NASA-CP-2283-1983 Shuttle Performance Lessons Learned part 1《航天飞机的性能 吸取的教训 第1部分》.pdf

1、e N84-10115 NASA Conference Publication 2283 Part 1 Shuttle Performance: Lessons Learned Compiled b.y Jamei P. Arrington and Jim J. Jones NASA Langley Rpsearrh Crntrr Hamplon, VirBinio Proceedings of a conferenee held at NASA Langley Research Center Hampton. Virginia March 8-10, 1983 RIPRJDUCED BY N

2、 AT1 ONAL TECH N IC A L INFORMATION SERVICE U.;. MPIRIMINT Oi COMWtRCt SPRINGFIELO. VI. 27161 National Aero3autics and Space Adminis!ralion Scientific and Technical Information Branch 1983 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-e -7 7 Kev Wo

3、rd? (Suggested by Author(s1) Space Shuttle Flight data Aersdyrtamic performance Convective heat transfer Thermal protection systems e 18. Distribution Statement Unclassified - Unlimited Subject Category 16 1. Repwt No. 2. Government Accession No. - NASA C1?-21Si. PdrL 1 4. Title and !;ubtltle 20 Sec

4、urity Clauif (of this page) 9 Security Clauif (of this report1 Unclassified Unclassified SHlrTTLE PERFORMANCE: LESSONS LEARNED 21. No of Pager 22 Rice 660 A9 9 7 Authorbt Janes ,. Arrington and Jim J. Jones, Compilers 9 Performing Orpanization Name dnd Address NASA Langley Kcscsrch Ccntcr Hanpto.7,

5、Virginia 23665 - 12. Sporisorin:jl-S -d or implied, by the National Aeronautics and Space Administration. iv Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-CONTENTS YEFACE iii Part 1 _I- ASCENT AERODYNAMICS I Chairpersons: Tru E. Surber, Rockwell In

6、ternational William I. Scallion, NASA Langley Research Center LAIIKCH VEHICLE AERODYNAMIC DATA BASE DEVELOPMENT COMPARISON WITH FLIGHT DATA 19 ,J. T. Ilamilton, R. 0. Wallace, and C. C. Dill SPACE SHUTTLE LAUNCH VEHICLE AERODYNAMIC UNCERTAINTIES : LESSONS LEARNED 37 J. T. Hamilton LAUNCH VEHCLE AERO

7、DYNAMIC FLIGHT TEST RESULTS. 41 L. Y. Gaines, W. L. Osborn, and P. D. Wiltse AERODYNAMIC ANALYSTS OF THE LOFT ANOMALY OBSERVED ON ORBITAL FLIGHT TESTS OF THE SPACE SHUTTLE. 59 T. E. !iurber and J. S. Stone TECHNIQUES FOR ASSESSMENT OF ASCENT AERODYNAMIC CHARACTERISTICS OF THE SPACL: SHUTTLE LAUNCH S

8、YSTEM. . 79 Kenneth S. Leahy ASCENT AERODYNAMICS 11 Chairpersons: Barney B. Roberts, NASA Johnson Space Center C, L. W. Edwards, NASA Langley Research Center SLiPERSOliIC LOADS DUE TO SHUTTLE-ORBITER/EXTERNAL-TANK ATTACHMENT . 95 STRUCrURE:S C. . li. Edwards, P. J. Bobbitt, and W. J. Monta SHUTTLE R

9、OOSTER SEPARATION AERODYNAMICS 139 Hark K. Craig and Henry S. Dresser SHUTTLE LAUNCH DEBRIS - SOURCES, CONSEQUENCES, SOLUTIONS 159 ?lark K. Craig 4SCENT AIR DATA SYSTEM RESULTS FROM THE SPACE SHUTTLE FLIGHT TEST PROGRA. 187 Ernc.st R. Hillje and Raymond L. Nelson V Provided by IHSNot for ResaleNo re

10、production or networking permitted without license from IHS-,-,-DEVELOPMENT OF SPACE SHUTTLE ZCIWCION OVEKPKESSURE ENVKROWENT Ah9 CORRELArION NLTH FLIGIIT 24T4 7 i 9 S. Lei JVTYY 4ERODY NAMI CS I Chairpersons: Jamcs C. Young, NASA Johnson Space Center Bernard Spencer, Jr., NASA Langley Researh Cen;e

11、r SPACE SHUTTLE ENTRY L0NGIrUl)iYAL AERODYNAMIC COMPARISONS OF -, c, , FLIGHTS 1-4 WITY PREFI,IWT PEDICTtOYS.,. , Paul 0. Remere aqd 4. Viles $!hitnah A REVIEW OF PREFLIGHT !:STIYkTES OF REAL-GAS EFFECTS ON SPACE W. C. Woods, 1. 7. A;ri:igtoii, and !i. E. Yamilton I1 ! I1 SHUTTLE AERODYKAYI? CH4R4CT

12、EIIISPICS EXPLANATION OF THE HYFERSOXIC LOiTGITUDIN4L STABILITY PKORLLPI - ), LESSONS LEARNEn. B. J. GriEfj.th, .T. K. V.I-S, iind .J. T. Best SPACE SHUTTLE ORBITER P.EACTLW comm SUBSYSTEM FLIGHT DAT 4 -57; ANOMALIES J. S. Stone, J. -1. 5aumnis-!i, and R. P. Roberts ,- ANALYSIS OF SHCTTLE OSCILLA:IO

13、,u IN THE PlACH NUMBER = 1.7 TZ MACH NUMBER = !.O YAGE,. William T. Suit, Haro!d R. Csmptor., William 1. Scal!ion, James R. Schiess, axi L. 5cs. Gahnn APPROACH TO ESTABLISHING WE Emcr OF AEROELASTICITY ON A:)L:Y;.YI.- CHARACTERISPICS OF 1BE StAC5 SilUTTLS ORBICLR., _I . D. C. Schlosser sild 3. I;. E

14、oitlinlk EP haust sirnulation (i.e., plume simulation) is determined by wind tunnel testing. basc of substantial scope. The results fell short of t.he target, although work coilductet:. was conclusive and advanced the state of the art. Comparisons of wind tunnel psedictions with Space Transportation

15、 System (STS) flight data showed c.orisIderzbie differences. yield(2u an additional parameter that may correlate flight aiid cold gas test data. Data iruci tiic plume teclinology program and the NASA test. flights are presented to suhstmti ate the proposed simulation parameters. Cold 935 testing was

16、 concluded to be a cost - and scticdulr-effective data However) a review of the technology program data base has INTRODUCTION The wid tunnel simulation of exhaust plume effects on the aerodynamics of rocke:-piwered launch vehicles has historically been accomplished by using cold gases (usually, unhe

17、ated alr). Although accurate simulation with hot gases is current state of the art, the cost and schedule impacts are one tc WO orders of magnitude greater t.han for test:ing with cold gases. In addition, data quality for hot gas testing i.s limited extensively because of the short duratior. of stea

18、dy-state flow (10 - 100 m/secj. Thus, the choice to be made was betweer: hot gas simulation, a costl:!, .ow quality data base of reduced scope, and cold gas simulation, a cost - and sc!icdiile-.etfcctiii. data base of substantial scope. Cold ?,as testing was the prefe.rrecl choi.ce by a wide margin,

19、 even though the scaling parameters required to make 01.ci Sas simulate hot gas are not well understood. iiocket exhaust extensively affects the base drag of a launch vehicle. For desigii purposes, thc effects are determined by wind tunnel testing. The following factors must be considered for any ae

20、rodynamic test: 1. Geometrically scaled model 3. Fr!Je-st-ieani ?!ach number Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,- 3. Boundary layer development (Reynolds number) However, if the rocket exhaust gases are to be simulated as well, additional

21、 The plume diameter is initially too small to factors must be considered. engine chamber pressure (Figure 1). significantly alter the forebody pressure. Thus, the primary effect is the entrainment of the base flow by the high-velocity gases in the boundary of the plcrnt and the subsequent reduction

22、of power-off base pressure. As the plume grows in sizz, it begins to block the base and increase the base pressure. Ultimately, the boundary layer will separate, and a recirculating pattern will develop. For multiple engines, the plumes will impinge upon each other and deflect exhaust flow into the

23、base. Three or more engines can reverse enough mass into the base to choke the volume enclosed by the engines. The effect of the plumes can actually increase base pressure above the power-off level. The exhaust plume phenomena vary with increasing rockct The following design options are available fo

24、r use in plume simulation: 1. Hot gas, by combustion 2. Cold or warmiheated gas 3. Solid-body simulator Hot gas testing can be eliminated as a viable option when cost and complexity are considered. Short-duration techniques (detonation/shock tubes or small solid-propellant wafers) are required to ge

25、nerate the hot gas, and only three to five data points may be obtained for each shift in the test facility. Short-duration pressure data are alvays of lower quality than continuous pressure data. In addition, specialized support personnel are needed to implement the short-duration techniques require

26、d for hot gas testing. A hot gas model costs 3 to 10 times more than a cold gas model. The use of a solid-body simulator can also be eliminated from consideration. The base environment is not known before the testing. Therefore, the configuration of the plume shape cannot be determined to enable des

27、ign of the solid body. addition, a solid-body configuration cannot respond to changes in angle of attack OK sideslip; and finally, entrainment of the free-stream flow and aspiration of the base cannot be simulated. In Cold gas testing is used almost exclusively for launch vehicle plume simulaticc. A

28、 cold gas model can be operated continuously to obtain 70 to 100 data points per shift in the test facility. Therefore, the Space Shuttle Program chose this technique to determine launch vehicle plume effects because of cost and schedule effectiveness . This paper discusses the development of the te

29、chnology correlation techniques used to define the plume simulation parameters and the impact of the flight data on this technology program. 2 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-BACKGROUND In 1972, NASA initiated the planning phase for t

30、he first wind tunnel test of the Space Shut:tle Launch Vehicle (SSLV). At this time, the technical archives were r;urvey.d t:o determine the appropriate rocket exhaust simulation techniques. Those data acuniulated through experience with the Saturn launch vehicle were chosen for study. indicated a d

31、eficiency in the technology at that time. The base drag was r;ubsta:itially overestimated by the predictions from wind tunnel testing (Figure 2). Finall,y, t:he surveys concluded that the simulation techniques and the simulation parameters were not well understood. Therefore, in planning the testing

32、 for the SSLV, the following approach was adopted. A comparison of the wind tunnel predictions with the Saturn flight data 1. For the initial SSLV test, model nozzles were designed and test conditions were chosen such that the cold gas plume shape matched the analytical estimates of flight plume sha

33、pe. 2. Simultaneously, a technology program was initiated to enable understanding of the flow phenomena and develop a set of simulation parameters. 3. These findings were input into subsequent SSLV tests. The current status of the technology program is best described as “termi.iatcd incomplete.“ The

34、 technology program (Figure 3) yielded substantial knowledge on how to correlate the cold gas base pressure. The simultaneous evolvement of the technology program and the SSLV program are shown in Table I. Note that the last output from the technology program did not feed into the SSLV test prcgrain

35、, i circumstance of the technology learning curve and the SSLV test program timing. Therefore, only an assessment of the latest technology could be made at that time. The result was a substantial difference between the predicted and the actual flight SSLV base pressure. TECHNOLOGY PROGRAM The object

36、ive of the technology program was to determine a set of functions that would rorrelate base pressure data generated by wind tunnel cold gas tests with full- scale j-light base pressure. A substantial empirical data base was obtained using generic models with some geometry variations to assess config

37、uration effects on the base pressure. The key independent variables were simulated gas, nozzle geometry, and geometric configuration. Hot, warm, and cold gases were used. In most cold and warm gas technology tests, air was the simulated gas; in some instances, however, Freon 1:CF4) was used because

38、its variation of the ratio of specific heats through the exhausl: plume was similar to a prototype full-scale rocket engine. The hot gases for the technology tests were generated by burning solid propellant charges in the models. Simulated model nozzle area ratios and nozzle lip angles varied from t

39、est to test, assuring that internal geometry was not an explicit contributor to the correlation functions. The external configurations (Figure 3) consisted of: 1, Cone or ogive noses and cylindrical afterbodies with single or triple nozzle bases. 3 Provided by IHSNot for ResaleNo reproduction or net

40、working permitted without license from IHS-,-,-2. A triple-body configuration, used to assess the effects on the center body (similar to the External Tank on the Space Shuttle). Difficulties were encountered in correlating the plume technology test data because of limited variations in nozzle geomet

41、ry and test conditions. Therefore, the decision was made to supplement the data base with analysis. The Addy program was used as a controlled experiment to generate additional data. The substantial empirical and analytical data base generated throughout this technology program was then analyzed for

42、correlation by plotting the base pressure data as a function of reasonable candidate simulation parameters. The successful simulation parameters were those that would coalesce the base pressure data to a simple function of the assumed simulation parameter. of parameters is shown in Figure 4. These p

43、arameters are defined in Figure 5. An example of a “winning“ set The results from the Addy program (Figure 6) demonstrate the failure of the initial expansion angle, 6 j, to correlate the data. number, M is introduced into the simulation parameter, the correlation is remarkably improved (Figure 6(b)

44、. have the same one-dimensional exit Mach number, Mex. data are once again uncorrelated when M is allowed to vary. Obviously, the correlation parameter must also contain M as a variable. This approach was continued until it became apparent that the simulation had the form If the plume boundary Mach

45、jy However, all cases shown in these two figures Notice in Figure 6(c) how the ex ex This knowledge was then applied to the technology program test data. Proceeding in a similar manner with the technology test data, the results (Figure 7a) show only a fair correlation with 6 The results plotted in F

46、igure 7(b) show that M.6. correlates the data with the same exit Mach number. The effect of the JJ exponent or y of this exponent results in the excellent correlation quality previously shown in Figure 4. j* equal to unity is shown in Figure 7(c). Assigning 0.5 as the value j The final result of thi

47、s analysis was the preparation of a preliminary table of simulation parameters (Table 11) for the SSLV. The caveat, however, is that neither the hot gas technology test nor the hot gas analytical data from the Addy program agree with the SSLV parameters in Table 11. Much of the hot gas test data was

48、 of questionable quality and had to be discarded; however, the few data points available show a definite offset from the cold and warm gas data (Figure SI. program (Figure 9). fully correlate the data. testing; furthermore, the data points were too spotty to extrapolate to all Mach numbers and base

49、configurations. This effect was substantiated with the Addy Obviously, an additional function of (T IT These data came too late in the program to impact any SSLV IC is required to c t- 4 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-A temperature correction was not incorporated in th