
00
4C
-I-
L6--
4z Do4O 0 o
sop~01
z0 =
C
4:-
I-I
aZ.IE
9L.
Of-
140
F. Turbine Aerodynamics
1. Turbine Aerodynamic Design Approach
The split expander and expander cycles demand high turbine efficiency to reduce engine size and weight.In addition, stable operation free of high vibratory gas loads that could cause bearing side loads should beattained. A full-admission reaction turbine was chosen to fulfill these requirements because this type of turbinehas proven more stable and efficient than partial admission impulse turbines during the development of the RL1Oand SSME-ATD turbopumps. Figure 102 shows that the performance characteristics of the AETB turbines fallwithin the area of demonstrated RLIO and ATD turbine experience.
The turbines are arranged in a back-to-back and counter-rotating configuration in the fuel turbopump toeliminate interturbine pipe losses and significantly reduce the second turbine vane gas turning losses. Low lossinlet and exit volutes are employed to reduce the first vane gas turning losses and eliminate the need for exitguide vanes. A constant static pressure gradient is designed into the volutes to eliminate circumferential pressuregradients that cause bearing side loads.
Mechanical options were chosen for high leakage efficiency and are necessary for low aerodynamic losses.A radial tip clearance of 0.003 inch at the maximum power running condition is the most dominant mechanicaloption chosen and is necessary to employ reaction turbines to their full efficiency. A passive tip clearance controlsystem, that maintains a cold case that shrinks on a dynamically and thermally growing turbine blade tip, isexpected to produce the desired close tip clearance at the full power design point. A thermal tip clearance thatcontrols the tip shroud radial position by using cold impinging hydrogen will back up the passive system.
The integral bladed disk and shroud will reduce parasitic leakage flows around the root attachment and overthe blade tips. The delicate machining of these blade flow channels will also allow 0.010-inch blade trailing edgethicknesses with a channel dimensional tolerance of +0.002 inch within a 0.050-inch throat gaging dimension.
2. Oxygen Turbine Description
The oxygen turbine elevation is shown in Figure 103 along with its design parameters. It is a conventionalsingle-stage reaction turbine with a volute inlet and discharge. The inlet volute flow enters the first vanetangentially, and reduces the first vane gas turning losses, as shown in Figure 104. The low axial velocitythrough this blading annulus, coupled with modem small blade manufacturing methods that allow small gaginggaps of 0.059 inch, trailing edge thicknesses of 0.010 inch, and exit gas angles of 4.8 degrees, enables thisrelatively low specific speed design to have high efficiency in this application. The tight tolerance control ismade possible with the advanced blade gas path machining capabilities being applied. The blading also has anintegrally machined shroud for low leakage and strength as a result of this machining process.
3. Hydrogen Turbine Description
The hydrogen turbine elevation is shown in Figure 105 along with its design parameters. It has conventionalsingle-stage reaction turbines with a volute inlet and discharge. The counter-rotating secondary turbine is mountedback-to-back in a manner conventional to current advanced fighter engine dual spool turbines. The reaction levelof the primary turbine is high enough to provide enough residual velocity from its blading to power the second-stage blade without any significant acceleration through the second vane. The angle of flow into the second vanevaries very little over the operating range as shown in Figure 106 so that efficiency varies only eight percentover this range. The high reaction first blade has a significant change in inlet angle but its inlet Mach number isso low that no significant inlet loss occurs nor is there any inlet separation. Initial computational fluid dynamics(CFD) analysis has shown this to be true and it will be confirmed as the final design is developed and tested. Theblades have an integrally machined shroud for low leakage losses and for shroud strength at high wheel speeds.
141
4. Turbine Methodology and Verification
The turbine aerodynamic methodology and design codes employed are listed in Figure 107. Analysisduring the preliminary design progressed through the meanline analysis. During the analysis, overall engineperformance was traded with overall turbopump size, cost and reliability. The efficiency, number of stages,diameter, airfoil stress and pull on the disk were selected in conjunction with the needs of the engine cycle.Once the engine cycle is finalized, the output of the meanline analysis will be used as input to the streamlineanalysis which defines the turbine radial flow, pressure and temperature maps for the case and disk structuralanalysis. The streamline analysis will provide the airfoil gas dynamic environment and lead to the design ofthe airfoil cross-section contours. These contours will be radially faired and the detailed analysis of the airfoilstress and pull will be performed to confirm that the initial estimates of disk stress, reliability and airfoil endwallcompatibility were maintained.
Of primary concern is dependence on turbine mechanical and manufacturing techniques which need to beverified in the test bed engine to support the turbine performance goals. Blade tip clearance control during engineacceleration transients, turbine tip shroud strength and integrity at full speed and temperature, the control ofparasitic leakages, and the manufacturing of small blade gaps and tolerances will be verified during componentand engine testing.
142
co 00
03-i4 c. 0
-2c ) 0
00a o 2
ccCL - 0 0
EE E0 = 12
~LL 04-
I CD
Q CD
*0 00 0o
7Z 0
x -v aJ I'0 cc z
F- .2 C'J (a
0
II-7wL X
< 0
< 0.C(n
0.OC c'j CD)
-62no
0-=
143:
C\ N- CV co co (DCJNt cm 4 0 r . 0 m t 0 w
0~
o c. -oCo ~ Cw. C C C
144
z
0 0
x
5 Vl
0
0)) cmoV* 0)CV)
.~~~o ........
0
0>>
0 >0
145
0O () LOl c N 0 ~06 cr U") (V0 N0 It IT
C.). CM 0! U' C
co 0) =r - 0 1. L
& - -M I -LO 0) c
a. Cf, c N6.
.*a 0
0~ I.- Co~ o )u ~
aD (D ~ , E CL0U). _)
a),W~ ~ c 0 < . %O 0I
146
z x~IU
lot'1I "1'
In.~. 0 CJ
CM
: 5 0 .
WG i-mc
cis,
00 CYCI
a w t
tZ
147"
4) .2 -~ 0
(3) a, wu.. C
~~ -0 )C
E z 0 - C)V
C~h4 a, 0)cD0..
c -a 0)
0)0 C~,~ r- 3:0 00 0
- - '0 0 C L y qC D
4 ..w = M iWu s
c CD ....0
o ~ .a.§..U U ,O )
r a c L~ 0 CaEna) 75ai c
.0 -2 a, 0 CD c
0,~ 0 , . B .- 5 -
~ -- 0 m0
d. u c 0 U) a, uUa, ) .!- 0)c
SC D "0 C /)fl n
.m > iZ E) c G.2 0 L a
m~ V 0' 0)3 E r
( w 0) U) - ) )"
>wD0-M h00CCL.-W 0 5 E.3: m m -Et5
148 'O ofl Q-
G. Bearings
1. Design Conditions
The rotor support system for the oxygen turbopump includes two ball bearings and one roller bearing. Theball bearings are axially preloaded with 300 pounds by means of springs to prevent skidding and provide radialstiffness at the pump end of the rotor. Two ball bearings were selected to provide transient axial load capabilityin either direction. The ball bearings are designed to carry transient loads up to 2500 pounds. The roller bearinguses negative internal clearance to prevent skidding and provide roller guidance. The roller bearing provideshigh radial stiffness to the turbine end of the rotor for increased critical speed margin. The roller bearing designis identical to the fuel turbopump roller bearings. The pertinent boundary conditions used to design the oxygenturbopunp bearings are provided in Figure 108. The rotor support systems for the primary and secondaryhydrogen turbopumps include two roller bearings on each rotor. Roller bearings were selected primarily basedon the high radial stiffness requirements. The roller bearings have been designed to carry radial loads up to350 pounds; the maximum expected load is 163 pounds. The roller bearings use negative internal clearance toprevent skidding and provide roller guidance. The pertinent boundary conditions used to design the hydrogenturbopump bearings are provided in Figure 109.
2. Nomenclature and Background Information
The AETB bearings have been designed to meet or exceed the conventional Pratt & Whitney (P&W) bearingdesign guidelines, and cryogenic specific guidelines that have been established under the SSME-ATD, RL10,and XLR129 programs. The major guidelines are listed below.
a. Ball Bearings
Internal Radial Clearance (IRC) (total clearance between the rolling elements and the races) - The internalradial clearance of a ball bearing sets the contact angle between the balls and the races.
Hertzian Contact Stress (load deformation at the ball to race contact forms an elliptical pressure area.) -
Based on 15 hours of SSME-ATD testing in L0 2, the maximum steady contact stress should be limited to 388ksi. The limit for liquid hydrogen (LH 2) use is based on five hours of testing at 483 ksi with RL1O bearings.For transient axial load capability, the SSME-ATD program has demonstrated 60 cycles at 665 ksi in LH 2.
Stress Velocity (SV) value (the product of the contact stress times the ball slip velocity, and a measure ofthe application severity at the ball-to-race contact.) - Based on SSME-ATD L0 2 and LH2 experience, the SVvalue should be less than 2.OM psi-fps to minimize wear and heat generation.
Ball Excursion (the distance a ball attempts to move circumferentially from the cage pocket center, where thecage speed is the average ball orbital speed.) - A ball bearing operating under combined axial and radial loadwill have a variation in ball-to-race contact angle around the circumference. Since ball orbital speed is relatedto contact angle, a ball will undergo an orbital speed variation as it travels around the circumference. This,in turn, varies ball-to-ball spacing. As ball excursion increases, ball-to-cage interaction loads and cage-to-cageguide land loads increase. In a cryogenic bearing, moderate ball excursions are desirable since the ball-to-cageinteraction is the mechanism for lubricating the bearing. The maximum ball excursion should be less than twotimes the ball pocket clearance.
B l Life (the number of hours that 99 percent of the bearings in a given application will operate withoutexhibiting a rolling contact fatigue failure of subsurface origin.) - The calculation is based on the Lundberg-Palmgren life theory. The life theory does not directly apply to a properly lubricated cryogenic bearing, since
149
wear or surface initiated distress are the predominant distress modes. Also, the calculated life is conservative
since the life theory is based on test data and materials and manufacturing technology from the 1930's.
Figure 110 provides graphical representation of the ball bearing nomenclature described above.
b. Roller Bearing
Internal Radial Clearance (the total clearance between the rolling elements and races.) - The roller bearinguses negative internal clearance for preloading the rollers.
Contact Stress (stress due to interaction between the rolling elements and the races.) - The undeflectedcontact between a cylinder and a surface is a line. The load deformation creates a rectangular pressure area. TheSSME-ATD program demonstrated 18 hours of operation with a contact stress of 300 ksi.
Edge Loading (stress caused by interaction between the roller comers (edges) and the races.) - Highstresses result when the roller is loaded at the edge. Edge loading is reduced or eliminated by crowning theroller. Based on the XLR129 15-hour durability demonstration, edge stresses should be less than 200 ksi.
Minimum Roller Preload - In a cryogenic bearing, roller guidance is controlled with roller preload. Theroller preload is obtained with negative IRC. The roller preload required to keep the rollers stable is determinedempirically. Based on SSME-ATD and XLR129 experience, the minimum roller preload for the AETB shouldbe 220 pounds.
Figure 111 provides graphical representation of the roller bearing nomenclature described above.
c. Cryogenic Bearing Experience at P&W
A broad experience base has been accumulated in the development of cryogenic bearings. Figures 112through 114 summarize the pertinent P&W cryogenic bearing experience gained through rig testing in the RLlO,350K, XLR129, and SSME-ATD programs.
3. Design Description and Trade Studies
a. Bali Bearings
An extensive trade study was performed to .ct the configuration. The trade study included bearing size,rolling element size and quantity, race curvatures, and IRC. The study also included the existing RL1O bearingto determine if it was suitable for the AETB. Figure 115 shows the contact stress and SV value for a 27x58-mmbearing and the 35x62-mm RL10 bearing. With the expected radial load of about 50 pounds, both contact stressand the SV value are within experience guidelines. As shown in Figure 116, the existing RL1O bearing hadadvantages over a new design and was, therefore, selected for the AETB.
The RLIO bearing is suitable for LH2 use only and requires modification for L0 2 use. The cage wasredesigned with L0 2 compatible materials. The L0 2 bearing cage will use Salox-M bronze-filled teflon insertsand a K-Monel shroud, a design proven in the SSME-ATD program. The inserts provide a transfer film lubricant,and the shroud provides a structural support element which is tolerant of rubbing in L0 2. For the LH2-cooledbearing, the inner ring material will be AISI 9310, which has better fracture toughness than the AISI 440Cit replaces. The AISI 9310 also provides increased stress corrosion cracking resistance over AISI 440C. TheL0 2-cooled bearing must use AISI 440C instead of AISI 9310 for the inner ring due to L0 2 compatibilityconcerns. Promoted combustion testing in the SSME-ATD program has shown that AISI 9310 is not suitable for
150
use in a L0 2 environment. Concerns of AISI 440C are discussed in the risk assessment section. The materialsselected for the ball bearings are listed in Figure 117.
Bearing temperature control is vital to success because it assures stable operation and prevents localizedsurface distress at the ball to race contact points. To assure this temperature control and establish an adequatecoolant flow rate, bearing heat generation must be known. Analytical heat generation techniques have beenverified during the SSME-ATD program. Figure 118 shows the predicted fluid temperature rise across thebearings at various L.02 coolant flow rates. Note the selected flowrate of 2.0 pps in relation to the knee ofthe curve. This clearly defines the safe operating area. Similar analysis on the LH2-cooled ball bearing wasconducted and a coolant flowrate requirement of 0.1 pps LH2 was established. Detailed thermal analysis isplanned during the final design to show load and flow margins and transient behavior of the bearings. Thethermal modelling is based on a NASA approach using the SINDA Heat Transfer Code.
A design summary of the L0 2 turbopump ball bearings is provided in Figure 119. Experience guidelinesare also provided to show design margins. Final verification of the ball bearing design will come under anIR&D program. A rig test will be performed simulating AETB operating conditions to demonstrate bearingdurability. In the test rig, both the L0 2-cooled and LH2 bearings are tested along with the interpropellant seal.The rig is shown in Figure 120.
b. Roller Bearings
A detailed trade study was performed to optimize the rolling element size and geometry for the AETBoperating conditions. Figure 121 shows the effects of radial load and negative IRC on contact stress andminimum roller preload for the nominal geometry of the selected configuration. The XLR129 and SSME-ATDtest data were used to establish a safe minimum roller preload for the AETB. Specifically, the minimum rollerpreload was selected based on a ratio of roller energy of the proven designs to the AETB condition. From thisapproach, a minimum roller preload of 220 pounds was selected to ensure adequate roller guidance. For the worstcase design radial load of 163 pounds, the minimum roller preload is greater than 270 pounds and the maximumcontact stress is 358 ksi. Although the contact stress is above the previous test experience of 200 ksi, the AETBlife goal of five hours is far less than the 18-hour life demonstrated at 300 ksi. The negative internal clearancecould be reduced to decrease the contact stress but this would reduce the minimum roller preload margin.
Material selection for the roller bearing was based on the SSME-ATD program. AISI 9310, the race material,has excellent fracture toughness compared to through hardened steels. Figure 122 shows the superior fracturetoughness of AISI 9310 compared to AISI 440C. Arnalon was chosen for the cage material, again based onprevious roller bearing experience. Armalon is a glass fabric laminate filled with Teflon. The glass fabricprovides the structural integrity, while the Teflon provides the lubrication.
Roller bearing coolant flow requirements were selected using the same methodology as the ball bearings.For the design point in the hydrogen turbopump, a coolant flow requirement of 0.2 pps was selected. For theoxygen turbopump, a coolant flow requirement of 0.1 pps was selected. Cooling curves for both turbopumproller bearings are provided in Figure 123.
A summary of key roller bearing design parameters is provided in Figure 124. Final verification of the rollerbearing design will be provided under an IR&D rig test program. A cross section schematic of the test rig isprovided in Figure 125. The test rig will simulate turbopump operating conditions and verify the bearing life.Three bearings are tested simultaneously; the center bearing reacts 100 percent of the applied radial load whilethe other two bearings react 50 percent of the load. Axial thrust balance control is provided by a thrust piston.
151
4. Bearing Methodology and Verification
Negative internal radial clearance provides the restraining force to provide roller stability. The negativeclearance produces additional load between the roller and the races, which increases contact stress. For theAETB roller bearing design, the small size and high speed requirements necessitate heavy internal preload. Theresultant contact stress can be as high as 359 ksi or approximately 20 percent above SSME-ATD experience.Roller stability margin also needs to be addressed since the required roller preload was determined empirically. Adynamic model is currently being modified under a NASA-MSFC contract to analyze negative internal clearanceroller bearings. This model can be used to verify roller stability. The ultimate verification will be providedunder the IR&D rig testing.
High assembly hoop stresses in cryogenic bearing inner rings can cause stress corrosion cracking (SCC)failures. AISI 440C has excellent general corrosion resistance but is susceptible to stress corrosion cracking.Figure 126 shows the contributors to SCC and steps taken to minimize the risk. Testing at P&W's MaterialsEngineering and Technology Laboratory has compared an AETB (RL 10) AISI 440C inner ring to early ATD AISI440C inner rings. Although this test is still underway, the data presented below shows a substantial differencein AISI 440C processing.
Stress Tune to Failure
AETB 440C 35 ksi 255 days (still testing, no failures)Non-Optimized 440C 35 ksi 19 days
AETB 440C 50 ksi 255 days, still testing, no failures)Non-Optimized 440C 50 ksi <1 day
By maintaining hoop stresses below 25 ksi, the SCC risk is very small.
Figure 127 shows various computer models used in the design of cryogenic rolling element bearings.Methodology verification of some of these models has been carried out under the SSME-ATD program. For theAETB, all of the models will be anchored or verified under the IR&D rig test program.
152
I I .Iix x
I- x
E m il00
o4
00.
o- 0
153~
0. 1-
CLL
Cl) M M f
II cc1, Z
G)WI W~ k3-N
00 0
154.
=04
coE~ X)(* 0 0 0
Co 0 to~ V)g0r 0 ~>>
Cdj 0 .- CL 00~L
0 ~ Of >~ 00
0 _ 0 0 ox co c u).a
LUN
0 Z
cc _30
O E .~7 0
0 cc
A. -5
cr 0
00
0
Cc EI - 00
01
,-0,____cc__ cz
0 =Q
155
E
2- 2U.0a o
I- 6
0 '0
w woc
2-6x '0 .o)o.)
000.
co 0 0 cam0O
4iE 4v~
a 0) co
cc 0 0 jc <_ 0 2c
0 C5 '
01
(3 5c- O .2 p( )-E 0
S. o . 0 0.
_~ -I,
0 32
156
I0
LL. IL U.
coco
L()
O
00O _ _) CoQ f
ZZC
3: & 0 0 0Cox c_ .. ' - 0L
X Au E ALox c"~
E0 ww M-o Lqo0 coig CO - 0*~
oL W Imo0O
E cogou-D %o
(0 -- I
E LLLL w xV .1 w 01 X~ x - q 0 -..
m V-t . o__ - ,L
-'U Co
0-CJU to0.
LL.
157
0
0 z c
o 00a
z C0
0 a0
_ Cu U
z 0o X
o 0 4-f'(
0 LE 0 0
0 Ld 0.
E 0_w 0o E
C~l (MCEI
0 co 2 oCO) cL
E x 0 o Co 0
o CO o c p' 9 0u
0) 0) 0)I .0)wJ 0
S w
CfCo
158
z z
x x
*~
Ow0co '-0 C 0
0 U,_X -C 0X N o XC
x U)) I Scri Ero.O QDCO0) %1-0
co 0%.1 00 *L. 00
0 dCU LO n 7 2 E 3:L6C Iq
)~ EE (6 L6c
E N co cE 6xw o 0 L
x xx LI. EEXb _ocL..CDv _n _
LU a " IJJS."- 2= CIO
ww www
159
it X16M00
0) SW
i5-
M 22:
LO
CL co
0 0C) 0
Coc
0160
0)
0.) CE
LEO
CMC6Co %_
x CToo 0
E QCJ
00
C /
coc
COl0 CD CC0 C0
c 0 CL _4
x CO cn m
CO Lo-DC
4-a C~ LL01 a) X a) (DU
V Lo L.I 0 )
V Orn161
oOE< E
00
CO 0 a:
0SE a
w Vo 00
0
00
C
EEE n jC 0a, -. &co xLCu u
4-4 t
162
00
*- ri __ _
x N .'t~ 0 -j
CL 0)
0)
V~ 9r- 0 Ia
LL U)_ _ _ _ _ _ _ _ _CE
CL g
16
Vd
cc
CLV d
0. C; 0 -- A_
cc
-o C4z-(0
0c S E0 C
-U x a.Ea)1 0CLG .
-J g co * .9M = Cs) CI 0 '4.
4)164
0)
x
0)
01)
.j.
165
IAI
.Eo
_ , IIe IC
0 <M
0 0 v -0 "
cca:Cc
._ 0*C , LII o °
0 0
Ct
*0 'a1
0)
as6
cU166
LLL
o ) aaQ
0 LL 4
Cn-
0cc
000 167
C.Cu
Eg2 zN
o..
U- CL
CL .
=*~~jLLL-o 0
0 '5m 00
_. 3..
0)
168
Cc
C, D
CL -9C? sq
CL Cy C CS .0
x V08U. 0= 1 ) CV
000) Cls c C
0 Mi
E o.
169(
IM M
(Adc
ca
U-0 MLo co 0
0 go
Cu Wco 0R6 0 .CL L)o oc
IOO 2 I)
~C.. 0) 0 .-C 0 C~ 0
00
CC
cv 00
cr. Ev&0 ~Q00 E 0
a or O C
~zEC) 00I_i E01 S0
-0 E0 8Co 1 Co0~
Xn 0 4)0 00a
0L 0
E~o v a. (0 A
VI-C L.C C.)
000
oCoa 1-04 Iw 0 171
00
U E -
0M C
cS 0
(I) C&-C 0)
05~3:
C cc,
00
U,.SE C
C C172
H. Combustion System
The combustion system consists of an injector with igniter, combustion chamber, and a conical nozzleextension as shown in Figure 128. The dual-orifice injector and milled channel liner combustion chamber arebased on an existing design completed and detailed under a P&W Space Engine Component Technology Program.Although contract work on the components in preliminary design included only the detailed layout of the exhaustnozzle, the design of all the hot section components is described in the following sections.
1. Injector/Igniter Assembly
The AETB igniter uses the same design approach used in the P&W RLIO engine, SSME-ATD hot gassystem prebumers, and the Advanced Launch System (ALS) Technology ignition system. Figure 129 shows theH2-0 2 torch igniter design that will be employed.
The torch igniter consist of a Haynes 230 mount flange housing with a oxygen free high conductivity (OFHC)copper combustion liner and a Haynes 230 structural jacket. The ignition chamber diameter is constricted from0.500 inch in the chamber to 0.220 inch at the exit to produce adequate igniter chamber pressure for ignition ataltitude. The liner operational life is predicted to be adequate with GH2 cooling. The same design features areincorporated in the SSME-ATD igniter which has over 1000 seconds of operation with no problems to date.
Various ports on the mount flange allow installation of the spark plug, instrumentation, and inlet lines. Theigniter is mounted through the center of the injector using stepped studs.
The injector assembly, Figure 130, will be manufactured from ferrite controlled 347 stainless steel (347SST). It consists primarily of an injector housing with a fuel manifold welded on the outside. In the center of thehousing. various cavities are machined to create the internal oxidizer injection manifolds. Sixty-five dual-orificeelements are uniforn-ly spaced in a circular pattern with allowance in the center for the torch igniter. Ferritecontrolled 347 SST was chosen for its ease of machining, weldability, brazeability and ductility. The ferritecontrol helps reduce the risk of post-weld cracks in applications where no filler metal is added to the weld.
A separation plate is brazed in the top of the assembly to separate the primary and secondary oxygenplenums. A welded dome closes the secondary plenum and provides for installation of the igniter. The fuelplenum is created with a porous faceplate welded to the housing and brazed to individual fuel sleeves. Theporous plate provides transpiration cooling of the injector face.
The core of the injector consists of the 65 L0 2 elements and fuel sleeves The elements, Figure 131, are ofthe dual-orifice tangential entry type and are brazed into the top of the housing. Primary LO, enters each elementthrough three holes equally spaced, and secondary oxygen enters through three equally spaced axial slots. Onthe bottom of the housing are nozzles machined from the housing forging prior to the sleeves being brazed tothe housing. The annulus created by the nozzle OD and sleeve ID meter the fuel into the combustion chamber.
The injector has been analyzed for acceptable structural integrity at the design point by both conventionalcalculations and a 2D boundary model. BEASY. The model included both thermal gradients and pressure loadsfor the injector. Figure 132 summaries the factor of safety for the injector.
A chugging model was created and run at the 5 percent, 10 percent, and 20 percent power levels. Theanalysis predicts no chugging will occur at these points, as shown in Figure 133. The model represents thepropellant feed system in terms of inductance-resistance-capacitance (L-R-C) theory. High frequency combustionstability analyses were also conducted and adequate stability margin is predicted.
173
2. Combustion Chamber Assembly
The AETB combustion chamber, Figure 134, has a contraction ratio of 3:1 and an expansion ratio of 2:1.The chainber consists of a NASA-Z copper alloy liner with 120 milled coolant channels on the outside surface.Liner cooling channels are a constant 0.040-inch wide with a maximum height-to-width ratio of 5:1. Wallthickness between hyd.-cen coolant and the hot combustion wall is a constant 0.030-inch thick. The passageheight is set to allow a maximum wall temperature of 1460 R without exceeding the allowable budgeted cyclepressure drop. At the normal operating point, the maximum wall temperature is 1355 R. Maximum heat fluxat the operating point is 51.7 Btu/in.2-second, occurring 0.50 inch upstream of the throat. This configurationprovides a minimum predicted life of 200 cycles. No coolant two-phase flow instabilities are predicted in theliner or nozzle coolant circuit, since coolant pressure remains above the critical pressure of hydrogen over nearlyall the thrust range. At 1,000 pounds thrust (20:1 turndown), pressure will drop below critical pressure but notbefore the hydrogen temperature is well above critical temperature.
The liner has an electroformed copper outer jacket that closes out the milled coolant channels and providesstructural support for the chamber. Coolant manifolds are welded to each end of the chamber. Both manifolds,consisting of a ferrite controlled 347 SST material, are welded forming an internal primary distribution manifoldwith crossover ducts to a minor manifold, which is created when the jacket and manifold are joined.
The inlet manifold of the chamber interfaces with the nozzle extension, and the outlet manifold interfaceswith the injector. Both of these joints incorporate a pilot snap fit. The snap is used to control radial movementduring operation and to center the mating assemblies. The injector face extends into the chamber 0.7 inch toprotect the uncooled portion of the liner.
The liner has been analyzed for acceptable structural integrity and life. The inlet and outlet manifolds havebeen analyzed by a 2D NASTRAN finite element model. A summary of these analyses with the calculated factorsof safety is shown in Figure 135. The proof pressure condition was also examined and found to have acceptablemargin when pressurizing the combustion chamber and cooling passages simultaneously. Coolant pressure andcombustion pressure will be applied simultaneously during proof pressure tests with the throat area sealed off.The divergent section of the chamber will be exposed to ambient pressure.
3. Exhaust Nozzle Assembly
The conical nozzle extension consists of 160 coolant tubes brazed into a structural jacket containing the inletand exit manifolds. The nozzle cross section is shown in Figure 136. The base mater,.a for the assembly details,Haynes 188, was chosen for its ductility, weldability, and good strength in hot hydrogen. It will also facilitatebrazing during nozzle assembly, provide high-temperature capability, and meet heat transfer requirements.
The 160 coolant tubes are brazed into the inlet and exit manifold with a structural jacket joining the two.Each coolant tube is joined to the inlet and exit manifold by a braze joint. On the inlet end the tube will be hookedto fit into the inlet manifold. The tube exit will be an offset square socket joint that will fit into a machinedannulus ring. Various combinations of tube attachments were examined and the current tube configuration wasselected based on cooling and fabrication considerations. The uncooled portion of the nozzle is protected bybeing recessed into the inlet manifold of the combustion chamber.
The inlet manifold also contains one end of a spring arm that is used for controlling the radial thermal growthcaused by the 600'F temperature differential between the cold chamber inlet and hot nozzle inlet. The springarm between the two manifolds is designed to accommodate the relative thermal deflections of the manifoldswhile eliminating seal sliding and maintaining acceptable structural integrity.
174
A preliminary structural analysis of the spring arm was conducted by first examining the axisymmetricloads, then expanding the analysis to include asymmetric loading caused by transient pressure loads, weight, andinterface loads. As shown in Figure 137, a safety factor of 1.24 is indicated. Buckling analysis was completedby evaluating shear forces on the spring arm from axial, transverse, bending, and torsion loads, resulting in abuckling factor greater than 10.
175
C\j
IU)
I xx-i wwI a
NNxNL LU
00
Q)'
OCL)
Ccu W
U U: C:
176
CL
LL-
177C
ii a
178
Al 1 3I2 .
LL
179
LLI-
___ ___ C'
--
V)V
C/)C
118
25 -AETB 5/. Power
20 -Assumptions:Pc -64.9 psiaOIF -3.54
15 -Tc - 5035 R
Open Loop 20 zT -50R C*- 8197 tt/secMagnitude 10 (oia) All 02 Through PrimaryRatio - dB Nominal -,g = 0.62 msec
Nominal 7ign 0.16 msec5 *0 % 5 HzNominal T mix -0.40 msec
I Unstable A PLo 2 =9 psia
AETB 10% Power25 - Assumptions:
P,-122 psia20 - OIF -4.5 Operating
Tc -5968 R Curve Ti2- 759 RC*- 7363 if/sec
15 - Nominal Tg - 0.65 msecOpen Loop Nominal Ti - 0.16 msecMagnitude 10 - Nominal 7 mt -0.40 msecRatio - dB A PLo2 -23.5 psi (Primary)
5fUnstable ~
0 Region I
25 AETB 20% PowerAssumptions: 50 Hz
20 Pc - 237.8 psiaO/F -6.1 I
15 Tc -6130 R 100 HzC*- 7580 ft/sec
-~ 700 ROpen Loop 32% of 02 Through Primary (NmnlMagnitude 10 Nornalrg- 0.62 msec
5 Nominal 7 mix - 0.40 msec
*Unstable APo-83sa200 Hz0 RegionTH-20R
-40 0 40 80 120 160 200 240 280Phase Margin - Degrees
Figure 133. L-R-C Stability Curves at Power Levels of 5%, 10%, and 20%
181
*0J
z
X+T
1'1
182
LL.
d (U
tI CwwUQ
w<<
cc 00 00C>00 0N) qC.q N 4 4 IV'
CS0 CN" A A C A A A A * -
.0C.) -% l)V L
00
zi4W
'C I183
'46
184
cCD
- N.
( F
c-
L.0
LrL.
43-.
185
I. Hydrogen Mixer
In the split expander cycle, the hydrogen mixer, shown in Figure 138, mixes the warm hydrogen fromthe turbines with the cold hydrogen from the first-stage fuel pump discharge. The combined flow then entersthe main combustor chamber injector fuel manifold. Good mixing of these streams is critical to maintainingstable combustion and unitorm flow through the individual fuel elements. At the design point, the flow intothe mixer is split 60/40 between the hot and cold lines. The cold hydrogen flow is controlled by means ofthe fuel jacket bypass valve (FJBV). The percent of cold flow bypassed is lower at lower throttle conditions.For instance, at 20 percent thrust, the FJBV is completely closed so all the flow into the mixer is the warmhydrogen from the turbines. When bypassing cold flow to the mixer, the mixer must effectively mix the hotand cold hydrogen, yet minimize system pressure loss. To achieve the required mixing performance, the AETBwill use an in-line mixer similar in concept to the one used by the Space Shuttle Main Engine (SSME) system.The AETB design will use a single tube for the high velocity flow. The Rocketdyne SSME mixer uses seventubes clustered inside the mixing line.
The mixer works on the same principle as a jet pump, i.e., a high velocity stream imparts momentum to alower velocity stream. The momentum transfer creates turbulence which promotes mixing of the two streams.The hot hydrogen from the turbine discharge forms the high-velocity stream while the cold hydrogen from thepump is the low-velocity stream. Using the established design procedure for jet pumps, the minimum mixinglength for the maximum jet pump efficiency was calculated to be 10 inches at worst case operating conditions. Ifthe AETB mixer had used seven tubes, like the SSME mixer, the required length would be reduced to five inches.The actual mixing length will be 37 inches. There is a relatively high momentum ratio of 28.3 between streams.This compares to the SSME momentum ratio of 1.1. The area ratio of the AETB mixer is 2.5 compared to 2.2for the SSME. Due to the extended mixing length and high momentum ratio, the mixer design is conservativeand will provide uniform flow to the injector.
The mixer design incorporates the following features:
• The two-piece construction nearly eliminates the thermal stress problems that were evident with anearlier welded, one-piece design.
* The hot inflow is a separate piece of hardware, which provides the versatility of changing mixergeometry to evaluate alternative mixer designs.
" The parts are machined entirely from 347 stainless steel using only conventional machining techniques.
* Repairability is built into the design by allowing enough radial clearance around all tapped holes forthreaded insert repairs.
* A conservative LCF exceeds 3000 thermal cycles. v
* The cantilevered tube natural frequency is 3300 Hz. This is well below the vibration mode of eitherpump rotor and well above the low energy vortex shedding frequency of 66 Hz.
186
ci >/x
1 87
J. Control System
1. Requirements
The design of the AETB control system follows a flowdown of requirements from the general to the specific.The system configuration dictates the control mode selection, thrust, and mixture ratio control, which drivesthe requirements for control valves and sensors. The control mode, sensor and valve requirements, along withhazard analysis and safety input, are used to generate the control logic which will implement the control lawsand failure accommodation methods. This approach leads to a control system that will meet all system, hardwareand safety requirements. The test bed will be an oxygen/hydrogen split expander cycle of 20,000 lbf thrust at1200 psia chamber pressure. The test bed will be throttleable to a 20:1 ratio with mixture ratio control of 5 to7 and operation at a mixture ratio of 12.0. In addition, the test bed will be capable of tank head and pumpedidle operation, and will be able to be operated as a full expander engine.
a. System Description
A simplified schematic, Figure 139, is used as a reference to explain the control concepts. It shows thefunctional arrangement of the valves, plumbing, and turbopumps on the test bed. The igniter and purge valveswere omitted for clarity. The turbopumps are shown as separate pump and turbine sections and the nozzle andchamber heat exchangers are shown as separate from the nozzle and chamber.
On the fuel side of the test bed, hydrogen passes through the Engine Fuel Inlet Valve (EFIV) and into theprimary fuel pump. At the primary fuel pump discharge some hydrogen is bypassed into the Fuel Jacket BypassValve (FJBV) and, if necessary, some is recirculated through the Fuel Pump Recirculation Valve (FPRV) backto the primary pump inlet. The remainder of the flow travels through the two-stage secondary fuel pump. Atthe pump discharge is the Fuel Cooldown Valve (FCDV) which is used during shutdown and during cooldown.The hydrogen flow passes from pump discharge through the nozzle and chamber heat exchangers where itcools them and picks up energy to power the turbines. For high power full expander operation, some flow isbypassed around the chamber and nozzle heat exchangers through the Chamber Coolant Bypass Valve (CCBV).The hot hydrogen from the heat exchangers expands first through the oxidizer pump turbine and next throughthe two fuel pump turbines. The Main Turbine Bypass Valve (MTBV) and Fuel Turbine Bypass Valve (FTBV)are used for thrust control and high mixture ratio operation, respectively. Turbine bypass flow and turbinedischarge flow are combined with FJBV discharge flow in the hydrogen mixer, passed through the Fuel ShutoffValve (FSOV) into the injector manifold and into the chamber where it is mixed with oxygen, combusted, andexpanded through the exhaust nozzle.
The oxidizer side of the test bed is significantly simpler than the fuel side. Liquid oxygen passes throughthe Engine Oxidizer Inlet Valve (EOIV) and into the oxidizer pump. At the pump discharge some flow isrecirculated back to pump inlet through the Oxidizer Pump Recirculation Valve (OPRV), if necessary. Duringpump cooldown and test bed shutdown, the Oxidizer Cooldown Valve (OCDV) is opened to provide a flow pathwhen the injector valves are closed. Flow from the pump discharge then passes through the Primary OxidizerShutoff Valve (POSV) and the Secondary Oxidizer Control Valve (SOCV), which is used for throttling and tocontrol test bed mixture ratio. From there it passes into the primary and secondary injector manifolds and intothe chamber where it is mixed with hydrogen for combustion.
b. Controller Functions
The controller will perform the following functions:
Perform pre-start diagnostic system checks, purge the lines of moisture and air with inert gas, and chillthe pumps to where cavitation is not a problem.
188
* Start the test bed to the requested thrust level and mixture ratio setting. A goal of <5 seconds has beenchosen for the start time. Model simulations indicate that this is achievable.
For mainstage operation, regulate test bed power setting (throttling) and mixture ratio (0/F). Althoughno response requirements exist for the test bed, the bandwidth of the thrust control loop will be madeas wide as possible within hardware limitations.
At the termination of a test or if the control or facility declares an abort situation, the test bed willbe shutdown in a safe manner.
* Subsequent to shutdown, purge the test bed of fuel, oxidizer, and combustion products to make thetest bed safe.
During all test bed operation, monitor safety parameters and take appropriate action when safety limitsare violated. In most cases this will be a test bed shutdown.
c. Thrust Control
Thrust control will be accomplished by closed loop control of chamber pressure (Pc) as shown in Figure140. Chamber pressure will be sensed and the MTBV will be modulated by the control to achieve the desiredthrust, Figure 139. Feedback P, with lead-lag compensation for overshoot minimization will be compared toa requested reference P, and the resulting Pc error fed through a proportional plus integral controller. Theproportional gain will give fast response to request changes and the integral gain will give good steady-stateaccuracy. The lead-lag compensator time constants and the proportional and integral gains will be determinedthrough control studies during the final design.
Chamber pressure will be sensed by the controller to control test bed thrust level. A sensor accuracy ofthree percent of point has been specified. To achieve this accuracy over the wide throttling range and to sensechamber light during engine start without excessive sensor complexity and cost, multirange transducers will beused, Figure 141. Low (<150 psia), medium (<500 psia), and high (<1500 psia) range pressure sensors will beused to cover the entire pressure operating range. The control will gradually phase out the appropriate rangesensor in the transition regions to avoid discontinuities. Each sensor will be accurate to one percent of full scaleand, as seen in Figure 141, will provide three percent of point accuracy over its range of operation.
Since the chamber pressure sensor performs a critical control function, redundant sensing capabilities willbe provided. The high range will have dual pressure sensors, and redundancy in the medium and low ranges willbe accomplished by using the sensor in the next higher range as a backup. Some thrust control accuracy in thebackup mode will be lost in the medium and low ranges, but redundant pressure sensing will be accomplishedwith the addition of only one high-range sensor.
The total achievable thrust control accuracy is set not only by Pc sensor accuracy, but also by valve positioningaccuracy as well. The MTBV will be positioned by the control until P, is within the sensor inaccuracy so itscontribution to thrust inaccuracy can be neglected. However, the FJBV and the SOCV positioning inaccuracieswill contribute to thrust errors. The allowable valve position error is 1.5 percent of stroke. The steady-statemodel was run to determine the sensitivity of P, to SOCV and FJBV areas. The total root sum square thrusterror for valve position inaccuracy and Pc sensor inaccuracy is shown in Figure 142. The total P, error is shownfor ball-type and linear valves since a determination of the valve type has not been made. Figure 142 is validfor the nominal 6.0 mixture ratio point. Thrust inaccuracy at other mixture ratio points will be determined aspart of final design.
189
d. Mixture Ratio Control
The options for mixture ratio control are to schedule O/F open loop by positioning the SOCV as a functionof requested chamber pressure, or use flowmeters to directly measure O/F and position the SOCV so that O/Ffeedback equals the O/F request.
Open-loop O/F control was chosen because it represents the simplest design, avoids possible thrust and O/Floop coupling problems, and meets the requirements of the test bed. A sensitivity study on O/F error similar tothat performed for P, error was conducted. As seen in Figure 143, the estimated O/F absolute error will be keptwithin 0.12 for power settings above 15 percent. This study was performed at the 6.0 mixture ratio point andwill be repeated at other O/F settings. At this time the effect of open-loop mixture ratio control during throttletransients has not been determined. However, rate limit logic and other control compensation techniques can beimplemented should transient inaccuracy prove unacceptable.
F'-ure 144 shows the operating range of other test bed parameters due to valve positioning inaccuracy. Asshown, no system limits will be violated with the chosen control modes.
A schematic of the overall control mode is shown in Figure 145. As discussed above, a thrust request willbe transformed into a rate-limited P, request which will be fed through a proportional plus integral controllerwith lead-lag compensated P, feedback to control thrust closed loop. Mixture ratio control will be open loopwith the SOCV and FJBV positioned from a bivariate table lookup based on requested Pc and rate-limited O/Frequest. The FFBV will be positioned as a function of O/F request to achieve the high mixture ratio point. Therecirculation valves will be opened, if necessary, as a function of thrust and O/F setting.
e. Component and System Protection
The control must protect the test bed pumps, valves, thrust chamber, and other components from damageand must operate the test bed in a safe manner. To accomplish this, a large portion of the control function isdedicated to execution of fault accommodation and safety monitoring logic.
Prior to engine start, the controller will perform diagnostic self-checks to determine its own state of healthand will continuously monitor itself during engine runs. Sensor validity checks will be performed and eithersensor redundancy will be provided (for critical parameters), or test bed shutdown will be initiated when out-of-range signals are detected. Limited capability will be present for sensor in-range failures and verificationwith other parameters will be performed where possible and safe to do so. Valve actuator simulations willbe incorporated into the logic to check for slow or stuck modulating valves, and pre-start rate checks will beperformed on the discrete valves.
Throughout operation, the control will monitor engine parameters for violation of safety limits, and willshut down the test bed if any violations occur. Monitored parameters include fuel and oxidizer pump metaltemperatures during prestart, pump inlet pressures, pump speeds, and pump vibration levels. Pump bearingcooling flow temperatures will also be monitored for excessive heat generation, and interpropellant seal healthwill be determined by helium inlet pressure and He/H 2 and He/0 2 discharge pressures. Oxidizer turbine inlettemperature will be monitored, and in addition to its control function, sensed main chamber pressure will bemonitored for limit violation and the inability of the test bed to achieve requested chamber pressure.
Although not specifically monitored by the control, protection from other system anomalies will be inherentlyafforded by the design of the control logic. Combustion instability will be avoided by design of the valve schedulesto maintain proper primary to secondary L0 2 injector flow split and primary L0 2 injector delta pressure. Becausethe primary fuel pump is designed for twice as much flow as the secondary fuel pump, the control must also
190
properly position valves to avoid secondary pump choke and primary pump stall. In addition, during test bedshutdown, FJBV positioning is critical to avoid reverse hydrogen flow from the mixer to the secondary pumpinlet when the FSOV is shut, Figure 139.
f. Software Development
The control logic software will be developed using the standard industry 'waterfall' process smtured toDOD-STD-2167A. The 2167A process starts with an analysis of the system requirements. Next, design of thesystem proceeds along with software requirements definition. Then, software design and production of sourcecode begins, and finally, test and certification of the software is completed. At this time, analysis of requirementshas been accomplished, and a Control System Requirements Document (CSRD) and Software RequirementsSpecification (SRS) have been published.
The software development process is concurrently engineered with inputs from the customer and variousP&W functional groups such as Propulsion Systems Analysis, Controls Engineering, and Safety Engineering.The development proceeds in parallel with hardware development, ending in system integration testing andoperational testing and evaluation.
Much of the AETB software will incorporate already proven codes from the National Aero-Space Plane(NASP) program such as system executive, I/O software, and monitor software. Unique to the AETB will bethe control law software and bench test simulation software. The control law software will be comprised ofthe program executive which handles process control, initialization software, monitor support, processor healthsoftware, and the actual control laws themselves. The control laws will consist of engineering unit conversions,thrust and mixture ratio control loops, failure detection and accommodation logic, and engine safety monitoring.It is anticipated that the majority of the control law software will be dedicated to the tasks of failure detectionand safety monitoring.
g. Valve Requirements
The valve configuration is selected to provide control over thrust and mixture ratio in the split expander,full expander, and tank head idle modes. Additionally, valves are selected to start, shutdown, and inert the testbed prior to and after operation.
The design requirements of the control valves are shown in Table 14. Fuel and oxidizer valves were sizedfor the 125 percent thrust point with the exception of the OCDV and FCDV which were sized for shutdown flow,the FTBV which is sized for 12.0 O/F operation, and the MTBV which is sized for tank head idle operation.The purge valves have been oversized for commonality and will be provided with downstream orifices for flowcontrol. Position feedback through LVDT's will be provided for fully modulating valves, and position of thesolenoid valves will be indicated at the critical position. Fully modulating valve effector loop bandwidths are setto give adequate response so as not to impact upon major loop (thrust control) response. Model studies indicatethe system response to MTBV area has a bandwidth of 0.7 Hz. The MTBV effector bandwidth is set at 5 Hzwhich is sufficiently responsive. The failsafe/depower position of all valves has been chosen at the shutdown orlow power setting. Valve slew rates were driven by the abort shutdown requirement.
h. Sensor Requirements
The control sensors have been selected as a result of cycle and throttling studies to sense the operatingconditions to meet the performance, operability, and safety requirements of the test bed. Sensor designrequirements are given in Tables 15 and 16. A summary explanation of the requirements is presented inTables 17 and 18. In these tables, signal loss refers to control action if the sensor input has been determined to
191
be invalid. The redline/permissive column refers to the action that the control takes if the sensor input indicates
a violation of redline or permissive limits.
2. Electronic Controller
a. Objectives
The brassboard controller design objectives arc focused on efficiently supporting test bed operation. Duringtesting at NASA, the brassboard will be used to support test bed demonstrations of the engine cycle andinvestigations of technologies such as advanced health monitoring and electromechanical actuators. Supportof the objectives requires a flexible and expandable design. During preliminary design, several requirementswere given to hardware and software designers to meet the test objectives. These requirements included thecapability for system growth, adaptability to Input/Output (1/0) modifications or additions, and the ability toeasily reconfigure the test setup.
Control system growth will be required to support control changes and investigation of advanced healthmonitoring sensors. Currently an open loop start schedule with a steady-state closed-loop thrust control andopen-loop control of mixture ratio are planned. Changes to the control cycle that increase throughput or I/Ochanges should be accommodated by the controller.
The controller must be adaptable to 11O modifications. Technologies such as electromechanical actuators thatrequire testing with the brassboard will require that current 1/0 hardware and software be changed to supportthe tests. The changes are best accomplished with modular designs of hardware and software components. Inthe event that processing and 1/0 changes are significant enough to warrant a growth of two or three times inprocessing or 1/0 capacity, the brassboard must have the capability to accept additional channels or functionalgroups. The level of processing and 1/0 required for the current design can be accomplished with one 19-inchrack of computer hardware in one functional group.
b. LO Requirements
Input/Output requirements determined during preliminary design are shown in Table 19. The brassboarddesign accounts for all required 1/0 and provides spare channel capability for each of the 1/0 types. Since the1/0 circuit boards are modular in design, further expansion beyond the spare capability is possible by addingadditional circuit boards. Five spare slots remain in the baseline system to accommodate additional processor,sensor and effector interfaces.
c. Test Stand Interfaces
Test stand interfaces were designed with safety and test time optimization as the main objectives. Safetyconsiderations determined that abort signals from the brassboard and the facility should haN e the capability toterminate a test if an unsafe condition is detected. A discrete signal interface will be used to implement abortsystem communication. Upon assertion of the abort signal, both the facility and the brassboard will take action toshut down the test bed with minimal delay. As a precaution against AC power loss during test, an uninterruptablepower source will be required from the stand.
Several features were added to the brassboard system design to optimi7e the time for test bed operation.Propellants, stand power, test bed hardware and test personnel time are all costs associated with operation ofthe test bed. The brassboard system design has provisions to perform preprogrammed test sequences basedon input from a time code generator and the Monitor System user interface. Predetermined test programs canbe loaded into the monitor system to accomplish accurate and repeatable tests. The sequencer function allowsverification of test setup parameters prior to actual testing. Data transfer between the brassboard system and thestand computers can be accomplished with an Ethernet Local Area Network using the DECnet protocol.
192
d. System Description
The brassboard system consists of three major components, the Brassboard, the Monitor System, and aBrassboard Test System (BTS) as shown in Figure 146. The configuration shown depicts the system in theverification configuration, where software and hardware testing is performed prior to engine test. During enginetest, the BTS is replaced by the actual test bed.
The brassboard design is based on a functional group concept. A functional group consists of a 20-slotVME card cage, processor board and a full complement of 1/0 hardware. Up to four functional groups can beused in one system to distribute processing and 1/O functions as shown in Figure 147. The brassboard contains19-inch wide rack mounted hardware with two separate card cages for circuit boards. Each cage holds up to20 boards and is an industry standard VME design. As shown in Figure 148. five spare slots are provided forexpansion capability. The AETB brassboard contains cards in only one of the card cages, the other is providedfor expansion capability. Other associated hardware such as power supplies, fans and connector panels are alsoprovided. One processor board and one set of 1/0 hardware as previously described are included. In addition, aglobal bus board is used to provide timing functions and a link between functional groups.
The monitor system is a MicroVAX-based computer system with a 1553 interface to communicate with thebrassboard. The monitor's two main functions are to provide a vehicle interface simulation and a user interfacefor brassboard operation as shown in Figure 149. The vehicle interface simulation will send commands tothe brassboard that would normally occur during an actual mission. Commands sent to the brassboard includeprestart conditioning, start, throttling and shutdown commands. In return, the monitor will receive feedback fromcommanded parameters and engine sensor data. The data can be stored in the monitor for later graphic analysis.All communication initiated from the vehicle interface simulation can be automated by building command filesand storing event sequences prior to test bed operation. The user interface functions of the monitor include storageand downloading of control programs, examination and alteration of control constants and memory, real-timedisplay and bar graphs of engine parameters, parameter versus time plots and other general computer functions.
The BTS provides a real-time simulation of the test bed for hardware and software verification. A completeset of 11O hardware to mirror the brassboard I/O set is provided to simulate sensors and actuators. The sensorand actuator simulations are interfaced with the engine simulation to provide a complete test bed simulation.To the brassboard, the BTS appears and acts like the actual test bed. The BTS will be used extensively duringverification at P&W prior to test bed operation. After the first complete set of test bed hardware is used foracceptance testing, the BTS can be used to verify logic changes prior to operation with the test bed.
e. Circuit Board Block Diagrams
The processor board, Figure 150, contains two processors which share the same bus. One processor isused for 1/0 and the other is used for control laws. A 1553 avionics data bus and VME 1/0 bus interfacesare also provided.
The global bus board, Figure 151, provides a link between optional functional groups. The board containsglobal RAM accessed by all functional groups for communication. System functions such as clocks, real-timeinterrupts and switch discretes are also implemented on this board.
The Linear Variable Differential Transducer (LVDT) board, Figure 152. provides signal conditioning for upto 16 channels. A dual-coil input for each sensor is input through multiplexers and full wave rectifiers to an AiDcircuit and the conversions are made available through the VME bus to the 1/0 processor.
The torque motor board. Figure 153, provides current drive for up to 12 channels. Current commands arereceived from the 1/0 processor through the VME bus, converted to a pulse width signal, filtered, and sent tothe current drivers. Each channel has wraparound current sensing to detect torque motor or cable faults.
193
The low level interface board, Figure 154, conditions thermocouples, RTD's and strain gauge pressuresignals. The board provides excitation or cold junction compensation for each signal and A/D conversion. Anew board will be designed to optimize its use and reduce the number of boards required. Proven circuit boarddesigns will be used in the new board design.
The frequency board. Figure 155. provides pump speed monitoring by measuring the frequency of speedsensor signals. The board is a modification of a commercially available design. The same type board can beconfigured to condition flowmeter signals if necessary.
The discrete interface boards, Figure 156, condition switch inputs and drive relay/solenoid outputs. Twoinput boards and two output boards are provided to meet 1/0 quantity requirements.
The analog input and output board, Figure 157, conditions high-level inputs and drives high-level outputs.The board will be used to measure a conditioned vibration signal from the turbopump sensor.
A processor throughput and memory usage study was performed for the brassboard processor board. Thestudy was based on measurements taken on the NASP controller which uses the same hardware. The control lawand I/O processing estimates are based on the I/O channels. As shown in Table 20, 62 percent of the availablethroughput and 41 percent of the available memory will be used. If throughput or memory requirements growbeyond the current hardware limit, additional modules can be added to share the control and diagnostic tasks.A processor upgrade will be evaluated during the final design phase. This would provide additional capabilityfor anticipated advanced sensors.
3. Valves and Actuators
a. Objectives
The valves and actuators use brassboard controller commands to control the flow of propellants and purgefluids for engine operation from pre-start conditioning through rated power to post-shutdown conditioning and alllevels in between. The valve and actuator configurations are being designed to meet the objectives of providingsafe, reliable, low-cost flow control and allow component replacement and flexibility to meet the varied test bedoperating conditions and configurations.
The valves have been separated into three categories:
I. There are five control valves that control flow of the following fluids:
• Fuel Jacket Bypass Valve, FJBV - cryogenic H2" Fuel Pump Recirculation Valve, FPRV - cryogenic H2
" Secondary Oxidizer Control Valve, SOCV - cryogenic 02" Main Turbine Bypass Valve, MTBV - warm gas H2" Fuel Turbine Bypass Valve, FTBV - warm gas Hi.
Each of the control valves is modulated by a hydraulic actuator to provide variable flow control as requestedby the controller.
2. The main shutoff valves are as follows:
" Engine Oxidizer Inlet Valve. EOIV" Engine Fuel Inlet Valve, EFIV
194
* Fuel Shutoff Valve, FSOV* Fuel Turbine Shutoff Valve, FTSV* Primary Oxidizer Shutoff Valve, POSV.
3. The ancillary shutoff valves are as follows:
• Oxidizer Pump Recirculation Valve, OPRV* Oxidizer Igniter Shutoff Valve, OISV* Fuel Igniter Shutoff Valve, FISV
* Fuel Pump Cooldown Valve, FCDV* Oxidizer Pump Cooldown Valve, OCDV* Solenoid Purge Valves, SPV, 7 Per Engine.
The main and ancillary shutoff valves are positioned either full open or full closed by pneumatic actuationas requested by the controller.
b. Control Valve Configurations
Each control valve assembly consists of a control valve element and an actuation element as shown inFigure 158. Based on supplier recommendations, the control valve element is either a common ball or poppetvalve design featuring reliable, dual dynamic seals with an interseal leakage drain. The valve is positioned bya hydraulic actuator, which is thermally insulated from the valve element.
The hydraulic actuators are off-the-shelf units which use Electro-lHydraulic Servo Valves (EHSV) for outputeffector interface with the brassboard and Linear Variable Differential Transformers (LVDT) for valve positioninput interface with the brassboard. The EHSV is a three-element device which converts the brassboard actuatorslew rate command from an electrical signal to a hydraulic flow rate which slews the actuator piston. The firstelement is a torque motor which uses the brassboard supplied current in the stator to move an armature whichpositions the first-stage servo device. The first-stage servo, which is the second element, ports high-pressurehydraulic fluid to position the second-stage spool valve, which is the third element. This second stage portseither supply or return pressure to each of the two control pressures to either extend or retract the actuatorpiston. A feedback spring connected to the second-stage spool valve nulls the first-stage servo device such thata nearly linear current versus second-stage control pressure flow rate is created. Designs of both the LVDT'sand EHSV's will be tailored for aerospace applications.
The hydraulic actuators will use 3000 psi MIL-H-22072 operating fluid which is a LO:-compatible water-glycol mixture. The actuator interface with the control valves will include an insulation device to help ensurethat the actuator operating temperature stays within the allowable range.
The actuators will include a failsafe positioning feature within the EHSV's by biasing the second-stage spoolvalve to create an actuator slew rate in the safe direction. For all cases in which electrical power to the EHSVcould be interrupted, the EHSV will provide an actuator slew rate which positions the valve to the preselectednormal position of either full closed or full open, as listed in Table 21.
Five supplier proposals for potential control valve configurations were received. Data provided in theproposals show that the AETB valve and actuator requirements can be met.
195
c. Shutoff Valve Configurations
The shutoff valve assemblies consist of a valve element and an actuator element as shown in Figure 159.Based on supplier proposal data, the valve elements are either common ball or poppet types. The actuationelements use off-the-shelf pneumatic actuators with position switch feedback and solenoid valve interface to thebrassboard controller. The solenoid controls the 100 to 1000 psi GHc operating fluid supply to the actuatorpiston to oppose a preloaded spring element, which positions the valve in the normal position, as listed in Tables22 and 23. Since solenoid power is required to move the valve from its normal position, any loss of electricalpower results in all shutoff valves slewing to their normal position as a failsafe feature.
The EOIV and EFIV applications will use existing RLIO Propellant Inlet Shutoff Valves, as shown inFigure 160. These RLIO valves were ranked against supplier proposals during the proposal evaluation andwere determined to have the highest rating. A pressure switch will be added to the RLIO inlet valve actuatorhelium pressure supply to meet the valve position feedback requirement of the AETB which exceeds normalRLIO inlet valve requirements.
Six supplier proposals for potential shutoff valve configurations were received. Data provided in theseproposals shows that the valve requirements, as listed in Tables 22 and 23, can be met within the scope ofthe AETB program.
d. Ancillary Shutoff Valve Configurations
The ancillary valves, except the S#PV purge valve applications, will use designs similar to the main shutoffvalves. Two supplier proposals for ancillary purge valve applications were direct order catalog items. Thesepneumatically actuated valves are direct-drive solenoid actuated valves which meet the valve requirements listedin Table 23. These direct driven valves are different from the main shutoff valves in that pneumatic supplypressure is not needed as servo pressure to actuate the valve. The electromagnetic force created within thesolenoid actuates the valve. A preloaded spring opposes the solenoid and provides failsafe positioning in theevent of loss of electrical power.
4. Sensors and Cables
This section is divided into three topics. First, the control and safety sensors are discussed. These sensorsare actively used by the electronic controller to operate the test bed or to monitor safety conditions. Deliveredas part of the control system, they are required to meet the same life requirements as the test bed. Next. apreliminary list is presented of the performance instrumentation that will be monitored through the test standdata system. If desired. these parameters could be incorporated into the facility abort system. Finally, the cablesand electrical interfaces of the control system are discussed.
a. Control and Safety Sensors
A total of 35 control and safety sensors are planned for each test bed. The parameters used for control andsafety include 14 pressures, 16 temperatures, 3 pump shaft speeds. 2 pump vibrations, and 22 valve positions.The locations of these sensors are shown on the flow schematic, Figure 161.
Cross sections showing the locations of the hydrogen turbopump sensors, Figure 162, and oxygen turbopumpsensors, Figure 163 are included for reference.
The speed sensors are described on Figure 164. The design will be based on the speed sensors used onP&W's SSME-ATD program. However. the AETB speed sensors will be custom designed to meet the required
196
speed and envelope restrictions. The custom design requirement imposes a longer lead time on speed sensorsthan any of the other sensors, thus their procurement will start first.
The 16 temperature measurements consist of 13 bearing coolant thermocouples (TIC), two pump skin T/C's,and one fluid line bulk temperature at the nozzle coolant exit. The bearing coolant T/C's arc described in Figure165. These T/C's are required to measure small changes in bearing coolant temperature and will be useful indetecting an impending failure. The absolute accuracy of each T/C is not as critical as the T/C-to-TIC variation.To obtain the least sensor-to-sensor variation, the T/C's will be fabricated from a common lot of wire.
The pressure transducers are described in Figure 166. Three types of pressure transducers were selected tomeet the system accuracy requirements. To achieve the required accuracy, each type of transducer employsa different temperature compensation method. This approach permits cost effective pr(xurement of thesecomponents and uses standard methods of pressure calibration.
Turbopump vibration will be sensed using industry standard accelerometers described in Figure 167. Detailspecifications for each sensor are listed in Table 24 (speed sensors), Table 25 (thermocouples), Table 26 (pressuretransducers), and Table 27 (accelerometers).
b. Instrumentation
The preliminary performance instrumentation list is shown in Table 28. The list contains 28 pressures, 21temperatures, 2 flow rates, a voltage and an amperage measurement. These instrumentation sensors are shownon the flow schematic that also contains the control and safety sensors, Figure 168.
c. Cables and Electric Interfaces
A simplified control system interconnection diagram is shown in Figure 169. Fiber optic cables are usedbetween the Control Room and Test Stand Sale Room. This type of cable provides electrical isolation and willthus prevent damage from a difference in electrical potential between the two locations.
The control system shielding and grounding plan, along with a typical cable assembly, are shown in Figures170 and 171, respectively. The grounding configuration, in conjunction with twisted pair shielded cables, providesprotection from radio frequency interference (RFI), electromagnetic interference (EMI), and potential differencescaused by lightning.
197
a0 >
w
X 0)
-OE,
cn E 7R(a -
0 00
198
w
w
CL <)0co
wU wz wli LI
0
z
+0 zcr--
1-- 0
-i U) U) w
r. < <
0 UJ 0
CL.
wuzzww .LL.w
cr.
199
-04
00
CE'
Ca.wAom
0'rv
3: oo
20
0
04.4
to
0
III IdJI 0
II ClAD~a.
tAD Into
-I
II.
- Il *0 -ii 0g Z --J I Sal
,0
.0 to
to
to
I..to
___________ .0
0LNIOd JO iN33N3d - ~OM~i3 ~d
f
to
201
0'
>LLI> z
jI
< UU
i
2020
0 ~ILzr w
Laa-
*00.
IL I
Cfl
_ D
o* 1-0 0 9*0 0*
L I. -W w
D C
- L
u. I L Ix
£L~L ML I00L 2O 0
< I.- II <~ WI L L I >w
* z
8 - dY1N 9 O MN X'
a.Lt.
*203
z
LUJ
-JI
0
I-
20
0L.0.0 0
-)D 0 a)Q
O.8 000'00 0 00 0 0 0 0000000 -- - ---
0000 0 00 0 0 0 0000000z Z, Z: =z:: =- -
00000D C 0 0 00000000 0000000
0 0 (Dc~) 0 OOO D OO
zwui < <<<<<< <<<<<
Ow
L-)
4!o U
za 4i.2 _ _
'A- .C C 0) CC CC CC CCCCCCCm)M M om m
ir!. c C/)U () U)CCcnC) nU W W)O)WWWW U) OWC/ ) (U) (
u.0 .0.0..0.Q.~0 0 000 00 0 0
Lu 0' 0Lr 0 U) no0).a) )U)U .0.(()
>LZ- 0 ~CJL 0 wwta a 0 0 Ol 0 0t. 0 0a 00
U20
- -
LI
L-
C\1 0 0 0 Q0 00
z
Cu 0)
___ - C c
Z-4) C 0 (D o 0 0)C) CD C-)) >0)
-: O-O)~ 00) 0)c 0 00) Oc0
Z D amm.oo 00 (9 cX0a0 00 LuU
'aw < CL 0
m ).
c0)4
0)~0 LLLLa0. ol- E u
i*-*m E
L) 0)0)0) CD0 )t 0W 0.~ ED0Ei
ozC 0)Cu 0)CZ'J CA
206
I- w
w 0:
LiL
I-L
0r Cj000000 00 0000000 0000O
N~ ~ N Z fOow o ),folU C) I ,),C),
w 00000000~ 00 0o o a 000Cwww"" w 000 00 Go N..... 0 O GoGo
< E,
-) -- - - ,
0) 0L V, 6- o-- -L-L00L0000 )a O -4CVC Go 2 ~ 0 00 0 0 0L 0 J
C) 0)-
- >>0<o o o w 00 000000 0))0 0O.-
DO:3 0)0 0(/ ~) L0 U0) 0 0C)U) (1)U)0U)0U)
-LA
"q 0 00 t noLO0)E2
w 0) ~ *~ ~oofflfl II0 0~-00 00 0 mmEE 2 -;; aIO~ 0 D DCDc c D D a
0.0..0hL 00. 0 aCC)
CL CL CCL E E E U) U
207 ) i 6v r
z 0 0
6E . . U) U) CD) U)
0 0
U) CL0) 4 U 0
0 o
< c 0. > C 0. 05s- -
o~0 t C 0-0~~F C ) 0
< 0 >~ 0. 0 l
c 4) 4-L --
0 LL > .QL)
< 00 C/ - c
(3 4- L.LL 4) 0 )0 0X LU0
u 0 0 Fa 0L 7.a 0
z W)
0 > 0
000 ~ 0>
z0 **00
F- 0.0 0.U C0CU0 -0
z EE 7a r_.
4))
m r t208
w ~0A
-h %)- a.
uC E
0 0
C/) -D 40 0
Cl) 0 (D2.0 0
0) 0 0) 0 ~ CM- ocnU I0 -x coO ... w :-uSin a
4)Ul)0
0-0C
LL LD- L L) 0: 0 L. )L
u - a
00
CC4) 4) 4
C" 00U 0
os 0 00 0
C (ai- 0 0
0~ ~ 0 ~ a)- ~ ~ ~ c -u 0..0o
0 0.
4-o) (D0 U). NCI
C/) Q. U) - U) CLo u) 7D MUa) 4) E _ _ 4)Z )
z: 4) 4))ch CL a_ 0 m a 0li- 00. Z0 u -C
209
>0.
-t
COU
CS
i::.> 12
210
LL0
'~ 0 I~ 0 Lg 3
(J 0. I (
mi 0- 00 uJ 0 (3 >0- - - -
N1-0 1-.
0
CD 0
ILI
u -
0C.) 0
0 co z wM, c ~0
Clfl
OM 00 0 L LI0 U,.-.
1-2CD 0 LI
211
0
2 0
WcrI 4r
I-..-c a 8o~ aO0 1.-04
0:) Z9
40 OCo M W C
4 a.I-w
< IL C
w z -
C.)o
OZO Co M Co
LL Z0UU l
~0 0 00 z UzP 0 CA
CL P 0
4J 0
Z __S m cCO 10
91-. C 'C3 0Z z
00
U.-
21
m
0
zw < IL 0~N 0 1. 0. .
w >-m Lam )PUOR~ -J -- 0 I-
LU4 W LU w
4 I
LU 1M M- T-UZLU F- V OA a evvVvv q
LU CC MLIJ o I.
LULW g
L.
213
I- C) Cl,
(., lo II - a
Cl Z
Z LL
m CK 0
LL w,M-
W CL z
214
(Ua C
c w 1 '= O E
(U; v- o -aN 0.E
cr C E
WJU
a 0
%..J 0 QVI
0 Ef a. c-
0 P0
L215
r 0,
.00
as
.0
Ci
-E
A .2
0
.22
.!P
o Ll.
- 216
w
!,_ I
0I -- - -
I* -
U ~ U 217
0
cU,
00
00, 0
0
v-C4 N N
U.
0 '
c h.
a-a
041
ICI5-0 -K0E C e-
- I E
X %)
0e- 1- O
C
'C
I
02 -.q
caI
219~
-j I (
I z In
I CI)
I I.-IW z II3 Lu
I I
I II I
CC ILI lIC 0
zII
ZI'z I
> d
x
C'4'
U..
220
(Ac
0-'J
-Ju iN mmmmm m m m ..0 i40
-1,,IoI oI
I, -- '- -€
I I
3
I
I - I,i = I,
I
L
015 >
, e, -
. l I Q6, lI
>1 -1
0 0A Q
C11 3 °", I o I I 1
II , TI
2.21.
01
o z-.0 200O 0W
Wr Qds I
I-Wcc 01
0
ID.
V..
ILa
221
CL 0!
=
II
r-- --- --- ---
I r C U
1.0 a.I I
U * l a
I, •!i -
U 'I
I. IL
CLL
311 za _
I 222
00
I -
-2 -I - -- - -- -
WI 0 0~4J4j
a-I-
.. .... ...
thI) a4 t
0*~ '0 CA
22
< 0 0Lcc <
0 zz
>- ,r0o
>-o>
CL a
-jJ I.-
w CC
0 > -1
224
> 1.0
LL Nt-'mt ~n 0
N~~~~ 0%Ll.- ~ .
>L r-- cc -- %04 .~h 0 ?xco Go N -
N 00 I! U,Uf -T -0402
4 1J
> tA 0
ko to Ln 0.7.ON-~ 0 C
o 14 0N~~ W
> 17 17 0' m -'
LLh d -O
04
1.we0640~ a0
- m N 0 20c .1
-N .I %- .No N 0 z Co x UN\ .0 '-MS
0% M o 0% c .0. ol
d 0 64+
04.' 0
> 0% 0q4-n m 01 4
V.n % -'2 0% ~ C
-n Si. c a c 0 x ?A
0%. gn 41cc Un 5 04 -
-n N r-5 id
0T 0% cm M- 05.c0~~ 0- -0 0
Z U 0 A I U) U fa540 o 0s 0 0A
A 0 0J~~ 'A U 1A0 0 a "a 0 0 14 op0
o 10 4 a "4 5- C Cq 0 .O. \ o.9.i ole 0 04 0+ r
0 60 ) 5.090- lo 0.% "a a 0 0 n Pld~.2 1- 0-qP 4 o u A41o
2 4 Q. ago No U) q225
0 c
0+31----
0
Iz CL-
~r
0~0
zz
Lin M LO.
El -CCZ
0 > ED
22
xl m rn N : w - 3w 0 x0 i .Lr, N 7N 13O Nu O0 LI = .3 0 I
a.
> ~ ~ ~ ~ ~ ~ ~ 0 C, LATD 1 - %D 3 w z 0,OE ~ 12' N 0oL a cz9 co CI w
Sal
> 4 n n rLn oA C9 G 0Coo~ 0 nA -3z10 Ln L) La0 wCD~ 0 0 CL
cc 0 0 542m 0.
0Q
_%, 00nC L0n NA rA .al N 0 %~
S~L 1( 7 A a
0~~~~ Q1~. 0 ~ A 0I
LA',.. C' fn - M N 0 A.= ..2 0L LA ua wI
0.W 4J 0
12 a-zA 3Im V c 00 NC 20 ..2 IC 0
4) 0 4 0? 0 C
20 m$4 0-4 6r-4 IM 4 j.140U) W 4 - 0 L , 41 -
A W 0c~r-4PA 0 . 4 15 VIaz U 0 -- 000 1
64 -- .w 4p "P q -q t4 0u 0.0 4 0 . 0 (v toIUAP- M 4
UpM + -4 FA > C 41 4 - *,F4 "
3 )F 04 VI M A u . - 00 -40.-NC 4) 0.0 r40 54 Uk 0
rl U9 $1 c 4 0- 0 0 4 .4
>1 L SlS0-FI 0 a . -3 P1 VI a. m1
NS '- 1 4 15141227.
'2 0 0 0 -3 z S43\ \N7\oa 0 0r w* m M
It 004-z2n &. 6n CL Q0- \
-' 41
41 U) 17 4 1
CD) zzCz 0 0n61 0 0 .0
0 0 w U *-4
C.) 00r
w 14) 0.4 41
'-LnA0 Q) 0 Zin1' 4
h.)0 ff) 'D M U) N~ %
cm C-4 cli Go mja 0 K00 0~ \N L
0 ~~~ 4a .. -20.)~ 00or-
FA1 $41
W > 6.4 06L)
%0- el-l - w 0 0 04 9 4)f L) 0-Cl i K t'
>'r~ co n 0 04 17 F
x -N C.) C0% N-o1 -3x V 0 ;0%4)0 i40 - ) (A0 4.144 A 4
0 0 0 0
- %D2 0.5 4.6 0"Nz NI 04 r- V
U)A~ ?A..4 '41 4 0~
C% C4 (U # 404C-4 0 0c q$ 0 ou
in; on c 4 0 b I- 1- 0 40 91 W0:1 ' 1 - - 0. 01 U05 0-4
0z o z0 G .t 0r N ) . 4 ) 0. M.4
M 4a G 0 0 c 0 4 r- I
M 4J4 4
o =' in 05 C z 0 ) 44 0 A ;
-3 0 0 ,% 0 n'. y 0.0) N LA S141 *OGvU rr- 002z 0 .3 0041 0 "1
.m " m.S7.1 0 0 w 10
Ln .04 0 4141 a 0 4.400. M0
.444- CU 0541 0
0 04 C64 4100 0A =4 045 Ch V
0 -'-U 1 &1s4 >14.44~~4 ;hul V91 I - 400~
0.14 4 Ii* '.. .~ uAo
41 044 14 1-4 0 0n 0 A V) e'4.s0 '4.- 100 C: 0 c4 A 0 fO 4 10.
0 A 4 90 4 0e 0 .,1 .4. r4 2 F-J 4 -!u15. 0 LiL '00 - 4 414 0112
III 041 \. 4J 4 r 0 ' 40.'-'A14 41 -A 0 04 Q44 -r -
5.. 54- 0-.I- -06F 0 M t a 0. CULA W0. \ -l W 0 00e02 9 a 44 W( a) a-11 a
0 A 4.14 4.1 04 4) a 4oP 1 0% 4 a x a-I>-C, a .I N41 1 4 h1 1 4 14 m 4 r41 0 U) a .4 *
OJ #A 44- 4 1 - A *A 41 0 4 4 K 6 I - - - C UV- G 5 G 4 0 *4 0 (A 4) > s# -A aC ) 0 F -4 0 0 A 0 .4 ) c .m 0 . El ; I .' ' '
228
CL
00
.2.
fSS
CL.
229
ILL
le L-
lLll I230
00
o-zo
-0.
-IY
Cowhui 0
uj231
(J) ,)
cnwnCim
a-
LU
<( <LLJ <C\I
c'Jo00Iw
A;-
Coj
LU zUIr C/-
Cl)232
cn (9
DZ <- L Ow a L
o u Z Z Z(D
CL 0C) Cl) -.1 L -
ui0 cl L j
Z U-ji U 0-(U
CDe Co O.
co cm ..
-i o L.
< < 0 ; < z Z CL
co233
a.w 2 -1
wz 0SCL z
LU Z 0< 0
-LJ I. ir (D0.' -J z z
Ow U . I',' -:
L0 a.(z0 . V-%J z cn0. L -o s 0: LL L Lif
u~~ CAo Z0-UJ0. > < Lu,
S-L wC
_j I) U C) o z LU Clix < w SLu U to0
< c/i r. w i cm
X Lm xz Cl1
LLJ m(i :3+1 Lo. 0 3(
CLa
LL234
0a:
ui Wi z
w/ a: r, a: I
0- 0 zU-J- - Cn C C
.~z pwl wuz~~~0 wLZZc L
LL LL Lw-,LIL LJC
MW0 0~ cJ wu 0 0
0-~0 0 -- 0U JU
a:Wu 0 MJ- w
0 < Z
a: <wc 0C
CC W a: w-. LDH-0- : U<L
-- 0
s0
N
235
az
I-J 0
ix) I-
LU V-
0 ..4 Zo z
CL) 0L a
LU LLJ U-i0 L_ xciC
0 Lo
2I-l236
iil
ICme
al
Cii
I
237
...... .
ti i
a '
z! z z5
8i IL .
"'. .
_ _ _ _i di _i _i _imw
9- ,,. - - - - 9- -Cd -,d ,p ,, . .- ,,.- . ., . - C- -.
lll l ,i ,i ,i i i____ ___- I -I -W- I -I n -E U
iii I I i i
238l
E
0c 0 ~
- z z
U ;. U Lz -r V-o -4
a - ? - aVwa,ro IT 1 , Nci D
Ni -i I Ni ji~jijiL l
*9 I
00 - N N0
W6 IL j*
239
a
us
O
WE
.a VI -I;,
d -i - i
A
uU
a-.
18AO
I0 LI 0 00 0N
2 224
R
4 LU
X "? 1 10 Xi scour s,
II I ;-- S Is i -
U.- U . vi}
80k 812 ., - , o 1- -0 ,
L, . 0 e (A
ct z
J~0
'U: U ,In ,. hn U, " " .|l W4 ! ,- -- -
111111+ ____ i J ijII I
U 4, i _: 00 . U. v 0, v AL
A1 a
V ~ @ IC I 0 0241
So
U. UE
C0 In 110.?~
10402 Tz 0 soo .'as
w w i w i w i w+0
* ;~ ' '4 in
224.
II
tlli
0 § "
30 03
z
U
SO L U.0 z
8I CLE --
K
ell)
00
UEII !JI:IS
24
0z
cm c (M3 CL a cm-
c c a E E a cm cm cm c mc mc mc _
P- a. a. a. m a 002. 2 2
-c - )
E E - I- " L .-
cm CM
CD 0
Ln do L G n o cc c
w§
m .0ix.W - -20w- Ta 0 a 0I
a~~ ~ ~ at U), 0 , 0 0 , 0 (
-C,~ LUL - - z >2.a CL I- >
0 - E S EC LC C 60
E Eu-& LL E LCLC E m B %.
CL 0 0. L& u L- LCLCL u-~aCA .-.- 0 m M EE0= 3
w . U. U. - o C m c
CL 0 0 E .0 c CL4
E uo
wO(D D~~WWW- WWW >>ZU~i=W J K o. u--------u I-- -u
u- u- LL L. ) u. LL. Cl) U. L.U ) L. L. L L. A. U-
La.q U--q LL. CL 1-- 0. U) LL
244
= u u =u =u 00 Cu CM C M 6CLUE = -
g-, cm 0) -3 -2 -2 22 E 00 E ECL 06C~~ CL C _, _ cm
CL L u u Eu EMIMI)
~E N N N Cu m C C ~ CCL~ CC 0
-- C C.= C Qc~u rI
- _ cmI - CM C C C4C
C~
LV
CL C C L L ICL E Cu Cu Cu Cu C..
q I !. . C
880U LUE's'M) 0 u (a-
00 m lb t xx L N Em
~~ 4M C&L MI
ujuj;,Mc w 0 ) %.~vw R~l~ N-& *h. 0
E c b a 2 'ZR 3 t3ZZZ z m x 0
-------------------------------- -j W -
- - - - - - - - - - - - - - - - - - - - - - - - -
.0 z U)CL'(A45
o co CL 0 .>
U.m U. (a #
0 00
0 00
M(
Is. C L CU E a m aU qU C 4VU
0 e0
M ~ MI-U mCu MCoI L .W
CEE'Z"Uo-E
E~)U
Iz5.246
CL LL
CL I--
a. 0. rEVa.
.2 mmm
> >E H cow I--CIA
a. z CL
(58 a.
80L I tCY a.z a.> a.a. x a.
IL
U.
LL co CLE >ca ca 8
CYChoi 040 CLM8
CY CLca
> > ce CIL E0. 0 EL
.2IL 0- tz E
CL U- >
OLLV uj
C4Is ZJ->
CL 78 1EU- 0LL CL I-. I z IL 11-- 0 > 06
a. P- (9) 0 @ G)
.00
247
c. -1
iij
0- - w -PI
=0s x m a
248
CLC
CL
040
LL~
E CL
cc80z
-aC
A
2490
0DzPz
0 -1
L(0 ui w
z0 -c
j wLLJc/5w - W w o I azz
LL 0. _j
) )
0 mCL z I 31 111E 0 0<L 5 5 LJC
0 0 x
250o
SECTION VSYSTEM MECHANICAL INTEGRATION
The system mechanical integration is driven by test cycle configuration versatility and test stand size. Theoverall dimensional limits are dictated by the configuration of P&W's E6 altitude facility which limits the overallwidth of the engine to something less than six feet, Figure 172. The NASA testing facilities are larger than E6and cause no dimensional restrictions on the layout envelope.
A bolted frame assembly is the heart of the system mechanical integration. Figure 173 depicts the framepartially assembled. It consists of a mounting pad configured to mate with E6, a top plate, 8 top rails, 8 siderails provisioned as appropriate with pump and controls mounting features, 8 primary bottom links, and 16connection links. Completion of the frame assembly occurs as the rest of the engine is assembled. Figure174 shows the thrust chamber assembly installed. Assembly continues as shown in Figures 175 through 179with pump mounting provisions, L0 2 and fuel pumps, valves and actuators, mixer, and major plumbing linesinstalled. Small plumbing lines, w- ing harnesses, instrumentation hookups, etc. (omitted for clarity) will beinstalled during final assembly.
The test cycle configuration versatility is achieved by locating the components most likely to require accesson the exterior of the frame. For example, to prepare for the high mixture ratio demonstration, the two flangecovers are removed and the fuel turbine bypass valve (FTBV) is installed, Figure 180. To prepare for the fullexpander cycle demonstration, the FTBV is moved to the oxidizer turbine bypass position and a spool pieceis installed where the FTBV was located, Figure 181. The fuel jacket bypass valve (FJBV) can be moved tothe combustion chamber bypass location for full expander testing at P, above 750 psi. As shown in Figure182, the pumps are also readily accessible for removal. The thrust chamber assembly, while more difficult, isalso removable while the engine is still on the test stand, Figure 183. Methods of removing the thrust chamberthrough the top or bottom of the frame will be considered during the final design phase.
The estimated overall engine weight is 2200 pounds with the breakdown as follows:
Pounds
Thrust Chamber Assembly 550Valves, Mixer, etc. 550
Turbopumps 450
Frame 250
Plumbing 200Miscellaneous 200
TOTAL 2200
251
The static seals selected will be provided by Furon, Inc. and were selected for their demonstrated highreliability in P&W's SSME-ATD program. Deformable metal seals will be used on all threaded (MS) bosses.The cryogenic seals will be Raco face seals and the hot seals will be Omni face seals. Both the Raco and Omniseals have fluoroloy jackets and MP35N preloading springs, Figure 184.
Configuration control for the AETB will be achieved using the existing P&W configuration managementsystem.
252
0*
uuk
wSI o
a:I
250
0z Cz
P 0-,
0~c
oo
-- -
0 0U) z LE
0
254
wm
0 0w crcU)
3 X UC)H- Z
HWw-J).
N
HO zF-.
255
-zo
2 z z
Do~
3
z 256
-i -j
0(1 01
a p-
:D257
wLLJ z
(D DA
w >-
LLL
LLI D JLLI
LLJ.. 258
wz
Hw l
NWE-
0 >
< 00
wXO4moo> 0
259
crW
0><
'CJ LL 00U) L
Wa 5 L.
Z _jm0 co260
F- U)
-i -
261
z
5 0.
owS
0S
a: > w
262
yrt oc
C/)LI.
263
LUL
co)
T-.
Coc
264
Co
0
cc
U)
LuIU)
0C)
CnZ
-26
SECTION VIRELIABILITY AND SYSTEM SAFETY
The Failure Modes and Effects Analysis (FMEA) and the Preliminary Hazard Analysis (PHA) qualify theAETB preliminary design expected hazardous events resulting from an AETB hardware failure or a hazardouscondition. To address AETB failure modes and their subsequent effects, P&W reliability established a FMEAteam. The team is composed of personnel from Reliability, Manufacturing, Materials, System Safety, Design,and Maintainability. This team worked through the current drawings of each of the AETB components to'brainstorm each conceivable failure and the effects of that failure on the system. Using this team concept,the Product Assurance group instilled the concurrent engineering process into the FMEA to ensure each failurewas exposed and addressed.
The AETB FMEA ground rules derived for the analysis were as follows:
* Single Point Failures - The FMEA considered only failure effects from a single failure occurrencewithin each component.
" Bottoms-Up Analysis - The FMEA is derived from investigating failures at the lowest hardware levelpossible within the current design phase.
" Most Probable Failure Effect - Each failure is investigated considering only the most likely systemor subsystem effect.
" Hardware/Functional Mixture - Since the preliminary design was not detailed enough in some areas,the loss of a particular component function was investigated instead of a hardware failure.
" Criticality Classifications - Five criticality classifications were used in the analysis:
-CRIT I - Major loss of AETB hardware
-CRIT IR - Loss of a single redundant element, both of which if lost would result in a major lossof AETB hardware
-CRIT 2 - Loss of mission/test
-CRIT 2R - Loss of a single redundant element, both of which if lost would result in a mission/test loss
-CRIT 3 - A posttest hardware repair or unscheduled maintenance action resulting from a hardwarefailure.
Having established the failures and their effects, the FMEA documentation was then initiated. Thedocumentation began by drawing the reliability functional block diagrams as shown in Figures 185 and 186.The purpose of the diagrams is to document physical and functional interfaces, double check the FMEA systemeffects by tracing a potential failure to its highest level, and to provide a reference showing correlation of thecomponents addressed in the FMEA to their placement within the AETB system.
The FMEA document provides a full description of the failure modes and effects uncovered by the FMEA.The document is used to provide all the analysis findings to P&W and NASA-LeRC. The documentation formatused by P&W provides charts for each failure mode allowing quick and concise evaluation of the failures, theircauses, their effects, and the controls in place to prevent the failure or mitigate its effects. A sample page ofthe document is provided in Figure 187.
266
The final step in the FMEA documentation is to prepare the Critical Items List (CIL). The CIL provides asummary of all the CRIT I and IR failures uncovered during the FMEA. The purpose of the CIL is to highlightthose failures and to generate, as a part of the FMEA process, all rationale for retention justifying why the failureshould not be a concern in subsequent AETB testing. The following table summaries the FMEA findings:
CRIT Classifications
Component 1 1R 2 2R 3
Injector/Igniter 1 0 3 0 0
MCC/Nozzle 0 0 2 0 0
LO2 Turbopump 2 0 7 0 4
Hydrogen Turbopump 1 0 6 0 3
Controls 0 0 55 0 0
Ducting/Mixer 0 0 3 0 0
TOTAL 4 0 76 0 7
Of the 87 failure modes analyzed, four have been identified as potential CRIT 1 failures. These four failuremodes are within the injector housing, the hydrogen turbopump primary and secondary blisk, the L0 2 turbopumpturbine blisk, and the L0 2 turbopump bearing. These failures will be monitored through the design to identifyand apply proper design considerations and/or controls.
The FMEA and CIL will be updated throughout the AETB test phases to ensure proper attention is paid to allinterfaces. The FMEA and CIL documentation for the AETB preliminary design was delivered to NASA-LeRCin December 1990 as P&W FR-21322.
The PHA was performed by P&W's System Safety Group. The PHA is used within the design process toidentify hazards early in the design process, to ensure all identified hazards are recognized and addressed, toaid in the formation of controls for the hazards, and to track all identified hazards to closure. These hazardsmay be the result of characteristics in the design, a hardware failure, environmental effects, or human error. ThePHA also considers the hazardous conditions occurring at various phases of test bed life including handling andtransportation, test bed assembly and mounting, test bed operation, and test bed maintenance.
As with the AETB FMEA, ground rules were derived for the PHA prior to initiating the analysis. The
ground rules used for the analysis were as follows:
• Reference Document - MIL-STD-882B is the reference document used for the AETB PHA.
* Hazard Groups - The AETB hazards were categorized into the following hazard groups: fire/explosion,projectiles, temperature, pressure, vibratory energy, rotational energy, and electrical energy.
" Worst Credible Hazard Effect - Each failure was investigated considering only the most likely systemor subsystem effect.
" Hazard Severities - Four hazard severity classifications, established from MIL-STD-882B, were usedin the analysis:
267
Description Class Mishap Definition
Catastrophic I Death, or system loss requiring complete replacement of the test facility.
Critical II Severe injury or occupational illness requiring hospitalization, or majorsystem damage requiring removal of the AETB to complete repairs.
Marginal III Minor injury or occupational illness requiring first aid, or minor systemdamage which can be repaired with the AETB installed but will requiremore than two days.
Negligible IV Less than minor injury or minor system damage.
Hazard Probabilities - Five hazard probability ratings, established from MIL-STD-882B, were usedin the analysis:
Rating Probability Definition
A Likely to occur frequently during testing of the AETB.
B Will occur more than twice during testing of the AETB.
C Will occur more than once during testing of the AETB.
D Unlikely but can reasonably be expected to occur during testing of the AETB.
E Unlikely to occur, but possible.
Upon qualifying the hazardous events to the lowest level rause, the System Safety Engineer completes aHazard Control Sheet (HCS) which is part of the Hazard Control and Tracking (HCAT) System. This form,Figure 188, is used to track each event and subsequent cause through a sequential status until the event is closed.Until final closure, the HCAT is in either an Open (acceptable hazard controls have been identified but havenot been implemented), or a Closed status.
To close an HCAT event, acceptable hazard controls are identified and proof of their implementation exists,and the appropriate authority accepts the residual risk. The HCAT may also be closed by the appropriateauthority accepting the associated risk with no additional controls necessary. The authorization to close theHCAT is derived by the Hazard Risk Index, a combination of the hazard severity and the hazard probability.The following table summarizes the AETB closure authorities.
Hazard Risk Index Acceptance/Closure Authority
IA, 11, IIA, IB, liA NASA-LeRC
ID, IIC, liD, IIIB, IIIC P&W AETB Program Manager
IE, lIE, IIID, IIIE, IVA, IVB P&W AETB System Safety Manager
IVC, IVD, IVE P&W AETB System Safety Engineer
The final step of the AETB PHA was to compile all of the HCS forms within the PHA document. Thisdocument provides NASA-LeRC and P&W with a listing of all conceivable hazardous events uncovered duringthe preliminary design phase. The following table summarizes the findings of the PHA.
268
Category
Component I II III IV
Injector/Igniter 0 6 2 0
MCC/Nozzle 0 3 4 0
LO2 Turbopump 0 6 5 0
Hydrogen Turbopump 0 3 6 0
Controls 0 5 2 0
Ducting/Mixer 0 3 1 0
TOTAL 0 26 20 0
P&W System Safety group completed the PHA and submitted it to NASA-LeRC as FR-21321 in Decembert90. The PHA will be updated as the design progresses, to a Sub-System Hazard Analysis (SSHA) and afstem Hazard Analysis (SHA) per the AETB System Safety Program Plan. These documents will be completedke month prior to the AETB Critical Design Review.
269
U,,-z0
z
LL
z ULU
U- -
UI-
z <~
LLac
270
o 0
LL. 0 LL ju
A. x
z 'U3L
0zl
_2 L~
427
C3
0 L..
L&.J C~ C: L- 7
L LO
LLJ U L.<I
<~ >,
U>
La)O
z LL a)a.
Va)
u
U-L)
LA LLDz= > a
m a,-
LL
LO,
L- 7y- n '4-a 3 O o- a,
4-< 0 Q LLL Ca) -a, 3
C "CCUlJZ *o,'
Lo3 -
<3flf
272
*UNITED HAZARD CONTROL SHEET
PRtATT&W~ffMEYGovernment Engine & Spamn Propulsion
Hazard Level Category HRI __ No.IStatus IPage 1ofi1Program Phase Date
System: Subsystem:Operation/Phase:Hazard Group:
References:Hazard Description:
Potential Effects:
Assumptions/Rationale:
Hazar Control Considerations: Reference:
Remarks/Disposition:
Figure 188. Hazard Control Sheet
273
SECTION VIIAPPENDIX A
PRATT & WHITNEY
AETB PRELIMINARY
DESIGN REVIEW
January 29-31, 1991
Detailed Steady StateCycle Sheets
274
N..IN*M~iMM4~$ aNNNKfMN
4= W lo &NPO*O A4 ,.TA T- on onoano A N N "t
CA au: oa# o9 , V o
*~~mu.,!W ..~ pip! ~6.14m 4% . -o 4 t0p on0 lMP 0 M o4N~N~r N n M MIt NWItN N OA. K PP ,P D
u 0a 0 p tw P n9 a a m .a . 1N ft K 0 N~ z4 m 44 .0 *
1- 0 a cc0 4a
4A C 5 .1 0:N4.NN4..am40 0! 9: 0: asia 0: IC! a "0 0 e 0 aP
9L LL a sa In 4 a0o
mum x t-0uzxL@.~ - -a x aa* w ulij M
-- - - -- aN0 w 9u
in aY NW New aM 4 m m r
a a.14
.00N. C;14 0aaNI-N0,-ONKaew'4u0*M 1, 69114 ,4.Q~4.NN04NWia4IN
Itat4~N4'd4. C! C!aaaa..~'1~4400 maC~. 4 0-
W IA'aAa'
W~-0WiS ~-N ~iNN.NWm .S JW mu4
a 44~ aauua u>~ Ea w uJ wl .m . Nm4.,m .m
X0 i-S 0. 000 ZZZZLuZZZh.NNN I- oil
Wa WU Ii-N.mEa4aa.iJ -4AmP.a0.MSi4Na~ i-=mm14-.0.Nha ~ ~ 1 N.d-aNN....aw N N Nwa
aP
- 275
4c..j 14
I A W.40 W 0 &n0 U
alitll I W z44 4 - 0 l2W0 n oK A1K stN0 0 Do
uj :! .. J0 0 0000 00 00 0 00C C ; ;C1.4 1.-
- 04 0 J 14 0I- %t 4 02- 0
.. Ca ea , 41. 'A Wkla s I. j4 to82,~ 0000St!99 40,Nm IA tfL 00 00000 t*N
-Km A sW0NA d t-4, 0 -A 1.. 4 fagt- It .2 *00. (4 0 .AU3 in -
4K 1- 04. .f .-4 Z1 4 % t 04 D% D 04 0Q D f-a U; .0. 43; a14.z...m.4 1.4 1.* P
W>( >14 >U > * >L -- -J 4--
! - -t-) a zl- 8k> > 4> u- W* haI 1.-
01 0.4 ... >4 U1WW hauu 99 >M - >Ia U4 in.4 * 43 -ACC U
W d) 4 .4 mJ 8.J.4W 4ow wa Swwwtn o
W04-0--J W40m0go14do
at~~nw~ a 5 2 > 510 AL0- MN- - 1z11.0.a 2IL *1.- -J1. 14 0 u 11"
0.
-J~m *twn
'S to4 -4JVI) z
04 , 0~4m ha
0 0 *W . . A1 Y ,k* ~ ~ ~ ~ ~ C '. (04*Yl41141
0 44U. 011
Ml 0 o0 -
i VI 0 FA0 0 u .4p,4
0- a- #A .4W n240 cc2-.g~.~
20m .4 *,;m e .4 Q- ha 0 0-N1114l;-2 :S 0 z7A1 r40 Z 3 W: 0,4 @i % '04 $A
w IL cm 4 i 4 a " Si.-4 I~ Oro-4 O P
@-1Sl.'. 0t00 IWO 0 m *Y 00 - i N0
u LL
I m. I.-
hqa 044.
W Ua &IU IL 440 m .
MA Amb. 81 W0.
2.ft 9W4 @5 OE11.If WW4. LU &b £ 0 ~ a N j G
276
3-3
'A *0 PA
I .0 .A
.~~~ .. . .n .. .. .. . .3 3 .p.. . . . . . .C51 9 I 4w4o4 44
44 99 4
#. ;h 4 2 m4
on
6 IL O_
3. X
Hui3.U w~u **-~ w.;m-
&.S93.364.e
43. *30 310.1277.
- in 4p. 0 1* 0n
ta o a"on'
I ;C C; Qmo -* C o o 04 rC;ma4 4tenO; 6 C ;6 ;
&a.J0 000 00 000 00n n
to4 W4' Ah ICAa0 lcen -t0 w 4 A*
ml ul.4 0!! Im W ID n CA,A *et A , 0'V @00.0 N.O '49 04 m W 4Sl9w 1 l .t. pnO 0 .% Nb* P3 .'.' .4 .4 OlC.JCeaCrl ej
a"'A88a - on 0: 34* 49 4
w -19 'WL >h-
>a 6- -ta goi 0 ci
*I-4 e0010 a - .40Ln.4 .1*,iema 0 10 qa tbto so mi A O -4 j ---a0 CL q.4 IL 00.
Il-S.fHa. I-I
0 - ' Z'- 8us cc 4-u -
>14maJ >0 U~ *KW iZ W.J A 4 8
> .J>Zw * K 0 Wi I .- WL j
s0 W.00 m> A33~ x
h4 tUma "0 *OO Z'N- I-. -WI-h-84CC
tH .0.4 3.. m. **-4 .
H- ma -*0 x x x0 -n M. IL 0006U.
z4 N 4 m.1 ma VA 3.4 in *cir .4 *en8
maw i 4 * *OeP- , * 0, " P0. M 0 * M:N
04 u0ma~~~~ 0; C ec'enn w OD,
0 30 0 0 a
ma~ 2, i-.0 n0 W0?4 Z0 *.o *.,t4
0 -pn ty N* M I NT ka 0 A0.
0* N8
a 4
son zpo . a,- r, 34*0 I .4 ,-ed ee
*. .ma. m 0
4 m
U.. b44 040 aU
-m PI-S UL Oh 0 .1
0 ~ ~ ~ ~ ~ 0 mamaI.d a -
2784 0
-a -a p1 t .o 00 . 10@ p
#4~v . : .4.aEE#.0C 4.044 0*00eo~m o~
Do-~..
V.I.0 1. .01.0 .1 '
I-.J
ow t*m * e*0 0 a % NNHu AMMf"HE 22.3Haa" AS 1
aZ a
I- 4-S a
a,
u UIH 1a3;4 0 I99 &
.4 . SMmm o aaM o
iz~ a40
000 *a 2 1i. 444#
p I- m- - C 0i- Z -0 Z ZX zz a-a00
z- III A
.'a i ..... .... .. .. ...
Ua %DI 0 00r4 WZ40 0 0. P
a II'> .41f - 00004 00 61. 9
00-.
cc 000 cof4Q-41 MO 0 #aae J uPsSr
.. d04~~o . .4 04W1 Nw40 z N* . NN4 No .w.e . aeat~~i ma * a M 10 L4o .. t OK' .t9.o wa
ea in4 U - Ci- 4 A0r4 .1. . .OO 0 N i ; Z m o we e O m m r1'AC .- 49m N- 0 & c 3Co "
ma.. 0 *440w*r4 0 N 4 - U . jr - j I---096 V40 A O U
z 1aA44m,4 on. It n Ic4 ~ ~ ~ ~ ~ ~ ~ :0 T4... -u-~.*e ~ 000 N>I4 m .W..'. >.4PN >*ti .4- j A' -
-~ ~ U *oo z>5- ~ >-- I. us.j~~ ~ ~ ~ *41 Pw-9~ > ~ 9 ;W_
u) 49 w ".- -.W03 I - 9 o2
P 4 @2. us mam8dgm
.ze. a' 1.4X cc 0§ N 0 w(Y30 tu Mi mu i b4 z a0 1oaa x aam x2 mama i j
26 a0It . Ua. (2A I& IL U)J.aJi a ..i. IL Ua.a
rk 0 0. ID40&. a1 %t0 0 .N
-aL -in'j*.
.4 .5 8.4 MI
a S z0 .144e*10 4 04
g0 a
0 -SC N *e''N,0N 44 X
C; #4 0S00 a, .4~ 0aW - n00OM2 @'9 000 I e t
*- 0. 4" o
Ore *0* ,N M a . " ;0C; a' @ *4 a, a 9 @2@ t 0 0 MSe' 04 N.4W.p4
ma CS w42 La j-
0. ~ ~ ~ ~ ~ a 0. W24-D.**. .2-.9 L Mee~ .Op&. c We'.ao *Ue'w4 ea.w I-2'0Ow.4W. me... 4maO. w *u a. .400'4*04 -4mp490-M U*.4OaA1.4 2 LW 0 a ac,.W 0 W ) 044W 4 - .MC1a' 9 M - "40 z 5 P- 0SLM MW -0
280.
*4U *4....,*@oe a00000oo0000000400000 N PNNNN
OIL 1 a 044 4
T T
.~~~~ ~ ~ a0 0 ffd ft 0 0 ood b
ZW Z~~~~a . .....
.a 99 0 NN U = :0~u...444 0 pe 40,0" 2,144
NNC49 N @ N 4 40 4 4 6 444 -g 4646*1
4v1.Z P, I, S444.
u a
O6 NO I. o. a6 64 4 4
Owl44. U -- - --
* 28a
O NN 0 0* m1-4 wzccco04
N #44I
40 *120 * 1.go
*O#4 01 w1- 4NdbO 4 US140
1.- 0 -U 0Ct MW~O~ m~dd ItN -M am a N0 N NO 1IXL.0 - M Ix O 44 #00 .#4 m00 m~ 4000 -mai 0-. 0LIO -00 0I000n0
10 41 UW> r4- W 4O4 420 0 z *00U
0i f~ ILw~'# 0* L-I-~~l0.400.4 N~f#N ly FA -dl .t 0 P. OId O~0 00
w~~0C 00J dus l 2- (* C; -
OU>>O > U
to...........................4 . - E000 1.
W -C#12 W L614 N U.w 0 0 l "a i dbe21 inW O IdbON# .tW PIJO -jN# 94~e - 4-9-100. 400 "- CO (nu .- KV 0*-' ? Jw(A0L
z aI- a. P4 0c0U0 Ul 4 *6_ ow0* x 0 a I- 02d _ L. *i U 4i U Z i==x=
w>0 Ml Ou x .- .. w 4J. 1.j- - Ito 1 U4> .l- 6. 10 It U4 0- 1= 6 6L
(00> M1 *C 4
CD C;
00- >bi U; 14X * U4 le W 4.04Z0O ~ U 'D0 a4 4 41 4UP
m .4-~ 0 w ; a'wa00K> *00c a r r
z .14 ; C* .I 4 1I N 0 4 o N 406 4O 14C* 1 >O N O U .0 0 n c
FE=
ac r4 v4 -44P0 a .-m .0 . 1- 1-I- O
c;m wi 4 toH a c q- na N%
af 4u W D0'N. N*;c c C, 4 *
4K VA ON 90 D %T 4D 04 A w I -0
uaa N 40N a 'a0 U
£ 0 CO N .4.O N t
00 d*4 o# x 0 N 0 M Z-J CL LA 1-4U U) 0 -1 IL b -j -
0 ca 1 0* @1-1- * Z* 1- I. 0
-i 011. 5s M
0. .*
0 -J CD*#4-0--WC.a, *iIin ~ "w P-N -P0# 0t aW*400 "#4 = *II 0 .9~0. LU4# 0 w O44 Z w M wl- u u -MILILNN0 Ilba o Mi WMf#4db4,)b 0 490U -# *O~O .0K
0 w 14 ,4 *,0 . in 0* 0. vq 1I a w0O -I-S X .- 00 a F*oo, >~ x t #4NP--Wx#4#4
2821
.~~~~ .. .NN . .N ....
U2 =2ig: . . 0*oim ., VA a
04 a a 1a0a 00 4vm 0f 0:mc 22Z. =11100d
.0 ...... *
79O~*.4ONaS444S0N6Ia* ~ .
enNNN~~ a... @z.zeeas...4~
.a W- sww ma
x-*NISA4~~I
a m ax
-~ ~ ~ P 4 -44"
0.
.jX~C w 6.e-e.
aen~ew."'Iae lir
!!4
0. UW eM
U 283
.4 mi Nt00
.4 ImmW 4-000M000 tIt C - . 2
LU 'd U-*
on 00g U C. 0 m*0W..4I (AN .414-O LIE A t Wa,00ej0 N
m AI. Ae C 11 A -4 IG "I4C va0U Ctt (J 04 C NIxLL. 0 N. I.9. .- 0 to %t. 4 000 0.4 K4 00 0
Lu 9 C 0 0
IL 1. UA IU -M04 ON > >Cl.0N40
0_40mr40rl 4 N 4I.0 4-a, LL. 4* MVAIII -. * I Y , im Dr:Ll.
Vw oNIt LU W rAL0 .4 1
00 .4 N4 N0 m(fl*P C
9 *00= ==z x =
*00n Q fl I * - --- u -
us > 0 LL IL 9N jK> -1 0 w >0 NW w Z 0i Cz4 > Wl i > 4I 0i Z Q8_j
.> 0.>.4 xU 2" * w. -i ;- Ua wdo -49JZ 0 mw 8024uul-LUW80,
0n). 12 >t I tuAw-4 ( U4 4 -CCCuSa34La
&O .4.~ .0 Z l-X4 j .49 .4 0t (fl .4I .us.u .LA i02 in at0 to . LIO_ > .J4w 4i WW0 z - w u j i 0 U=_jw
HWM1 n 1-4 14 1 *-> o w o o xw L xao o .0.=Z=
w Ia. 0 "I. L6. LL L ... I
1-4 I-. - UI. C -
04 .6N ODT
CNNc * , *.4 4'A l04am wx In CUt.Ui40 ~ ~ _jaV CONeZ -A 1 4 LU
b. 4 * in w * ncri tatc
LU 0 a , *o .6 .. 4 .";.z" ," m
CK IW30 C.4&(W .4C .49
0 * 1-I .a
Ni L 0j In.0 4 0 m9 W nN 0OK 14 go a* 0.4 14 U
01 z* NN 00
Z ~ c *0 .4 LU 0* iAa Dj 4 1- -i0
P-N 0j C ; 0 t ~* .2 0 C, 4 .4 .
0Ua.
PA a.44
>U0 0L o 0*4 9 . M N . 0 a'a 4 m0 , 0I -0, ~.AN N LU LU* IZ ot- .m 9.4 N Mw a a. .4 0 Ih. 0 FA ItU10 0.MLU *0 'tO .4 I9 C* in in It C4 t W
I4 L (4**N .
u LL
10~~ go.cfN X Z1-xZu ogo
- N- IL -J
_j ~ ~ ~ ~ ~ .~ 1. 5 l uD- I
0 ow wW4.LU
5 -4 -4 WI.. Inz 0.4U x x C&
UWULUL 40J UW m L 4-
IL 00 0. LU P440L ~ 0 h0 0ILULwLU l0:
284
I .... . ......o.o.a.e.e.e.e. .o.o. .
U *4
s 4. 9M. ~a t ...
NO 04F ' @00 00 UU@ps 4 .444~~00 8u 0 WWJ9 t4 @ 4 4
~.a....................... .:z ...............
I.!fig~ M40 uW@OR If44W ddiUa 0-a -
N~~ ~ ~ -tWt A 'A om
'0~~ 1-Ua 40!
a -II4
Za~ ea~0. P. z ;~ 44W 1-4j3W.a ~ ~--4 w-
a a
aeS. I. - X- X- ZXXX ZR @3285 2~
4c1.4.4eco-0 WI
0 y -4-.4 -9 40~ N 10 w sZr tr- 0 0-4 10 40,0 0 .t
doll q .. .4.C .00 . 04.w 0~ -. a Co o c ;C ;C c C;
w.J0 I00 MC 0000 00 t
cc" A - lo L. 0 ejto t -444 . 4 .4
IL *UC zuI0. mu > C R
>4 to* z4 *Y ins~*.~ 06~ 060.0 r*I-.4oomoe...4emmIt-4Its0 cc a-*,nm N*AO'U"Knw -*-JW a N $ .l - P
00. (A. 4(4 4* Dt ftm OD wVZ ILN 0.CL I § 4 ;-o
-~~ ~ --- -- 7 --- ;00 8 *
u > LI L. 0 u -> i (Aw > 0 I-W * n
II~. -4 K >- tB -- S- - --
00> o z Lw zowoun 0 >U. m411 A4 - 4u wA
4~~~ ~ ~ ~ &*1&400xa C .c1 .-0.40. *(fl&Uwu WL
wha01 *-.9mt=1.J -A- -; mc.-cc~= s0 0 0 M- I-
I A!9-m ) I.- w w *= JJ - - -- l-
-J 0 No * . .. 4N.I. a4 40 a. D 4't0,Z I- x 0 .. O0M 'l0
1- 00z00~ ~ r4
9.0 0 1. OW 'I .NN M P.. .
*0. 0
Lm 4c4 C;, M* 14N 0w ej9 a, n-at-o. A iI-. z 0. a IOz a,
C; .4 *8* 00*S0 r ,I-~~L Cq 1 (3 C-.tf.00*
mum z54 .4mNm4.I0 IncNW 14Ntf.
04 0,4 ~ 4 .
IC -K
0. 0. 0* C; O.0i N*N4 4 0 V . * iC C ' 'l
Lu to a,' .4 re mC m* *N t .0 m 0z -j cm* 4N OS, Pei C N eN r-4.-- -4.4
I, I- zuZ I- M ZU05 I-J0.1L14W a.0m-W..J ZN1-- N- - W. - ON -
*~~1 a.a.- 'a
.. JSa Lwa .4 wuM. U IA w *w mu I.-
040.4 -0 .U m
CL I- m0.I mu ~0 04
. L-0 W4J 04.. CI- 44
WZ1m(fl 0I->Zms.( I~~~.- . &
286
44 1 6
0 ; *e pa w m~ C ~ EEe aeI-O O
C!~~~ ~ ~ ~ 0:..I.J8t1 0!4 1!1LI ~ ~ ~ ~ ~ ~ ~ Fe SISP.NE6..** 444 444N.6.6.6.4#VP 60um4 A. .SS "I U6 .m m pm m $.4 m
44 a
-9 aa t
4.4.4.4.4 I-
Ila
6-~~ c-W I
.4 L
287.N*rNN66.a46 444
49m4 T- 0~
C 40 40 0 644 Z
0 49 r.4 AX -0 4wuZt 0In MALnP- 'ts~. K- FA 4 40am r
do a
00004IN4 000 00000000
0M-ii 411 100.4P0
08 (A I n-
Oum~ > W *U- Y - -L. i m 4 4
> gALAIn CA 4 N 9.9.4 .. 0 P4 .. -j U "*~t 44 CC00 t O00C 40000C C0 P, P-
0Va4*DLAV..4LA *I-4f@0 Iag-*VI
. 4~ e-V a 04 V A t J M1 rlC
lo. !F 4.--
lz - -
-m U- U* -ul > 0 w U.0 * IxI- A*
> iinw > 0 1... "1 1.
*1 -9 > .4l - In---u08 8i4 I >. -j > -KllW q_.U
>~ .9 4I > m : 8 J40Jx0u ww J*Jw00>4 ~ ~ 0 K0 008
fr444 IU. It4 4K 5.4>0C'JN wOwLaLAZO.O4K uaWAW 4WMU 2c~JJJ(C J .J..J
*OK ft- I- &L -I-z
t3 Ii 1. >.' LC
& " OZ- 54- 01 0
22 X9 50 j -j.4
~0 U 5N;@ I-LACJa d0
ILI 0 50 V@w0 o4
V -1 .4. V 4t
z -l
0. *-o aA 0~04 Z* ! W! a * MP A. 0N P
do5 F t0 Z1 -*0 N D- *nNMC4
w Ins P4V ' '
In 414
50 LA * P4a in *0%a 0; -- Va'800Va-a 00* 4aV'D 0C
De - sU as
9L IZu0 IN MZ UO J*
Mi -m 4 ) r 4
.4 5L CL4M C- M.
U i4..J Ib~g q 4..a.JUa.-nZ u- u .- ISA.LU
im E ftO @@. 4. N0WWN4 44444 N@Nd10 *- Ir 0 N 4a4 7-00- ON. . .. . .PNN . . . grgzgz
0000000000000#0 000fttp
~~~~,~,- ZZj ee *.eas ma ee.U e..
*I. . . . .l . .e . .......*.U .. 0t
U ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ;00m .00 ." ...a- ace 4N N00 @ I4.EN
6w I.SO O 449494X. xx ZZ 01
040 OO. M~jOO a.. 44 ~ g a...aa
EN:A A~ Z" 0 - --
agal
-3 a
'a6 0- baw
228a
4 4-000,00
IA IA
tur44 0 C40 M 4I 4 W00us PA00 4WZ I N~l l.4 MA m m Mg'In' U4
4c *IC 91 G!'t -Cu S lm.4 4 r4 N N M OM .4r-N9l i 1 1 1- 0 30 C.K 0 0'.8 0 ~ 0000 00 00000 00
do~0 000 00 00 00 0
01-8
1.- (40 I-- -0 000m~~~ ~ ~ M SL 6ju ;C ;r q yNCiW 4
U- 11 U. r4 .I %a 49 -
us4 Ix.6P IA 5W~WO >4. 0. CL >
.a m -10.4 Ln _jmUYC
w 40 4 r4 N JO0 O IL 0 CL ;-lZoJ~~J4~
--- --- It :Cox xxx x xx0m~
>4 >A >0f> >IZ>,C,
-. -4 - -taw
1 .- *
>.A >0 SmK4 > ft SO>. uU8j>A mu4.~x 0 *w -40
00> om muw S0080
450 *..J * N0 *4 W
AK.4 0N4 .4 -j-4m i
$. = 'm 0 0 = =mI. 0 - .--U. (A mdi 00_ j-A & .u
0 ~ r4 'tr g tP
I.-~4 CJ mNt 0 At m * 0g jNw 0 *P10. 0, .. 05 ; 0
ix SA r4 %t1 rtQgQ@
D to I-S 00~ 03* PI'AN?
U I c
w do -I r It 81.00P Ci0
IxI
w~ la
I1*U 0 Y PAmuw -
*[email protected] a 00000040002:
ilia
fil -!"
HVI*& 026 0000000000000000000000
W W
43~ ~ ~ -1-lqx.-I ~ I.
a f;f M to -
i NoVVa:x X-~ZX Z fa~~aa-aF::48- a.
XX- X1-m 24 aMau a
ku- ZX -
291~I~a33
wZ 0 K4 0 10 a4.4om
4i 4-00t-
-A 0 f a 14
- L n14a 00 4 CY p-4( jUS
14 .1N0 t0 u - Ln mia 0* o4 1440 in4 Itin00ol 0 .V4 04
0L 4-1 1- . .. 0* r n M m -C iC %
4 -4 - N * 0 ee oco cc 00000 cc
I; LOw4 -1. r404400 -j -1-0 @i001 9 - r 4 r1r4Nt Nci
&.fAU 0 - Cc L.to0 -1 0 .44 4 4cZ£0O 0 - 0-
OU>. 0 UL wIa w
> 0 KN 4 %t 0 M A a M @.4.4 . m4 .9 . .. U44 0004t 00 00000 NN.1
.J m 04,040 on 4 ,.4 44
00I. 0. 041A 0 --- --
- - - - - - w- 5 - - - - -
-0
0 *i > 0 I-tu64
c>0 m I00 Wj4 >414 0. m -' w ZU U1 -J -J nJ -6 - 1> 1-4->=w 6. u- 0.cr. -- fl----044
> 4 U~. * m I 80>.W. IXU P4 1-01--cc0C uU0NJ4us~ IL L&JZ *U WZ 0 0 Ca, bc c b
w.. cc 1£ .OZ "Z w- - - IM
1. 39 Iio 0. * i- i-4'wmw 49 NUMuxxCij i - ij
I- 109 w41 =110E-.~ 0 0iX *
(11 -Jl-I 06UU *l 00 V i i - - .
1- ~ ~ ~ ~ ~ ~ t * .14.40 I1t41.4
c-* 00
4- 0! 't6 .41-0 *4 11 1. 00w 0Miy .4 M *4 01~4 40
4~~~ ii0 * 4
In~~ ~ 0. 41* -i U i M SnW
z eI-0
= 0 J D C5.1*.. 44 *y L"I K rn in- iZ11 * 4'.tU4 4 inN A*3.4 4@. *,a 1l00 Fn@a,0'4FA0 e
Mi 3 n11 P4F-.4 .444-4PA11 zs~ " . . *411 c, o 111
in e* m. 41444 14
0
B.
ne CL 0I1400 *1 rF wy Mi& 104.4 600a o04 40MI-S C t- 404 go1N44 C; Z 4 '1 4 a, W *0'.4m 1
-z -u *0 - -NI-U ILL. wCa.I j-
4L
*~ ~ -K Ca - Ia
I.-. ILU0
0. 4 0 4 01o
u U, WOE 0 Z -44 . 4cmiW V0. W£WI- M O wi I-U
1600CW404 U. b.£££W44.J .U Mie
292
I Ir
z 'o
a C,
-CZ WS
A.ab44on 0
la Id UZ a -- - a
IV: 4 .4-
.- ~~I t>, 4O 6I44 4
Gn I
ffl S293
"A 04"A
I a aa
.Z .ZZIC
I bu 10Az x 1" -4 -4 a
151 114.51111 1
w44 48
oo-
mK -A
=: IL I"
Ut0
5 0 6 *I*4* 111o
~ -uN9 4 44 146rIfl EEE~EEE
3*M.£
S 5294
4 4-000000-j f-(a &,
uj NI 11'0
4~l 04 In 4a Ieno a - 0
46 Oll 4 4 X0I 0wz0a nL F 0014tM O.J t In*~ C J 01P A C
0 - It In N N00>.0 (n I- 40r "n 0 Ui - *a t0*
491- -1*u14; 4 *CC ONO 4-.4N I04*41 -40';c F q 0 Qv00. LL4. 00 zI. n, -
z C*
.4-4. VAonr
?A 0 IL -K- (A 0 -4
- j > 9w ->0 4-W *6 0 u.4Z4>~~~ ~~~ -A(-l- >01-&4 0> > 0 WZ 11 wJ I- u M C
> >4 >Z )M i ZQ R -W 0-WW - - x00% >Olf" U. CZ L04 Z W 0 4 4U
0 0I >1.W 0 06 u1 0-4. -K4. uix <44 f4.4 IL A iIL0 0: 1.- X - * N 39-90: 400 44 40u
0. -- 00 - P.4 L 00 CA -40 m .40 0W wA Uc L lW j w0 s. 1,- 0w - 0 b4 0 --
m I*M59 .00.I.- z ~ 0 --. I0 .. 4mOM * . 1 1,- ON 1.<0 u0 NN
4u IX " A404wwwcu "u xC -1.4 *X . .1JCN N _.1.J -J -J_
I-L
4-N co wiI * c % ,a
0. a,~ 0 N IN 40 .Ja* In tn VA.40.4
4.4 -I 4p N0 041 -. 41. 1 0 l 40 It #14 #A e-
't In .4 W E C
I. Inf00- ZO.CGo (Y~
M IL.
'4j 04 *l :-aN 41N014 X M* t-i 3 0P a44- .0 1n,0 4p$.t *a in 4 0
w "A (n M * O ui. c;9 0Prt4 t* -. FAgo j W 0 0*D 11U .4 In1 -
01 111 ON 0 0
0 : 0* ' '- W .a .;*!c :r ; oN&
c*",1 0 'T 0.tDC; L Ltt .4 ri 41; in *a a , co tOr ,acc -3 n ,' O n NIno 0 u Z" CN*0.6 c, 0-.6 i -,4 A-4fj 0* 0. Cal 1 a g
.4* 0 a, .n
0. C. U*N *M CL -i C004* -J *i4*.~4 *0 '.04 4 W* C- N ma0.CA~a
~ .4.
UU LOUa
*j 0 M E -. 4
i.. 4 . IuiI .i IL - I I*& - Z - -uiu C j 11CLi
" ~ (A4. 0I C -u w o a 4- I4
W1 LL jU o IT4.4. LWW w -. Ix4wx- x 0 W w 40 U. IL
29510 Z4444Z4
~ ... 4..4.466USUa.a* a00000000 * a a a a-@oe
- g ... 440444444
T~ a
1 0
4~U M 4440044S3U4
0
- -a ----
iN WA I,44 N xn Q w 0 w9 0 N UIN 4 a a b.- to S i.NNN
X. X 1 - a
-R R 12A0.w a 00I 49N NNr. o Ntzzlmo K NP.0 a o.N. 9 o 4 m1 n. 490 94a
in-a 0"a @ow4 s0N 0~. 4 o o9 00 4 060 aa N 0N . 4.4
a4 a
a 006
0 . ~ U- -- K.vOT -f~t z~ 5070: 0
~~aZ NOUN UUUauUU O Z Z -a ~ B
229
4-00000
K 0 M 0 -9 W mfl0 -*L t 0a0 LnN 0 aILWU 41 0-I 4N,'O M D M0*1 4'>. C 1-, 51 t. 4 ;c;c;11 4% N 004 0 t a O t r4.4.C c q0to r P.4 0 gJoS .. ..
OLmO N W c;..
I'. V)U 0 N M onoe m. '4t* 4Y 0910 ~ ~~~ ~ ~ ~ -J J 4000a .o t 1a a NaN$
>0.0 PA> M 0 F (L.0. MwKU.g
> 4>09L AL >0
K 1. 1-44 -' 4. 0 0 0O 00% O 00000 N." @NICY f, r4 * J-4g#AQl
W0 (4 -4 so in. cC1V It--A qa4.I IL4 .4 NO I ~ z N zYM% nKC
WU U*K0
> -A mWW >0a 00 W -7--
4> CLJ>4 xw Z _* 1,J* K uW ixW Ul-WiuW W0do >4 419 Z, * 8O -CCU Z000DO 0o04110 > U &IK U'4 w -K4 K W 0 4 4c4 Uw4 w04 lW L6 U.W W 0 0 to ac~ . 13cXl 00 .4
C.)-I00 U 00 -. . -IDIW 814nvonW2 ~~~*00mN(0 --- am-
V R R M 0 W11.- 411 * 4 0 N4N0AQO CLu&.W-0- Ma ca U > Z Y004Zx~
-~ 1 i_ 11 (WA UM.,& 1011..4 *..I.n UJ. 1.
1...NI
K 0 *- -CIs * 40 -_ft4Ntyr-b
lZ m M .t 4 O-t OQ% r
00 *;;"
N!-4 *N4,4 *4A 40N4@0'PN OW~~~~* *..4t.. - Da,.0aS 8 W 0a tr O yJM (1* - 4.4
co 0 0*0 00
F_ OIr Ln 09 0 *LU *U *44N4N M*p @ rq M .,N40N40a C;4 -4~ *40 ' Zr4 VA *4,4 1-* NN co *;t
0 j 0* 40 N 0 -4t
0 0
0. I. 0*.4 0.v.4~ M iP, 4 M ; V e 0 N 'U W*40N N *40-4MONPOUIQ 0L OE I1 *n . , 0 .1044 ini i Z*.. . . .*a, 0 . *. 4 a,&. 0* M a I M004~~l.0 It. I-*.p.N *~*1
*0 @N t 4 * .4u 0 i0* a *N44mmm400 r
0040~OZ U0 N*44 104.4.
0 0 ! O.L6 I.
" ~ ~ I- a w p 41C ..41uI. U p W I-W I.
Ow'w. - 9 9L W 0 6
U' "A0 1-
40 W 0 N0 (4a
U.0o0.W$"?-j . owOw0
297
NASA Report Documentation PageSOaWS Adn-arlailo
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.
CR-1870811
4. Title and Subtitle 5. Report Date
ADVANCED EXPANDER TEST BED ENGINE May 1991
6. Performing Organization Code
Preliminary Design Report
7. Author(s) 8. Performing Organization Report No.
A. I. Masters, J. C. Mitchell, et. al. FR-2132910. Work Unit No.
9. Performing Organization Name and Address 593-12-41
Pratt & Whitney 11. Contract or Grant No.
P. 0. Box 109600 NAS3-25960West Palm Beach, FL 33410-9600 13. Type of Report and Penod Covered
12. Sponsoring Agency Name and Address Preliminary Design27 April, '90 - 31 Jan. '91NASA Lewis Research Center 14. Sponsoring Agency Code
21000 Brookpark Road
Cleveland, OH 44135
15. Supplementary Notes
Program Manager: W. K. Tabata
16. Abstract
The Advanced Expander Test Bed (AETB) is a key element in NASA's SpaceChemical Engine Technology Program for development and demonstration ofexpander cycle oxygen/hydrogen engine technology and component technology forthe next space engine. The AETB will be used to validate the high-pressureexpander cycle concept, investigate system interactions, and conductinvestigations of advanced mission focused components and new healthmonitoring techniques. The split-expander cycle AETB will operate atcombustion chamber pressures up to 1200 psia with propellant flow ratesequivalent to 20,000 lbf vacuum thrust.
Work under the contract began 27 April 1990. Effort during PreliminaryDesign focused on: (1) definition of the key methodologies to be appliedto the test bed design and to be verified as part of the AETB program,(2) development of transient and steady state AETB models, and (3) pre-paration of the AETB preliminary design of major components and systems.
17. Key Words (Suggested by Author(s)) 18. Distribution Statement
Space Propulsion Rocket DesignExpander Cycle Engines Rocket Test Run General ReleaseOxygen/Hydrogen EnginesLiquid Propellant Rockets
19. Security Clas. (of this report) 20. Security Claud. (of this page) 21. No of pages 22. Price'
Unclassified Unclassified 301
NAA OM IM OCT N *For sale by the National Technical Information Service, Springfield, Virginia 22161