AD-A275 889

Crashworthiness Analysis ofCommuter Aircraft Seats

DTIELECTE

S.FEBl 1M

November 1993

DOT/FAA/CT- TN 91/28/

i i

Document is on file at the Technical Center Library,Atlantic City International Airport, N.J. 08415

v:-. 9 -05271

US Depamnt of Trnspwatlo

Technical CeWVWtoAtlanticCiyitrai aAiprNJ045

94 2 16 130

NOTICE

Thi3 document is disseminated under the sponsorshipof the U. S. Department of Transportation in the interestof information exchange. The United States Governmentassumes no liability for the contents or use thereof.

The United States Government does not endorse productsor manufacturers. Trade or manufacturers' names appearherein solely because they are considered essential to theobjective of this report.

Technical Rkpog Documentation Page"I. R- port No. 2. G iawroment A~ccesion No. . Catalan Me.

DOTiFAA/CT-TN91/28

4. Tmtls and Subtitle S. Report Omet

November 1993

CRASHWORTHINESS ANALYSIS OF COMMUTER 6. P9rforming Orgoniztion CodeAIRCRAFT SEATS

S. Poetomsingl Urglanization Report NO.7+ Au~or•e)

David H. Laananen CR-R-910129. Performing Orgonization Nome and Address 10. Work Unit No. (TRAIS)

Arizona State University ll. Contract or Grant No.Department of Mechanical and Aerospace Engineering DTFA03-90-P-00447Tempe, Arizona 85287-6101 13. Type of Report and Period Covered

12. Sponsoring Agency Nme. and AddressU.S. Department of Transportation Technical NoteF•ederal Aviation Administration January - SeptemberTechnical Center 14. Sponsoring Agency CodeAtlantic City International Airport, NJ 08405 ACD-210

15. Supplementary Noteo

Contracting Officer's Technical Representative: Tony Wilson

16. Abstract

During the past several years, the Federal Aviation Regulations (FAR) havebeen significantly modified with respect to seat/restraint system strength,attachment of seats to the aircraft structure, and the means by which they areto be evaluated. Aircraft accident data, human tolerance levels, and aircraftstructural characteristics have been considered in the development of thesenew standards. Dynamic testing is now required for seats to be installed ingeneral aviation aircraft, transport category aircraft, and rotorcraft.Performance criteria are similar to those specified by the Federal MotorVehicle Safety Standards for automobiles but also include a limit on "pelvicforze," in order to prevent spinal injuries which may be caused by thevertical component of impact force. A category of aircraft that has not asyet been affected by the rule modifications is the comnuter type aircraft,which ,eat• 10 to 19 passengers. Since this airplare is closer in size togeneral aviation aircraft than to large transports, it is also covered by FARPart 23. The Federal Aviation Administration is currently involved in theconduct of a test program addressing commuter aircraft occupant crash safety.In support of this effort, a research program the includes full-scale aircraftdrop tests, sled tests of seats, and computer simulations is being conducted.This report describes the use of the SOM-LA (Seat/Occupant. Model - LightAircraft) program in modeling three commuter aircraft seats. The predictedresponse of the seats to a potential set of test conditions is described.

17. Key Words 1to Distribliteon Stetement

Crashworthiness Document is available to the pýitlicAircraft Seats through the National TechnicalRestraint Systems Information Service, Springfield,C(omputer Simulatton Virginia 22161

S.cu •Clos.f. (oft,,uT r• $.5,yVS. .,, C .e'. (of this ptge) Mi o of Pages 22. Price

Unclassif ied JUnclassified 20

Form 1>0T F 1700.7 (8-72) Iopre 6ction of completsted pPoge otlorue2d

TABLF OF CONTENTS

EXEC.TIVE SUMMARY vii

INTRODUCTION 1

COMMUTER AIRCRAFT SEAT DYNAMIC TEST REQUIREMENTS I

FULL-SCALE AIRCRAFT CRASH TIESTING 3

ANALYSIS OF AERO COMMANDER SEAT RESPONSE TO VE-RTICAL DROP 3

ANALYSIS OF PROPOSED TEST CONDITIONS 7

SEAT RETENTION 11

CONCLUSIONS 11

REFERENCES 14

laooosion For

DTIC TABun~aoun•od n-

H-�; A a-d/Cl-SI ali

LIST OF ILLUSTRAIONS

1 Dynamic tests under consideration for commuter-category aircraft seats. 2

2 Floor warping requirements under consideration for commuter aircraft seat tests. 2

3 Aero Commander 680E aircraft during 26.8-ft/s drop. 4

4 Forward section of Aero Commander 680E aircraft following 26.8-ft/s drop. 4

5 Finite element model of Aero Commander stat structure. 5

6 Aero Commander seat structure deformation predicted for 26.8-ft/s drop. 5

7 Aero Conmmnder seat frame deformation following 26.8-ft/s drop. 6

8 Cracking in region of seat frame deformation, Aero Commander seat. 6

9 Beechcraft 1900 seat model. 8

10 Fairchild Metio HI seat model. 8

11 Sidewall/floor-mounted seat configuration. 13

12 Two transport aircraft seats and dummies prior to sled test. 13

LIST OF TABLES

Page

1 Analysis Results for Lap Belt Restraint. 7

2 Analysis Results for Three- Foint Restraint. 9

3 Analysis Results for General Aviation Passenger Seat Tests, Lap Belt Restraint. 10

4 Analysis Results for General Aviation Passengei Seat Tests, Three-Point Restraint. 10

V

EXECU.TIV SUMMARY

During the past several years, the Federal Aviation Regulations (FAR) have beensignificantly modified with respect to seat/restraint system strength, attachment of seats to theaircraft structure., and the means by which they are to be evaluated. Aircraft accident data, humantolerance levels, and aircraft structural characteristics have been considered in the development ofthese new standards. Dynamic testing is now required for seats to be installed in general aviationaircraft, transport category aircraft, and rotorcraft. Performance criteria are similar to thosespecified by the Federal Motor Vehicle Safety Standards for automobiles but also include a linit on"pelvic force," in order to prevent spinal injuries which may be caused by the vertical componentof impact force. A category of aircraft that has not as yet been affected by the rule modifications isthe commuter type aircraft, which seats 10 to 19 passengers. Since this airplane is closer in size togeneral aviation aircraft than to large transports, it is also covered by FAR Part 23. The FederalAviation Administration is currently involved in the conduct of a test program addressingcommuter aircraft occupant crash safety. In support of this effort, a research program that includesfull-scale aircraft drop tests, sled tests of seats, and computer simulations is being conducted. Thisreport describes the use of the SOM-LA (Seat/Occupant Model - Light Aircraft) program inmodeling three commuter aircraft seats. The predicted response of the seats to a potential set of testconditions is described.

Vi.

During the past several ,ears, the Federal Aviation Regulations (FAR) have beensignificantly modified with respect to seat/restraint system strength, attachment of seats to theaircraft structure, and the means by' which diey are to be evaluated. Aircraft accideit data, humantolerance levels, and aircraft strctural characteristics have been considered in the development ofthese new standards [ 1 ]. Based on the recommendations of a joint industry/government/academiccommittee, FAR Part 23, which deals with small airplanes, was amended to require dynamictesting of seats and restraint systems for "normal and utility" (general aviation) aircraft withcapacity for fewer than 10 passengers [2, 3]. Performance criteria are similar to those specified bythe Federal Motor Vehical Safety Standards for automobiles but also include a limit on "pelvicforce," in order to prevent spinal injuries which may be caused by the vertical component of impactforce. The amended regulations apply to all new general aviation aircraft manufactured since 1989.Also, FAR Part 25 (transport category aircraft) was amended to require dynamic testing of seatsand restraint systems, although to less severe acceleration levels in order to allow for the largerstructures of those aircraft [4]. FAR parts 27 and 29, which apply to rotorcraft, have also beenamended to include dynamic test criteria.

Another category of aircraft that has not as yet been affected by the ule modifications is thecommuter type aircraft, which seats 10 to 19 passengers. Since this airplane is closer in size togeneral aviation aircraft than to large transports, it is also covered by FAR Part 23. The FederalAviation Administration is currently involved in the conduct of a test program addressingcommuter aircraft occupant crash safety. In support of this effort, a research program that includesfull-scale aircraft drop tests, sled tests of seats, and computer simulations is being conducted. Thisreport describes a potential set of test conditions and acceptance criteria and the concurrent researchprogram for their evaluation.

I O TER AIRCRAFT SEAT DYNAMIC TESTREQ ME-YE

A possible starting point for developing dynamic test criteria for commuter airplanes is toestablish a set of two dynamic tests and related acceptance criteria similar to those that have alreadybeen adopted for generai aviation but with more severe deceleration levels. For the first test, theseat would be pitched upward 60 deg on the sled, so that the impact velocity of 31 ft/s has forwardand downward components with respect to the seat. The deceleration pulse would have a peakvalue of at least 32 g, which should occur not more than 0.03 s after impact. In the second test,the seat is positioned upright, but would be yawed 10 deg with respect to the impact vector. Theimpact velocity would be 42 ft/s, and the peak deceleration, 26 g, occurring not more than 0.05 safter impact. The two test conditions are illustrated in Fig. 1. In order to account for the effects ofthe floor deformation that may occur in an accident, the floor rail on one side of the seat would berotated 10 deg about a lateral (pitch) axis; the other rail would be rotated 10 deg about alongitudinal (roll) axis, as illustrated in Fig. 2.

Both tests use a 50th percentile anthropomorphic dummy; the dummy must includeprovision f6r measurement of pelvic force, the f-)rce that is transmitted to the dummy pelvisthrough the spinal column. By means of extensivt, -•xerimentaion using modified dummies andcomparisou of those test results with injury data I -iary fection seats, this compressiveforce has been related to the potential for injury' to th iiun T`- spin( due to an upward accelerationof the body 151.

Suggested pass/fail criteria include a requirement that, although deformtatiol of the seatstructuce is permitted, attachments of the seat and restraint system must both reri'airu intact.Specific injury-related limits are a Head Injury Criterion (1-1C) of toX(), a femur load of 2'250 it,and a pe',Vic ,.-ompressive load of 150W 1). Upiper torso restraint wNould .,. required •ully for thefront (r•lot) seats, where the load in a singie snomulder belt should nort ex,,'eed V750 I of the ,sguof O loiads iM duai ,atips, 2(XX) l'.

Forward and Downward Loadirng Forward and Lateral Loading

zY

10060 x

Requii'ed Velocity and Acceleration

"32g

31 ft/Av 42 /ftt/s

0.03s 4 0.056

Fig. 1 Dynamic tests under consideration for commuter-category aircraft seats.

°2:• j• -- -...

FIiy 2 F-lx~r ww'aping riirenments tinder ((onsidicratUor1 COHI-IiIu(I, iITVA'-rt S'oi teso.

~.L1L-S AL A1CAFCRSIEU

In or-)ei to investigate the applicability and practicality of proposed FAR, amendments forcommuter-type ? 'W-raft, the FAA Technical Center embarked on a program of testing and analysis.Because the vertical coiiiponent ofJO impact forces can be a sigpificant part of the occupant injuriousenvironmen~t in an airplane crash, testing of full-scale -ircraft began with vertical drops todetermnine the nature of vertical iccelciations at the floor. The first two tests used airplanes at thesmaller end of the commnuter category, an Aero Commander 680E and a Ces'-na 421. Fullyinstrumented dummies were placed in all seats. Accelerometers were installed on the floor at majorframe locations. Each aircraft was dropped in a flat configuration onto a rigid platform fromn aheight of 11.2 ft, so as to achieve an impact velocity of 26. 8 ft/s. equal to the verticai component ofthe combined longitudinal/vertical test.

In the Aero Commander high-wing aircraft, the wing assembly crushed down into thecabin up to a maximum penetration of more than 20 in. at a time of 0. 18 s after initial impact, asshown in the photc,-xmh of Fig. 3. After elastic recovery of the structure, the cabin interior heightunder the wing was tound to have been reduced by more than 12 in.. The subfloo~r structuxe in thecenter of the aircraft crushed less than 0.5 in. so that the floor between the inboard seat tracksremained nearly flat, as shown in Fig. 4. Outboard sectioi.s of the floor were pushed downwardby the fuselage sidewail. The outboard seat au-ck on the right side of the aircraft, movipg with thefloor, was pushed down approximately 1.7 in. relative to the center floor section ar-d rotated 16deg about its own axis. On the left side of the aircraft, the outboard seat track was pusheddownward approximately 1 .5 in. and rotated about 6 deg. The seats (in the absence of longitudinalloading) remained in place on the tracks, although attachment fittings were bent atid the seat backstructuie on two of them failed under the aftward component of force from the dummy. Theacceleration mea-sured on the aircraft floor varied from one location to another but, when tikered inaccordapce with SAE Recommen(:ed Practice J21 1 [6], exhibited ptak values between 20 and 50g, in the range of the proposed 32-g seat test requiremenit.

The low-wing Cessna aircraft did not experience any significant deformation of the cabinstructure in the flat drop. In fact, the stiff wing structure limited crushing of thc subflwor structureto less than 1.0 r. out caused accelerationF at the floor that. exceeded 70 g.

ANISA , A L ( YMMF UL1NE ARQ1CLIQP

The response of existing commuter aircraft stat designs to a range of cras;" conditions hasbeen exam-ined using the SOM-LA computer program, which has been developed under FAAsponsorship [7, 8]. This programn combines an I I-mass, 2-9 degree-ofl-fr4erdori model of theaircraft occupant with a finite element mod~el of the seat structure. A.ý a chtek on vatlidity of the seatmnodel, the conditions cf the Aero Commander vertical drop test were first simulated, using theacceleration measured at the floor, and the pre~dicted- seat responISe Was 'com_1paTCd With teSi; ýsutThe SOM-LA finite element model of the Aero Commander seat stmcw~re, which cc!cststs mainlyof welded steel tubing, is shown in Fig. 5. Nodles I through 4 are attached ;o ihe Ilor he lapbelt is attached to the seat at nodes 17 and 18. Diring simulation of the 26.8 ft/s drop, the yitld,strength of tht- steel frame was reached at approximately 0.030) s An the 0.7 it. diameter tubuiarrnen~otrs tl-at run along ithe left and right sides of the scat (at nodes 19 and 20). A\ ni unrun- oeof'.1950 lb t'xerted Uy the dumm111y downward on the seat was predicted at a tinme of W.U5 ' At thattimle, tile side tubes weire bowved dow.nward approxirn-ately 0.52 in. io riodes 1 i ad 20, as s-hownin the side view pjresented as Fig. 6. All of tht single. passtenger seats Intlr in th(' aircraftduring the test e).perfienced deformation in the saine tegion ol the irame s r( -C~d vplcal

deformlation can bt seen in iFig. 7. Two of tle seat frames bent enougth to crack In 0ýe vu icliy ofnw~e,, 19 and 20 on the mnode,,as showiin i ~Ig. 8.

~gomw

Fig. 3 Aero Conirnander 680E awiraaft during 26.8-ft/s-drhop.

4 orwm (aIl sect I(oi ofiVro Connnswýdei- 68XO1. aii cr-af followint, 260.8-- t/- j

AERO COMMANDER 680E SEATTIME = 0.0000 SEC

"25

2 ,4

22

1 4

Fig. 5 Finite Jen-rnt model of Aero Commander seat structure.AERO COMMANDER 680E SEAT AERO COMMANDER 680E SEAT

TIME - 0.0000 SEC TIME - 0.0350 SEC

2

Fig. i 0Ac (a om mandei w.ai Si tturT deformation prdihct•d toir ,b. 8 li'i. cl U'p,

5,

Fig. 7 Aero Commnander seat fran~e. deformationl follow-mg 26.8-fit/ drop.

1-'Dllowing successful simulation of the Aero Comrmander seats that had been installed in thetest airplane, the SOM-LA pi agraxn w~is used to analyz~e the response of three, existing commuteraimcTaft seat designs to the proposed dynamic test conditions. The first was the Aero Commanderse~a descibe in the preceding paragraphs, the others were the passeniger seats used in two of thevrnost widelv used comimuter aircraft, the Bcechcnaft 1900 and the Fairchild Metro III. *The finiteekn-xrevi wodals of &h: Beech and Metro seats are illustrated in Figs. 9 and 10. The Beech seat isattached to the aircraft at nodes 1, 2, 22, and 27; the Metro seat, at nodes 1, 2, 3, and 4. Resultspredicted for the dummy in both dynamic tests are summarized if Table 1. No floor deformationswere applied in the simulations, for Attemnpts to apply the proposed floor wwrping requirements inthe program caused failure of all three seat models in the vicinity of the attachment points.Furtheimore, the SOM-LA programn has the capability to bypa~ss the finite element model andsimulate a rigid seat, which supports the cushions in fixed positions in the aircraft. In order todemonstrate the rigidity of the seats and the need for energy ab,,sorption in their strutures, thisoption was exercised using the Metro configuration, and its results are also included in Table I forcomparison. As noted in Table 1, the. Aero Commander seat structure failed during simulation oftest condition 1, prior to application of the full test pulse. Therefore, program execution wastcrininated before the dummy response reachied peak values of accelerations and forces.

Reftrring to the simulation results for test condition 1, the maximum pelvic for.ce presentedin Table 1, for every seat, exceeds, the proposed acceptance limit for compressive force (1500 lb).In fact, except for the Aero Commander case, in which the sear s~.rcture failed prematurely, thecompressive load predicted is more than twice the limit. The fact that the pelvic compressive loadfor the Beech and Mctro scms, is close to that predicted for the rigid seat indicates that neither seatprovides any inhe.'-cut energy absorption capability in its structure. Therefore, some kind of'.'ýrtical force - ttenu ati ng mechanism should be included in 9rder that the seats be capable ofmeeting the pelvic force criterion.

Table 1. Analysis Results for Lap Belt Restraint

LAccele-ration WNATIf Neck Moment1Seat Pelvis (g) Chest (g') Head (g) 1-1C (lb) (Ibib.)

Aesm (ýnýdi2 217.4 24.6 26.6 5. -2450. -G.0/44537

Bet'ch 1900 46. 5/ 161.3 31319 502k 4840. -3480. I-53.1/±84.1

Mtx 11 48.1 56,6 57A4 269. -3780- 31.7/+ 111.

Rigid 4T.5 57.9 58.8 277T -3820. -19414

Aem Orrir .28.0 41.3 74,9 9,5. +-2720. -46458e~hPY 97193-(,/184,3 64 9/388.' 10400- +2)940- 8./ 52

S. ISrg for- h~. adIng ~I Vun t-; nr i f ''IA1 .- C a.1ý\ bl~. n~~ . ut c1il.d IIIt OAb ý 0 I h lwq, pr -g I, I ge i

'A,1 1tIU~'~I~'wm Y mxgittkit amr~b ca~ rq of oll ~ iiI i

/ 17

I Is41426

27

ps 21

Fig. 9 Beechrmaft 1900 seat model.

N2

2 j/"V' ,

The tabulated reusahs priedic-ted. for test condition. 2. are. inconclusiveý with rcsxect to thkprcoposed pass/fail critetia. The maximumr pe~lvic force in those cases is pAoSiti'm, imnplying tension,to which the 'ý.juwy criterion dtes m~t apply. The 111C2 values vire acceptable, but no requiremnenthas been iVIcIlude ftr siimulation of the passe~nger enviro nmniit for head strikes, such as on the seathack in front of th,ý passengers. Bending mnomentt in the neck, also presented in TFable I ascormpute~d by SOM--LA, is not listed amiong tht. proposed criteria. However, the momentspredicted in sirnwation of test condition 2, ivi even, case, exceed the neck toleranc-1 limits proposedby Mertz and Patrick [9]. Simply stated, these limits are 65 lb-in. for flexion and 35 lb-in, forexten'sion.

'The baseline conditions being investigated would require upper torso restraint for the pilotseats. Therefore, the two test conditions were. simnulated a second time for the three seat designs,this time including a three-point automotive-type restraint system. Results are presented in Table

2.Dummy segment accelk'rations pr-exiciecl. for test coutdition 2 were reduced below those of TableI due to the prevention of head/chest impact on the legs, and all the predicted HIC values wereacceptable. Therefore, these numbers were not included in Table 2, but were replaced by themaximum shoulder belt load, which exceeds the 1750-lb acceptance limit for test condition 2applied to every seat. Also just as prtedicted in the lap beilt-only cases of TFable 1, the maximumpelvic force Iexceedis th e 1500-lb limnit. As noted in Table 2, both the Ae~ro Commander and Metrseats failed early, before the dummy reached peak response. Strengthening the seats to preventthese, structural failures would undoubtedly permit the tabulatedi fi-wces to tecome higher.

Because of the high pelvic loads predict Ad for t.est corndition A with its 32-g peakde~celeration, another set of s;niularions was perfor-ned icv which, the seat models were subjected tothe l1-Fs severe dynamic te!st conditions that are required by FAR Part 23 fo r general aviationpassengeiý seats. Although imnpazt velocities and impact 'vectorcorientations re~main the same, thepeak decelerations air reduced to 15 g and 21 g for tests I and 2, respectivety. Results for lap belt-only restrmin are. presented in Table 3, anid results for three-point res&rait (as reqluired fo~r 3all

gnalaviation seats), in Table 4.

Tfable 2. Analysis Resuill for Three- Point Resmranlt

Aero (2nxtr- -.2 1,70, -0.23i f A4,9 ` '1W

Beexh ) 412~ 41~M .~ 8

Mean-ý p,13 - 1260. L~/I-I

It' 4id -3830. 16.U + 12,1. 1'29.

Aroi (ndrl 56 10/. 12 2. 0 190.

Ileech 'i 1990. .38 2)( 2800.

i/Vu~ i0. 44 i.'~.> 160.d,

h-~~ ~ ~ rlh'i .0ý

.Iiabie Aralysis Results fbr Generala, Aviatiofl PaI,:senger Seat Tests, Lap Belt Restraint

"~cc el'i•-in' ?eafnTTke- Neck MomentISeat Pelvis (g) Chest (g) Head (g) HIC (lb) (lb-in.)

Aemo Cradr2 22.dJ,, 20'•ll, 17.7 23,1 -1530. -0.17/+49-5

Beech 1.900 24.1 23.8 22.2 35.0 -1530. -25.2 -I1.0

Metro Iri 25.3 27.2 25,3 61.0 -1830. -12 9/+62.5

Rigid 25.2 27.4 25.6 61.3 -.1840. -12.7/+62.9Toi 2

Aeiru Cmdr3 14.6 3.9 6.0 1.9 +,,-27.4 -27.6/+0.,54

Beech 1900 30.5/284.4 73.3/192.4 62.8/294.4 5,790. +26313. -82.91+129oMetro 1I5 24.7 4937 67.7 477. + 3060. -52.4,/+0.92

Ri id 50.0/222.4 50o8/176.4 64.0,377.4 10 3 67.5

Notes: 1. Sign convention for bending moments: 4 = flexion; -. = extension.2. Aero Commander seat stmictwe failed at 0.093 s, halting piogram execution,3. Aero Cornmandtr seat structure failed at 0 065 s, haltitig program execution.4. Maydirnun acceleration magnitudes are before/atter impact of chest & head on legs.5. Metro seat structue failed at 0 122 s, halting program execution.

"Fable 4. Analysis Results for (ineral Aviation Passenger Seat Tests, Tlree-Point Restraint

"Neck Moment1 I -- t-Lo -"Seat (|b) (lb-in.) (lb)

AIc, . 420. -0.17/-55.4 453.

bee4'(-1) -X63. --0.26[+51.6 715.

MdU.) tl - 1J8-9.30/0-68.3 240.

RigikI' 1880. -8,881 0 ,68,8 233.

Aý,ro (n itr. + 1 !9•, -6,03•/+0t,2•6 4S5.f~txh 9(•)-1730, 1.- I9/ 8 94A 4 6, 0,

IM e 'tro IH O '1 - 16 8 0 . -1 .2t / 8 . • (K I i (I - I 7i 6 5(1t,- 144,. :, )

... . .. ... . ......

L -..-A:l•t •If!i-IJ~ <. •t )- -, -lt-. .- . . .i•:i,. ftl.-k:<' ".... .. li . .. ,.----::

A ; ,.tlc , :!••. st•<r:i• t d ; ) <7 ;

.A major concern -in designing for c~caparit survivability is die inertial loading of the seat onthe fittings by which it is attacheA. to the Aiv,,ra;; Because of the ý.arge downward component offorce it produces on the floor, test condition .1 tends to keep the seat in place. Test condition 2,howcver, with its significant forward loading component, exerts an upward pull on the rear legs ofthe seat and represents the c~itical conditiont for seat strengti. The WL-deg yaw in the requirementsserves to create an unsymnietric loading that iil.;reases the severity of Iloading on. one of the rearattachments. The SC)M-LA program determiner, the loading at the seat attachnri,;t points betweenthe sea! stiacture and tie aircraft, results that can be useful in design of the attachment hardware.Furthkermnore, success of the design would L,- ultimately demonstrated by testing.

In the preceding section, it was mentioned that none of the three scats, being analy-.ed couldsurvive the application of the floor warpin~g coaditictns. Of particular concern with respect to seatrettntioa are those aircraft in which one side of each seat is attached to the side of the fuselage,while the other side is supported on the floor. This configuration, which is schemnaticallyillustrated 3n Fig. 1!, appears in somne of the most frequently used cornmuter aircraft, including theBeechcraft 1900) nd the Fairchild Metro. Fuselage deformation during a crash can causesignificant movement of 'outboard seat attachment points relative to the inboard legs, which mayexceed the, floor warpage conditions specified by the proposed amendinent. The invezhgation of aNovemnber 1987 accident involving a Beech 1900 showed that, although theý fuselage remainedintact, "all of the seats separated from their floor- and wall-mounted seat tracks." The crash wasfatal to both pilots and to 16 of the 19 passengers, and "the majority of the injuries sustzined by thepassengers were as a result of the secondary impact after the seats separated from their tracks."[110] Coi--irmidng the need for- energy absorption in th:-, vertical direction, the NTSB3 estimated thatthe average acceleration along the vertical axis of the aircraft during initial impact ranged fromi"19.8 to 35.7 g" and that "the vertical velocity chanige was itbout 42 feet per second,," The reportstates farther that '"somne injuries, such as aortic ruptures, were typical of a severe verticaldeceleration. "

Returning to die matLerT of seat retention, the NTSB reported ,Nit ;n a 1980 crash of aSwearingen Metro aircraft, "Despite the integrity of the cabin -re~a, all oA the 13 occupied passengerseats separated from their attachments during tile impaz:t sequen~ce, leading to the (11) fatal and (2)nonfatal injiuries to the passengers." I 11. AlthGugh passtriger seat configuratien rernains basicallytile. same in the cunernt Metrow III aircraft, attach,,ment hardware has been improved.

'Q LQLU-RQN

B~ased on analyses oil the pr-rpose.4 test coniditions, several conclusionis cart be dra.wn.First, the close spacing of passenger seat iows in cournnuter aircraft makes head impact. against tileseat back likely in an ac~cident with a sK.gnificant longitudinal acceleration, as represtnted by testncondition 2, A..hdough the proposed arnencelinet specifies a IIIC limit, it could ailso U.cciude itfoethod for evaluiating the actual passýenger env irownnent, &uch as by the Use. of' two seat rows, ;' ai4dynaiiic test. F or sac cases, tile possihnjrty 4,% ineck in~arv shoildd,,i ý, I oiisidcro RelelreAC~t12 dee s~study~ of the effkct oV,,,caz &,si, n paramrne!.., inctludislg s~row ' ptuing and seatback stiffness,~ the poteuwia for pa, sengci nr ntas~4 rr ;oi A~v ~ pre hru~svd data froyn LI'C tCu SI that Were condU!)cted tsn wo Sc"At vowsv as howl In V.12 1131.i~talrtw vclwciat'es wt ario (C:~ 44 [i's, and to2l ta~ pVc 16g 1 il&ad linpact:i1-Cchi.:ted bcomnputer kuni atioll j 'iur~ne HI I ll( iue '"4s!rlii anlyi ý rhovs 1W01 and wec'k'lonlients Iin :tASCLa~'if coi sddelabl V Nooe Ow I'mfh ~~C~iu(itd~'iAT. 11i

1'0i. curluIIIIC! :"i l e'st (,0ldi te.n 2a Ic" 2ý" dc. i he > lcf mt, ý 4TC:11inllt i eK, 11C U'W tofS tt~~sm oeý6i v : 1:dn.h alk Illm the airt':Yart,

TIte hlig'h pevc IV ' Clads icdc 1' 0i 1~ c S ) I _I. A i iii -1d t~ o ito V,(l c,imlc tcn lfwrv "';.hsorpitso~ lil bt mc Vli i * hy' il If-JA 1i ý'Icc 1 ie i,` c i c I

11,h c t (.c d v k ý~l S, wl tiý ,

Seat retention has been a problem in accidents involving commuter aircraft. The floordeformations produced by the Aero Commander drop test inJic-ate that the 10-deg-pitch/10-deg rollfloor warp conditions are no more severe than deformations produced in actual floor stnrctures;some aircraft may force even greater displacements on their seats. It appears from the SOM-LAsimulations that none of the three seats modeled would survive these w,rping displacements, sothat introducing new seats or modifying cuirent seat designs to accommodate these displacementswould certainly represent an improvement. The FAA Technical Center is planning a drop test of aMetro III aircraft in 1991L It would appear desirable to install on that afirraft some seats that havebeen designed, or at least modified, to meet the floor warp conditions.

The acceleration environment inside the aircraft can vary considerably from one aircraftmodel to another, as demonstrated by the drop tests of the Aero Commander and Cessna aircraft.A seat that might stay in place under the proposed floor warp conditions could break loose due tohigh inertial loads in an aircraft that has a stiff irnderfloor structure, such as the Cessna 421. TheU.S. Army approach that has been used in design of two helicopters, the UH-60 Black Hawk andthe AH-64 Apache, is to specify, in addition to design and testing requirements for the seats,crashworthiness requirements for the complete aircraft, including the landing gear and the fuselagestructure. Compliance with these requirements may be demonstrated by analysis.

Pig. I1I SidewaflVoor-mounted seat configuratiori.

l.

lo

AuIi: S 111 1 O ýd ,L

I. Softis, S.J. and INissley, W.J., 'The Development of Dynamic PerformanceStandards for Civil Aircraft Seats," 1990.

2, Soltis, S.1. and Olcoti, 1.W., 'Trhe Development of Dynamic Perfontnance Standards,for General Aviation Aircraft Seats," Crash Dynamics of General Aviaxion Aircraft, SP-622,Society of Automotive Engineers, 1985.

3. Federal Aviation Regulations, Part 23, Airworthiness Standards: Normaul, Utility,Aerobatic, and Comnmuter Category Airplanes, Amendment 23-36, "Smiall AirplaneAirworthiness," UTS. Department of Transportation, adopted 1.988.

4. Federal Aviation Regulations, Part 25, Airworthiness Standards: Transport CatepqryAirplanes, Amendment 25-64, "Improved Seat Safety Standards," U.S. Department ofTransportation, adopted 1988.

5. Chandler, R.F., "Data for the Development of Criteria for General Aviation Seat zandRestraint System Performance," Crash Dynamics cf General Aviation Aircraft, SP-622, Soc-lety ofAutomotive. Engineers, 1985, pp. 13-37.

6. Instrumentation for Impact Tests, SAE Recommended Practice J21 Ilb, Society ofAutomotive Engineers, 1974.

7. Laananen, D.H., Bolukbasi, A.0., and Coltmnan, J.W., Computer Simulation of anAircraft Seat and Occupant in- a Crash Environment, Volume)I - Technical Report, DOT/FA-A/CT-82/33-1, Federal Aviation Administration Technical Center, Atlantic City Airport, NJ, March 1983.

8 ý,Laananen, [).H., Computer Simulation of an Aircraft Seat and Occupants in a CrashEnvironment - Program SOM-LA/SOM-TA User M~anual, DOT/FAA/CT- 90/4, Federal AviationAdministration Technical Center, Atlantic City Airport, NJ, May 1990.

9. Mertz, H.J. and Patrick, L.M., "Strength and Response of the Human Neck,"Proceedings of the Fifteenth Stapp Car Crash Cotference, Society o1 Aa~tomotiv~l gineers,1971, pp. 207-255.

10. Aircrafx Accident Report: Ryan Air Service, Inc., Flight 103, Beech AircraftC,-rporation 1900C, N4OJRA. Homer, Alaska, November 23, 1987, NTSL3/AAR-88/l 1, NationalTransportation Safety Board, Washington, DC, 1988.

11. Aircraft Accident Report: Air Wiscorksin, Inc., Swearingen SA226 Metro, N650S,Valley, Nebraska, June 12, 1980, NTSPIAAR-80115, Nationa; Transportation Safety Board,Washington, DC, 1981.

12. Laarranen, D.H., "Analysis of the Effects of Seat Parameters on Passenger In~jiryPnr~mial1 in Tmnspont Air-raft Accidents," Cornp-aters in Engineering, Vol, 3, American Socilety ofMechanical Engineersi, 1988, pp. 559-561.Memorandum Report A.AC`- 1 1 9-81 -8A., Protect ion and Survival L.ahoxratorY, Civil AeromedicalIn1SiitUte, Federal Aviation Administration, Oklahoma CiAty, 1983.

14. Linder-hill, B3. and McCtillough, V-. "An I" terg y -Absorb ing Scai Dt sign for LightAircrafi,'' 13ýper No. 7203-22, Socxiety of'Autoriotive I'i-gineers, 1972.

15~, Garrison, J R. and WaldrupII. "'I'he Bell Model 222," Paper No. 770951,Secivety of Autonxotive I 'i'r 1917

16. Stricelv, aC w I )--s Wd l'ia'n' S" f'P., "Crashworthy H elicopter Seats and OccupwnnRt;raii ystem," Operiwnal Ut he iortetr viation Melictine, ACARI) COnference, 1roceedingS

No, 2 55, Nortlh At'lantRIic .`Yrat Ogi . ,nzon, AdvisoryV (.4rOup1 tot. A cfl)s xcc Rela(cýrch an IdYe 'ropflrONcuni ks'in ci~:iac, 1F'ru iJJce, 11"8

17, .ow' W-, Ok-spn anrd Test ' ~ ou u /cT&-ae 1" flClv F'bXst!)inU'i Seý'a~ U' ~j A\Q11 1-1ifO~ ie-8 ~pp KA 1r-hwmeioiV ii~ 1.'1- . Arn,

ketI c and 'I (k!