Apollo Experience Report the Development of Design-Loads Criteria, Methods, And Operational Procedures for Pre Launch, Lift-Off, And Midboost Conditions

8/8/2019 Apollo Experience Report the Development of Design-Loads Criteria, Methods, And Operational Procedures for Pre L…

http://slidepdf.com/reader/full/apollo-experience-report-the-development-of-design-loads-criteria-methods 1/18

E /73 -dW 7 5

N A S A T E C H N I C A L N O T E NASA TN 0-7373

mhM

C

APOLLO EXPERIENCE REPORT -

THE DEVELOPMENT OF DESIGN-LOADS

CRITERIA, METHODS, AND OPERATIONAL

A N D MIDBOOST CONDITIONS

PROCEDURES FOR PRELAUNCH, LIFT-OFF,

by Alden C . Mnckey and Robert D, chwartz

Lyndon B. Johnson Space Center

Houston, Texas 77058

N A T I O N A L A E R O N A U T IC S A N D S PA CE A D M I N I S T R A T I O N W A S H I N G T O N , D. C. A U G U S T 1973

-



~

1. Reporr No.

NASA TN D-7373

11 . Cont rac t or Grant No.Lyndon B. Johnson Space CenterHouston, Texas 77058

13 . Type of Report and Per iod Covered

Technical Note. Sponsoring Agency Name and Address

14 . Sponsoring Agency CodeNational Aeronautics and Space Administration

Washington, D. C . 20546

The JSC Direc tor ha s waived the use of the International System of Units (SI) fo r this ApolloExperience Report, because, in his judgment, the use of SI units would impa ir the usefulnessof the report or result i n excessive cost.

5. Supplementary Notes

6. Abstract

2. Government Accession No. 3. Recipient's Catalog No.

In this repo rt , the prela unch, lift-off, and midboost conditions that a r e important to the designof spacecraft a r e described. Original Apollo design concepts that were deficient in reali sticand accu rate analysis of spacecraft struct ural loads in a wind environment als o a r e included.

Important improvements in design cri te ria , mathematical models, wind-monitoring proce dures,and loads analyses a r e discussed. Recommendations fo r future progra ms a r e presented.

4 . "1 and Subt it le

rHE DEVELOPMENT O F DESIGN-LOADS CRITERIA, METHODS,

APOLLO EXPERIENCE REPORT

IND OPERATIONAL PROCEDURES FOR PRELAUNCH, LIFT-OFF,

IND MIDBOOST CONDITIONS

7. Aut hor (s )

Alden C. Mackey and Robert D. Schwartz , JSC

-9. Performing Oryanirat ion Name and Address

5. Report Date

August 197 3

6. Performing Organizat ion Code

8. Performing Organizat ion Report No.

JSC S-342

10. Work Un i t No .

914-50-20- 14-72

17. Key Words (Suggested by Author(s ))

* Apollo Spacecra ft Loads

* Spacecraft Design Launching

'Winds Aloft -Struc tural Design Launch Vehicle

Ground Wind

18. Dis t r ibut ion Statement

19 . Security Classif . (o f this report) 20. Security Classif. (of this pagel 21 . NO . of Pages

None None 18

22. Price

$3.00





auxiliary damper is attached to the vehicle f ro m the launch umbilical tower (LUT) to

reduce dynamic respons es of the vehicle during its exposure to the ground winds. Thedamper is removed approximately 4 hours before launch. During the prelaunch condi-tions, winds a r e the only fo rc e that can cau se significant response s i n the vehicle.

Before lift-off, the vehicle is responding to winds plus unsymmetr ical thrust

buildup and thrust misalinement forces. These force s caus e stati c and dynamic vehicledeflections and re su lt in large constraining loads i n the holddown st ru ct ure. At lift-off,the sudden re le as e of these constraining loads induces la te ra l and longitudinal str uc tu ra loscillations and dynamic loads i n the spa ce vehicle.

The winds aloft contribute significantly to the vehicle loading only during first-sta ge boost. The major wind loading occ urs during the per iod of high dynamic pressu re ;therefore, even though high winds may occur la te r during boost, they do not affect ve-hicle loading significantly.

The purpose of th is report is to document the development of the Apollo spacecraftdesign-loads cr it er ia , methods, and operational procedures f or the conditions in which

wind is a fac tor. The discussion includes the design cr it er ia and methods shortcomings,their impact on spacecraft design and operations, and recommendations for futureprograms.

During prelaunch, launch, and boost, the Apollo spacecraft loads are directly re -lated to the loading environment of the enti re space vehicle. Therefo re, the so urce s ofloading on both the launch vehicle and the spacecraf t are emphasized in this re port .However, it is the Apollo Prog ram impact of the spacecraft loads, not the launch vehi-cle loads, that form s the basis for the conclusions and recommendations presentedherein.

PRELAUNCH

The prelaunch phase is the period f ro m the final joining of the Apollo spacecraftwith the launch vehicle to the ignition of the fi rs t- st ag e main engines. Concern fo r de-sign wind loads begins when the Saturn V vehicle leaves the VAB or when the Saturn 1Bservice structure is removed.

The original design cri te ri a included the requirement that the freestanding(laterally unrestr ained) launch vehicle, with unpressurized tanks, must withstand99.9-percentile winds of the most se ve re wind month. The 99.9 -percen tile winds are

. those that would be exceeded only 0 .1 percent of the t ime during the worst wind month.

For the 99.9-percentile steady-state wind profile (table I and ref. l ) , the wind veloc-ities are referenced to height above the mean ground level.

The definition of the wind environment used for vehicle design did not include sev-

e r a l important effects critic al to vehicle loading. These effects we re wind-exposureperiods, variation in gust fa ct or s, variation in wind profiles , wind directionality, andwind turbulence.

2



TABLE I . - SURFACE-WINDSPEED ENVELOPE 99.9 PERCENTILE

FOR EASTERN TEST RANGE^

Height abovenatural grade

m

3.0

9.1

18.3

30.5

61.0

91.4

121.9

152.4

ft

10

30

60

100

200

300

400

500

Quasi-steady-statewindspeed

m/sec

11.8

14.8

16.9

18.8

21.6

23.4

24.7

25.9

knots

23.0

28.7

32.9

36.5

41.9

45.4

48.1

50.3

bPeak windspeed

m/sec

16.6

20.7

23.7

26.3

30.2

32.7

34.6

36.2

%ind velocity distr ibution with height is

where V = wind velocity at des ire d height

= wind velocity at reference height'Ref

h = altitude at which wind velocity is desired

= reference altitudehRef

bPeak wind velocity is 1.4 time s steady-st ate wind velocity

knots

32.2

40.2 *

46.1

51.1

58.7

63.6

67.3

70.4

3



The 99.9-percentile steady-state wind profile used in the original design wasbased on the assumption that the space vehicle would be exposed to the ground-wind en-vironment f or only a f ew hours; that is, the vehicle would be transported fr om the VABto the launch pad and launched within a f ew hours. However, most Apollo/Saturn vehi-cl es were exposed to the launch pad environment for s everal months. A real isti c ex-pos ure period would have resulted i n higher magnitudes fo r the 99.9-percentile winds;hence, the original cr it er ia were not adequate. The exposure-period definition was themost important deficiency in the ground-wind environment, but the other l isted defi-

ciencies also were significant.

To evaluate the complex effects that contribute to the definition of the winds in theboundary laye r of the atmosphere, a 150-meter meteorological tower facility was con-structed in the Mer ritt Island Launch Area at the NASA John F. Kennedy Space Center(KSC) in Florida. A complete description of the tower is given in reference 2. Dataobtai’ned from thi s meteorological facility a r e useful for the study of the ch arac te ri st ic sof the lower atmosphere wind and temperature profil es as a function of height and time,gust-shape phenomena, turbulence spec tr a, correlation propert ies of steady-state andturbulent flow, and calculations of diffusion pa ra me te rs that define the environment towhich the space vehicles would be exposed during the prelaunch period.

Although the environment definition used in ground-wind design c r it er ia w a s orig-inally deficient in several ar ea s, it was later improved. A complete description of thepresent definition is given in reference 3.

As wind flows by a cylinder, the wake behind the cylinder is disturbed and is char-acte rized by distinct vortexes, as shown i n figu re 1. Clockwise and counterclockwisevortexes ar e shed in an alternate regu lar manner. The Von Karman vortex sheddingcauses a harm-n~ica!!;~ ~ a r y i n g sr ce oii the cyiirider normai to tne wind direction

(r ef . 4). Originally, vortex-shedding load for design loads was assumed to be1.25 times the drag load resulting fro m steady-state winds. In 1963, a wind-tunnel

test program w a s initiated to define the dynamic response of the Apollo/Saturn vehicles ,particularly the vortex-shedding response . Two separa te series of t es ts were run onthe freestanding Apollo/Saturn IB and Saturn V vehicles. Vehicle response to vortexshedding w a s identified as a potential problem, and fur ther t es ts were planned.

Ft

L

F D = Drag force

F = Li f t forceL

Figure 1.- Vortexes shedding in

a wake.

In the fall of 1964, a third series ofwind-tunnel te st s was initiated on the Apollo/Saturn V configuration. The t es t model in-cluded the space vehicle and the LUT but notthe mobile servi ce stru ctur e. Testing w a scompleted on empty, intermediate, and fullyfueled vehicles. During the te st s, the

vortex-shedding response w a s identified asa critical launch vehicle loading conditionfor empty and intermediate propellantweights. The change in vehicle propellantweights al te rs the elastic vibrational char-ac te ri st ic s of the vehicle, which in turn

affect the vehicle vortex-shedding response.Several potential fixes, such as helical

4



st rak es, we re evaluated on the test model. These tested fixes, which would have al-te re d the aerodynamic configuration, were rejected, and an auxiliary damper was r ec -ommended. A fourth se r ie s of wind-tunnel te st s was completed i n June 1966. The te stmodel was modified to include the effe cts of the mobile s ervi ce st ruct ur e, and testingwas performed for t he empty and intermediate propellant conditions. The re sult s from

these t es ts we re used to develop a viscous damping sys tem that was inst alled on subse-quent Apollo/Saturn V operational vehicles. This damping sys tem was attached betweenthe LUT and the command module (CM) launch escape syst em tower (fig. 2).

As a resu lt of the wind-tunnel te st s, the original 1.25 vortex-shedding factor for

design was changed. Fo r vehicle ground operations with the damping sys tem attachedand unattached, the vortex-shedding load w a s assumed to be, respectively, 1.6 and3.6 tim es the drag load resulting fr om steady-state winds.

The f inal wind-tunnel test on the Apollo/Saturn V configurations w a s completedduring June 1967. The purpose of the te st was to evaluate ground-wind boundary layer

effects on the dynamic response of the vehic le. The te st indicated that ground-wind

boundary layer effects could be ignored in the prediction of dynamic response fo r theground-wind loads condition. Pr im ar il y, this insensitivity to boundary laye r effects iscaused by the s mal l vari ation along the vehicle of the design wind magnitude.

Figure 2 . - Viscous damper for

Apollo/Saturn V .

The third and fina l se r ies of wind-

tunnel te s ts on the Apollo/Saturn IB vehiclewas completed during the spr ing of 1965.In this testing , numerous wind azimuthswere investigated fo r full and empty propel-lant conditions and for empty conditions (themore se ve re weight condition). No criticalcondition was identified during the Apollo/

Saturn IB test; hence, no further testingwas required.

The high dynamic response s measuredi n the Apollo/Saturn V wind-tunnel testscaused concern about the abil ity of the vehi-

cle to withstand the on-pad environment. Ina n effort to verify the re sul ts f ro m the wind-tunnel tests, ground-wind tests were initi-

ated using a full-scale Apollo/Saturn Vfacilities verification vehicle. The objec- ’

tives of the tes ts were to evaluate the r e-sponse of the Apollo/Saturn V duringtransportation f ro m the VAB to LaunchComplex 39 and to measure vehicle responseduring exposure to launch pad environment

f o r long time periods. The pad 39 te st ssta rted on May 25, 1966, and were com-pleted on October 14, 1966. During thetesting period, the vehicle w a s returned tothe VAB (June 8 to 10, 1966) fo r protectionfrom a hurricane that passed through the

5



ar ea . Data from the tes ts were inconclusive in defining the vehicle response to groundwinds. High-magnitude winds were exper ienced only during the time of the hurric ane ,but no loads or accele rations were recorded during that period because of an instrumen-tation syst ems failure. Also, to preclude the possibility of a structural failure, adamper was attached from the LUT to sta tion 2560 in the S-II/S-IVB interstage. Withthe addition of t his damper, overall s tr uc tu ra l damping fo r the vehicle was calculatedto be 5.5 percent compared with the assumed 1percent fo r the vehicle s tru ctu re with-

out auxiliary damping. Dynamic response of the vehicle was re st ri ct ed severely bydamping of this magnitude, and the Von Karman vortex-shedding effect was never de-tected during the course of the test.

.Data f rom wind-tunnel and full -scale tests did re sul t in the viscous damping s ys-

tem being installed to reduce vehicle dynamic loads, but the data were available toolate fo r inclusion in space vehicle design. The installa tion of the viscous damping sys -te m eliminated the high dynamic responses to winds and resulted i n prelaunch loads be-low the struct ura l capability of the space vehicle. Two significant lessons that apply

to ground-wind environment and loads definition were learned f ro m the Apollo/Saturnprogram. First, an accura te definition of the vehicle exposure to the launch pad en-vironment is required . Second, wind-tunnel tes ting for ground winds should be timely

so that the r es ul ts can be considered in the vehicle design.

LAUNCH

The launch period is the time fr om the initiation of f irst -sta ge start sequence un-ti l all structur al response transi ents following vehicle rel ea se have damped out. Thest ruct ur al response and accompanyir.g inertia! loads zre caused primariiy by a lift-offphenomenon often called "twang. '' Twang is the vehicle dynamic response to the suddenre le as e at the holddown ar m s of the constraining shea r, moment, and axial forces.These holddown for ces are caused by the ground winds and thrust buildup before release.Therefore, the calculation of design lift-off loads is dependent on an ac curate definitionof the wind- and propulsion-induced loading environments and an accurate mathematicalmodel describing the st ru ct ural response. The advent of the la rge, flexible Apollo/Saturn IB and Saturn V vehicles was a unique opportunity to advance the s ta te of the a r tin calculating lift-off loads; however, in the calculation pro ce ss, se ver al deficienciesin design philosophy became evident.

Wind- I duced Loadi n g

The original launch wind cr it er ia specified that the Apollo spacecraft should be? designed for the launch of a Saturn IB or Saturn V exposed to the 95-percent-probability-

level winds (table 11). The peak wind velocities, gust prof ile, and vortex-shedding ef-fects were established in a manner si mil ar to those described in the prelaunch section.Late ral drag forc es caused by the design wind profile were applied omnidirectionally tothe vehicle throughout the design and design verification phases of the Apollo Program.

The loading caused by Von Karman vortex shedding w a s considered in the originalcrit erion to be 1.25 times the steady-state wind loading acting perpendicular to the wind.

During the design veri fication phase, wind-tunnel test data we re used to determine the

6



TABLE II.- SURFACE-WINDSPEED ENVELOPE 95 PERCENTILE

FOR EASTERN TEST RANGE^

Height above naturalgrade, ft

10

30

6 0

100

200

300

400

500

Steady-state windspeed,ft/sec

2 3 . 6

2 9 . 4

3 3 . 8

3 7 . 5

4 3 . 1

4 6 . 6

4 9 . 5

5 1 . 7

bPea k windspeed,

ft/sec

%ind velocity dist ribution with height is

where V = wind velocity at desi red height

VRef = wind velocity at reference height

h = altitude at which wind velocity is desired

= reference altitudehRef

bPeak wind velocity is 1 . 4 ti me s steady-state wind velocity.

3 3 . 1

4 1 . 2

4 7 . 3

5 2 . 5

6 0 . 3

6 5 . 2

6 9 . 2

7 1 . 8

*

7



Apollo/Saturn IB and Saturn V vortex-shedding character isti cs and to provide betterpredictions than the approximations in the original cri te ri a. The improved vortex-shedding predictions were included in the calculation of design verification loads.

Three major deficiencies in the Apollo wind criteria development became evident.First, although the design wind prof iles were based on numerous windspeed time his-to ri es , exposure time stat isti cs were not considered throughout the design and designverification phases. Had a 1-hour period of exposure to the winds been conside red, the

95-percentile design peak wind envelope would have been approximately 20 percenthigher than that of table 11. Second, although the vortex-shedding approximations we re

significant in magnitude but were known to be crude approximations, a wind-tunnel pro-gram'to provide a better definition of the unsteady effect s commenced too late fo r sup-port of the design phase. Finally, wind direction, which is particularly important i ndefinjing the vortex-shedding fac tor , was not considered. For example, although the

wind-tunnel data indicated a much la rg er vortex-shedding factor than the 1.25 designfact or, the unsteady effect ex ists fo r only a narrow range of wind direc tions becausethe LUT inte rfere s with the wind. It is apparent that exposure time , vortex shedding,and wind direction have varying effec ts on the vehicle loads a t lift-off that we re not con-

sidered in the design. As a re su lt , the probability of ac tually experiencing design windconditions at lift-off was unknown. The combined effects of wind direction, exposuretime, and vortex shedding were not investigated until a NASA Lyndon B. Johnson Space

Center (JSC), formerly the Manned Spacecraft Center (MSC), Monte Carlo statisticallift-off loads analysis w a s conducted late in the design verification phase.

Because of the shortcomings in design cr it er ia and data , the need to develop meth-ods to res tri ct the vehicle operations in a se ve re wind environment w a s evident. There-fore, an effective wind-monitoring procedure was sought so that the Apollo/Sa!xrn IBand Szb.~r::V vehicles couici be operated efficiently and safely during lift-off. The orig-ina l system used anemometers 60 feet above ground level to establish wind measure-ment limits. Wind distribution and vehicle dynamic effects could not be detected from

these wind measurements. Therefore, design-level vortex shedding and gust dynamicswere , conservatively, always assumed to be present. Because measurable vehicle dy-

namics caused by vortex shedding had never been detected on the Apollo/Saturn IB orSaturn V vehicles, this launch wind procedure resulted i n an undesirably se ve re re-striction. Several vehicle-response-monitoring syst ems that include the us e of sensi -tive guidance and control velocity change measurements, ground-based opticalmeasurements, and vehicle linear acce le ro me te rs we re investigated. Not one of thesesy st ems was found to be practical. The procedure finally was improved when Saturn IBload measurements at the S-IB/S-IVB inte rface and Saturn V load measurements at theOS-IC intertank region were calibrated to r ead out bending moment while the vehicle wasstill on the pad. The measurements and calibrations were checked by a vehicle canti-

levered calibration te st . Although load measurement s at the holddown st ru ct ure would

have indicated more accurately the enti re loading on the vehicle, the interface measure-ments provided satisfactory information indicating the presence of vehicle dynamicscaused by gusts and vortex shedding. As a res ult , the load measurement became thebackup to the wind measurement in the wind-monitoring sy st em fo r Apollo 7 and 8 andbecame the primary mode, with wind measurement serv ing as a backup, for the Apollo 9

and subsequent vehicles. The new procedure provided gr ea te r, more real isti c opera-tional capability; consequently, load mea surement ha s been recommended as the pri-mary wind-monitoring procedure f or future progr ams .

8



A subsequent wind-monitoring innovation was a change in the operational refere nce

height from the 60-foot anemometers to a 530-foot anemometer atop the LUT. Thischange was made to obtain measurements more representat ive of the winds influencingthe vehicle. The new refer enc e height for wind monitoring was first used operationallyfor the Apollo 14 mission.

Propulsion-l nduced Loading

The importance of thrust buildup dynamics to the vehicle design for launch releasewas recognized early in the Apollo Progr am. Unfortunately, although the cr it er ia Spec-

ified that th rus t buildup was to be considered in the calculation of lift-off loads, nomeasured data were available fo r thrust misalinement o r unsymmetrical thrust buildupin the initial Block I analysis. However, th rust misalinement effects were establishedbefore the first calculat ion of Block I1 loads. A stat isti cal analysis indicated that lateralloads caused by booster engine misalinement would approximate 5 percent of the ground-wind loads; therefore, fo r convenience, the ground-wind loads we re increa sed by 5 pe r-

cent to account fo r this effect. The portion of unsymmetrical thrus t buildup basemoment caused by off-nominal engine ignition times w a s not considered in the l ate ralloads until the design verification phase.

Although off-nominal effects were not included, the effects of the planned stag-ge re d engine ignition sequence we re included in the thrust buildup cr it er ia for axialloads calculations. The origina l Saturn IB engine start sequence was planned so thatpa i rs of diametr ical ly opposed engines would fi re simultaneously with a stagger time of0. 1 second between pa ir s. The Saturn V engine start sequence w a s defined so that the

center engine would ignite first, followed by opposing pa i r s of engines with a0.25-second stagger t ime between the center engine and each of the pa ir s.

Data provided by propulsion t es ts led to design verification t hrus t buildup c ri te ri aconsiderably different f rom those of the design phase. A new Saturn V F-1 engine starttime sequence w a s established with a 0.3-second stagger time between engine pai rs in-stead of the 0.25-second stagger time used i n the design phase. The cr it er ia defined

the F-1 engine start time as the time required for the thrust t o reach 90 000 pounds.The 30 dispersion on start time of one F-1 engine was ? 0.233 second. The Saturn IBstart sequence w a s unchanged i n the design verification phase. After propulsion t es ts

wer e evaluated, revised Saturn V and Saturn IB thrust buildup envelopes and the dis-

pers ions defining the unsymmet rical thrust buildup charac te ri st ic s were establishedand included i n the Apollo design verification. In addition, stage alinement t e st s pro-vided data to establish a Saturn V 3a dispersion i n thru st vector alinement, which was

used t o calculate design verification loads.

.

Several additional effects of propulsion-induced loading not included in the designphase were noted and evaluated during the Apollo Prog ra m. Fo r example, a Saturn V

failure that would permit the F-1 engines to st ar t out of thei r normal sequence waspossible. Th is type of fa il ur e could add significantly to the unsymmet rical thrust fo rc esat the vehicle base. An analysis indicated that the Saturn V start-sequence failure couldincr ease the spacecraft lateral structural loads by approximately 4 percent. Another

propulsion effect not considered i n the original lift-off design was that of control -system-directed engine gimbaling. The Saturn V control system is active from ignition

9



through release and, in fact, is used to direct the vehicle in a tower-avoidance maneu-ver shortly after re le as e. The effects of gimbaling did not add significantly to struc-tu ra l lift-off loads but should have been considered i n the design.

Mathematical Modeling

The original s truc tu ra l mathematical models fo r lift-off were based on lateralbeam models and axial spr ing mass models of both the Saturn IB and Saturn V fromwhich uncoupled bending and axial modes we re computed. The prel iminary Block Ispacecraft model included an LM that was assumed to be rigidly attached to the SLA.A sep'arate model used fo r Block I loads included the LM axially connected to the SLAby a spring but laterally rigid. This spacecraft-loads model formed the ba si s f o r the

preli$inary Block I1 loads. Later in the design phase, the models we re modified toinclude an LM represen ted by a mas s axially and lateral ly connected to the SLA bysprings. However, the lift-off condition requ ired mo re complex dynamic models thanthe Saturn IB and Saturn V beam models used for design. Accurate loads were not cal-culated until three-dimensional models were developed late in the design verificat ion

phase. The two-dimensional mathematical models were not sufficient, fo r example,to predict accurately the relative torsional motion of the CM with respect to the servicemodule (SM). A s a resu lt, the torsion loads used for early design were based on rigid-body roll accelerations. La te r, dynamic te st s of the Apollo/Saturn V wer e made usingthe dynamic t es t vehicle at the NASA George C. Marshall Space Flight Center (MSFC).These tests indicated considerable axial-torsional coupling at the CM/SM interface, s omore accurate torsion loads were calculated.

Although two-dimensional la ter al and axial I_oad,cwere zompiiied ana documented

of the Apollo/Saturn V was developed lat er , and loads we re updated. The Apollo 9 andsubsequent mission compatibility loads were calculated with the three-dimensional

models.

qn.. th qA ~ ~zaigiil------ verification phase, a 180-degree-of-freedom, three-dimensional model

In addition to the problems created by cr it er ia omissions and uncoupled models,the mathematical models used to calculate the design loads did not d escri be realist icallythe actual lift-off condition. Design wind prof iles were applied without re ga rd to direc-tion or exposuretime stat ist ics . Wind- and propulsion-induced forcing functions were

derived by root sum squaring worst-case inputs. Axial loads were calculated usingforcing functions different from lateral loads. The lateral loads were calculated in thepitch and yaw planes, and the maximum loads we re root sum squared. Therefore, al-*though a 95-percentile wind profile was used in the design, the method of root su m'squa ring inputs and outputs resulted i n design loads that had no valid stati stical basis.The probability of exceedance w a s not known until late i n the design verificat ion phase

when the Monte Car lo lift-off loads analysi s discussed earlier was completed (ref. 5).

A Monte Carlo solution is one in which the input data fo r many ca se s a r e selectedat random from probability distributions. Outputs then a r e analyzed statistica lly byconstructing probability distributions from the r es ul ts of all the cas es. The analysiswas performed primarily to establish the feasibility of a Monte Carlo approach to lift-off loads. However, the analysis al so indicated the conservative nature of the orig inaland final Apollo design methods. A single, important load, the maximum bending mo-ment at the CM/SM interface, was chosen as an example fo r the Monte Carlo analysis.

10



A moment obtained by the Apollo design method of root s um squaring the maximum mo-ments from thr ee ca se s was compared with the Monte Carlo res ult . Each of the th re eca se s was constructed fr om the selection of 3 0 wind, unsymmetrical thrust, and thru stmisalinement forcing functions. Probability distributions of these data we re prepared,and 200 independent s et s of forcing functions were selec ted fo r the Monte Carlo analy-sis. A compilation of the maximum CM/SM interface moment f ro m all cases estab-

lished a log-normal dist ribution. The CM/SM bending moment obtained by theroot-sum-square method (the design moment) was never exceeded in the 200 MonteCarlo cases and would not be expected to be exceeded at least 99.99 percent of the time.This example indicates the potentially conservative nature of a lift-off design load ob-tained by the root sum sq uar e of sev era l discrete ca se s derived from worst-case in-put forcing functions.

Another load sou rc e overlooked in the initial Apollo design phase was that resul t-ing fro m a conditioning and inert ing purge of the S M and SLA. The purge in cr ea se s the

SM and SLA internal pr es su re and, therefore, increases the CM/SM interfac e tensionloads at lift-off. The purge pr es su re w a s reflected in the s tru ctu ral capability analysis.

INFLIGHT W I N D RESPONSE

The inflight wind response period extends fr om the time the lift-off transients

damp out until the end of first -sta ge boost. The original design cr it er ia specified thatthe vehicle must withstand the winds associated with the 95-percentile idealized scalarwindspeed profile envelope, windspeed changes (wind shear) fo r altitude dif ferentials to10 000 feet (the 99-percenti le windspeed changes), and a discrete quasi-square-wavegust (29.5 ft/sec) superimposed on the windspeed change so that the maximum wind-speed did not exceed the idealized windspeed profile. A wind profile constructed to

these criteria is shown i n figure 3 . These wind cr it er ia wer e to be applied to the vehi-cle f rom any horizontal direction during boost.

Vehicle loads for this wind condition were determined using constant aerodynamicand inert ia l coefficients and included the effects of two elastic modes. An evaluationwas made of the inte rpretat ion of the inflight wind-condition c r i te r ia, and an agreementwas reached that the applied gust should be in excess of the 95-percentile windspeedenvelope. When th is change was implemented and its severity evaluated, a statisticalcombination of the wind shea r and wind gust was applied. The effect of this s ta ti st icalcombination was approximated by a 15-percent reduction of the wind shears and wind-

gust magnitude. The re su lt of this change also is indicated in figure 3 .

Inflight wind-loads cr i te r ia and analysis did not include the effects of turbulence,propellant slosh, wind direction, thrust oscillations, loads-minimization methodsthrough trajectory shaping, gust penetra tion, higher frequency modes, and time-varyingcoefficients. Sev eral of these effects significantly influence loads analysis . The mostimportant ef fect s not included in the design are wind directionali ty and biasing of the

command-pitch p rogram i n operational trajec tori es to reduce space vehicle loads.Only the first Apollo miss ion did not involve the use of wind biasing for load relief.For the first mission, the pitch prog ram was biased to induce loads into the space ve-hicle and to te st the vehicle structure. Subsequent to this mission, the decision w a smade to take advantage of the biasing technique and to at tempt to reduce th e inflight

11



Wind condition 1

S a t u r n P I S L A I

I I6-mlsec

wind velocity -c( II I

I -II II I-- - % I I

-4 -I

75-mlsec

,%wnd velocity\

--- %%%

- - \ I-%% I

99-percent

9-mlsec aust :

speed profile

88-percen t wind-

speed profile

Wind condit ion 2

Satu rn Y ( BIock II

I

I2.5-mlsec I

wind velocity4 I61.5-mlsec

%

wind velocity I I

%%%%- -% I I

-%%---I I-- I-%- I

wind shears I

I

/

speed profil e

65-percent wind -

speed profil e

Wind condit ion 3

Sa turn P (Block II) nd

Saturn IB Block I and Ul

Figure 3 . - Design wind c ri te ri a.

wind loads during boost. The wind bias is based on the stati stic al mean wind for thequa rter of the year when the vehicle is launched. Loads relief from the programed biasis verified with prelaunch wind measurements. Wind directionality w a s not consideredin vehicle design until design verificatio:: was CGiiipleieci in June 1 9 8 8 . In these analy-ses, the space vehicle loads were calculated using biased trajector ies. The inflightwinds were applied fr om the direction that would yield the most s ev er e spacecraft load-ing. The direction and magnitude of the winds were based on the wind data found inreference 6.

The primary wind-measuring system used at the time of the spacecraf t designwas a rawinsonde system. This system provided measurements of horizontal windspeedand direction a s a function of altitude averaged over approximately 600 me te rs . Be-cause the smoothing inherent in this sy stem cannot provide measuremen ts of sma ll-sc al e wind motion, the original spacecraft design did not include the effect s ofatmospheric turbulence.

A more accurate system for measuring inflight wind was requir ed to provide de-tailed wind profil es and information fo r st atis tical analyses to be used i n space vehicle

development. The FPS- 16 rad ar /ji msp here sys tem was developed as the wind-measuring device. The jims phere se ns or (balloon), when tracked by the FPS-16 ra da r ,provides a wind profile measurement fr om the surface to 17 kilomet ers altitude in ap-proximately 1 hour. The frequency resolut ion is fine enough to excite elastic-body re-sponses as much as 4 hert z in an analysis of the high dynamic pr es su re phase of boost.

A wind-monitoring team and system we re established by JSC and MSFC to cal-culate inflight loads o r resp onses on the s pa ce vehicle on which launch recommendations

would be made. Launch restr ict ions fo r the first two Apollo/Saturn miss ions we re

12



based on angle-of-attack dynamic-pressure a q limits that were proportional to thevehicle design loads. The a q fo r these missions w a s predicted using measured rawin-sonde data from the launch site. The predicted aq history w a s compared with the aqallowable, and a recommendation w a s made to launch based on thi s comparison. Thismethod of monitoring inflight winds did not include the effects of elastic-body response;

variations in axial load; nonlinear aerodynamics; propellant slosh; and variation in qcaused by winds, turbulence, thrust oscill ations, and gust penetra tion.

Beginning with the AS-201 mission, a new method w a s used to calculate spacecraftinte rface loads for comparison with struc tur al allowables (fig. 4). The launch recom-mendation could be made on the basi s of th is comparison. The FPS-16/jimsphere datafrom this measurement system were put into a vehicle flight-simulation prog ram fo rload calculation. In th is operational procedure, the wind measurements and flight- loadssimulations were begun 24 hour s before launch and repeated periodically until launch.This procedure gave a history of the winds and vehicle loads that provided the basi s ofoperational assurance during the actual launch. The flight-loads s imulations fo r both

spacecraft and launch vehicle on the Apollo 4 to 10 miss ions were conducted by a joint

JSC/MSFC monitoring team at MSFC. These simulations included the effects missingin the a q restr ict ion s, with the exception of propellant slosh. The elastic-body defini-tion w a s restricted to two modes.

Because the FPS- 16/jimsphere data measured before launch were t ransmit tedfrom the launch complex at KSC to JSC and MSFC, a JSC flight-loads simulation t h a tgave bet ter definition of spacecra ft loads w as initiated for loads calculation to make thelaunch recommendation. A JSC flight simulation fo r inflight wind monitoring w a s im-plemented on the Apollo 11mission and has been used on all subsequent missions. This

simulat ion includes the effects of four elastic-body bending modes and all the effectsmissing in the a q restrict ions except for fuel slosh.

Using the wind-biasing technique, inflight wind loads for the Apollo missions havebeen well below design values for all but the first mission, which was biased deliber-ately to in crea se the vehicle loads for a stru ctur al test . When the relieving pitch-planebias is cons idered, the design condition fo r the inflight wind condition is overly conserv-ative. The effects of wind biasing should be considered in the definit ion of the inflightwind condition for all future programs.

Several asse ssmen ts of cr ite ria for the inflight wind condition have been completedusing the wind data measured by the FPS- 16/jimsphere system. King and Ryan (ref. 7)made a comparison of maximum space vehicle loads resulting from the idealized syn-thetic wind profile to loads resulting fro m a large sample of jimsphere wind profile s.Although the maximum spacecraf t loads fr om the two wind descrip tions were of the same

magnitude, the most s ev ere spacecraft loading w a s in response to a high-turbulencewind profile , the maximum wind velocity of which was 26 m/sec compared with the82.65 m/sec of the design cri ter ia. This result indicates that, for highly flexible vehi-

cles , the influence of wind turbulence may be of equal or greater importance than thehigh-wind sh ea r and the high-wind magnitude and should be included in fut ure inflightwinds res ponse conditions.

TO combine rationally the effect of winds, propellant slosh, thru st oscillations,and tra jectory shaping in new vehicle design, a Monte Carlo analysi s of boost is recom-mended. A s discussed for the lift-off cas e, a Monte Carlo solution is one in which input

13



4Structural model

ngine and

actuator model

- - - -- -- ----- -

Final wind- I Vehicle dynamicsdata edit

iZ 6 % 0

- 0

-r -

s;g100

'pEX Ic L -

10 20 30 40 50 60 70 80 W 100

2 Flight time, sec

Figure 4.

-Prelaunch simulation for inflight wind loads.

Elastic-body

simulation

\nalysis

Launch complex

launch or no

launch because of

excessive winds

data for many c ase s ar e selected at random f rom pfobability distributions. For boost,probability distributions would be defined fo r all vehicle para mete rs, such as thrustoscillations, and the vehicle would be flown, by flight-loads simulation, through a largenumber of measured FPS- 16/jimsphere wind profile s. Using resul ts fro m these boostsimulations, a statistica l analysis could be made of vehicle st ru ct ur al loads; this analy-sis would be the bas is of design loads f o r the space vehicle.

14



C O N C L U D I N G R E M A R K S A N D R E C O M M E N DA T IO N S

Several conclusions can be made concerning the wind load design conditions of theApollo Pr og ra m. The prelaunch design conditions did not account fo r the following ef-

fects that could be cr it ic al to vehicle loading: exposure periods, turbulence, crawler

operation, vehicle propellant loadings , on-pad vehicle configurations, auxiliary damp-ing, and wind direction. The ground wind-tunnel t es ts we re conducted too late in theprog ram to support the original design of the spacecra ft. The Apollo design cri te ri aand analyses fo r lift-off were not reali stic because they did not consider accurate ly theeffects of unsymmetr ical thrus t buildup, on-pad purging, wind direction and exposureperiods, realistic description of vortex shedding, rea list ic rep resenta tion of thrustmisalinement, and three-dimensional st ru ct ur al dynamics. A Monte Car lo analysi sshowed that the root-sum-square method of combining loads fo r lift-off was overly con-serv ativ e and create d design loads for which the probability of occu rren ce is unknown.

The evolution fr om wind measurement to load measurement as a primary ground-wind monitoring sys tem improved the Apollo operational capability by incorporating

measu red vehicle dynamics ra th er than assuming the existence of conservative levels.The inclusion of a biased pitch program i n the space vehicle operations wa s significantlydifferent fr om the design condition and resu lted in an overly conservative inflight wind-response design condition. By analy sis, inflight winds of high turbulence and low mag-nitude have been shown to cause loads as severe as the loads resulting f ro m the designsynthetic wind profile; therefore, the original wind cri te ri a we re not adequate to ac-count fo r the effects of turbulence.

Emphas is should be placed on realis tic cr it er ia and analysis of the design windloads; therefor e, six recommendations are made.

1. New prog ra ms should include exposure periods, turbulence, propellant loadand pres suri zati on varia tions, on-pad maintenance configurations , crawler operations,and loads-alleviation methods in the design prelaunch condition. The se we re not con-side red i n the Apollo design.

2. Wind-tunnel ground-wind tes ting should be initiated e ar ly in the progr am toallow test re su lt s to influence vehicle design.

3 . The fu tur e design phase for lift-off should include development of a statisticalload analysis to provide rea list ic design conditions and gre at er operational capability.

4. Provisions for a wind-monitoring procedure that provides the maximum op-

erational capability consistent with a safe launch should be included as early as possiblein the design of new vehicles.

5. In futu re prog rams, the space vehicle loads relief for all conditions should

be included in the design as well as i n the operational portion of the vehicle life.

15



6 . A Monte Carlo analysis is recommended to define a more rational inflight winddesign condition. In such an analys is, the new vehicle could be flown, by flight-loadssimulation, through a large number of measured FPS- 16/jimsphere wind profiles, anda statistical analysis could be made using the r es ul ts of these simulations.

Lyndon B. Johnson Space CenterNational Aeronautics and Space Administration

Houston, Texas, March 2 7 , 1 9 7 39 1 4 - 5 0 - 2 0 - 1 4 - 7 2

REFERENCES

1 . Daniels, Glenn E . : Te rr es tr ia l Environment (Climatic) Cri ter ia Guidelines forUse i n Space Vehicle Development. NASA TM X - 5 3 0 2 3 , 1 9 6 4 .

2. Kaufman, J . W . ; and Keene, L. F. : NASA's 150-Meter Meteorological TowerLocated at the Kennedy Space Center , Florida. NASA TM X - 5 3 6 9 9 , 1 9 6 8 .

3 . Daniels, Glenn E . : Te rr es tr ia l Environment (Climatic) Cri ter ia Guidelines forUse in Space Vehicle Development, 1969 Revision. NASA TM X - 5 3 8 7 2 , 1 9 6 9 .

4 . Den Hartog, J . P. : Mechanical Vibrations. Fourth ed . , McGraw-Hill Book Co. ,I n c . , 1 9 5 6 .

5 . Schwartz, Robert D . : Monte Cir lc h a l y s i s sf Sti-iiztiii-iil Luads Eesuiting Fro mSpace Vehicle Lif t-o ff. NASA TM X - 5 8 0 4 6 , 1 9 7 0 .

6 . Daniels, Glenn E. ; Scoggins, James R. ; and Smith, Orve l E. : Terre str ia l En-vironment (Climatic) Cr it er ia Guidelines for U s e in Space Vehicle Development,1966 Revision. NASA TM X - 5 3 3 2 8 , 1 9 6 6 .

7. King, A. W . ; and Ryan, R. S . : The Influential Aspects of Atmospheric Disturb-ances on Space Vehicle Design Using Sta tist ical Approaches for Analysis.NASA TM X-53565 , 1967 .

16

Documents

Apollo Experience Report the Development of Design-Loads Criteria, Methods, And Operational Procedures for Pre Launch, Lift-Off, And Midboost Conditions