NLR-TP-2001-108

Development of a crashworthy compositeDevelopment of a crashworthy compositeDevelopment of a crashworthy compositeDevelopment of a crashworthy compositefuselage concept for a commuter aircraftfuselage concept for a commuter aircraftfuselage concept for a commuter aircraftfuselage concept for a commuter aircraft

J.F.M. Wiggenraad, D. Santoro*, F. Lepage**,

C. Kindervater*** and H. Climent Mañez****

NationaalNationaalNationaalNationaal Lucht- en Ruimtevaartlaboratorium Lucht- en Ruimtevaartlaboratorium Lucht- en Ruimtevaartlaboratorium Lucht- en RuimtevaartlaboratoriumNational Aerospace Laboratory NLR

NLR-TP-2001-108

Development of a crashworthy compositeDevelopment of a crashworthy compositeDevelopment of a crashworthy compositeDevelopment of a crashworthy compositefuselage concept for a commuter aircraftfuselage concept for a commuter aircraftfuselage concept for a commuter aircraftfuselage concept for a commuter aircraft

J.F.M. Wiggenraad, D. Santoro*, F. Lepage**,

C. Kindervater*** and H. Climent Mañez****

* Alenia Aerospazio- Aeronautic Division** CEAT, Department Atterrissage Crash*** DLR, Institute of Structures and Design**** EADS - CASA, Development and System Division

This report is based on a presentation to be held at the 57th Annual Forum of the AmericanHelicopter Society, Washington DC, USA, May 9-11, 2001.

The contents of this report may be cited on condition that full credit is given to NLR andthe author(s).

Division: Structures and MaterialsIssued: 9 March 2001Classification of title: Unclassified

-2-NLR-TP-2001-108

Contents

ABSTRACT 3

INTRODUCTION 3

APPROACH 4

DESIGN RATIONALE 4

Materials 4

Loading 5

Structural concepts 5

COMMUTER FUSELAGE DESIGN 5

FABRICATION 6

DROP TEST 6

Test configuration 6

Instrumentation 6

Failure mode 7

Lessons learned 7

NUMERICAL ANALYSIS 7

CONCLUSIONS 8

ACKNOWLEDGEMENT 8

REFERENCES 8

14 Figures

(13 pages in total)

-3-NLR-TP-2001-108

Development of a Crashworthy Composite Fuselage Concept for a Commuter Aircraft

J.F.M. Wiggenraadwigg@nlr.nl

National Aerospace Laboratory NLRPO BOX 153, 8300 AD Emmeloord, The Netherlands

D. Santorodsantoro@aeronautica.alenia.it

Alenia Aerospazio- Aeronautic DivisionViale dell'Aeronautica, Pomigliano d'Arco (Na), Italy PC 80038 C.CIT

F. Lepagefabrice.lepage@ceat.fr

CEAT, department atterrissage crash31056 Toulouse CEDEX, France

C. Kindervaterchristof.kindervater@dlr.de

DLR, Institute of Structures and Design,Pfaffenwaldring 38-40, D-70569

Stuttgart, Germany

H. Climent Mañezc80411@casa.es

EADS - CASA, Development and System DivisionAvda. John Lennon s/n, 28906 GETAFE, Madrid, Spain

ABSTRACT

Within the framework of Brite-Euram programme CRASURV "Commercial Aircraft - Design for CrashSurvivability", technology was developed for the design of composite air frames with respect to crashworthiness.The ultimate goal of the project was to develop computer codes for the simulation of the crash behaviour ofcomposite fuselage structures. A significant part of the project consisted of the design, fabrication and drop-testing of two representative composite fuselage sections, to generate the experimental data needed for thevalidation of the new code developments. The present paper gives an overview of the development, test andnumerical analysis of one of the fuselage sections, a one-bay section representative of a commuter aircraft likethe ATR-42/72. The fuselage section consists of the sub-floor structure, which is the major area that will becrushed during a potentially survivable crash. The structure failed in a mode which was not predicted. Thedeficiencies of the model were repaired and a post-test analysis gave satisfactory results. The project hasresulted in improved simulation capabilities. However, it cannot be concluded that the state-of-the-art is such,that the behaviour of new composite structures can be predicted accurately in the near future.

INTRODUCTION

Traditionally, fuselages of fixed-wing transport aircraftare made mostly of aluminium, a material withconsiderable capacity for plastic deformation, hence,an inherent capability to absorb energy in crashsituations. Since the last two decades, compositematerials are used more extensively to build aircraftstructures, although in civil transport aircraft,

applications are limited mainly to the empennage, theouter wing and secondary structures. However,application of composite materials to aircraft fuselages

becomes feasible rapidly, and the crashworthinessaspect related to composite structures has become aserious issue. The composites used for aerospacestructures are generally brittle materials with hardlyany capacity for plastic deformation, socrashworthiness is no longer a materials aspect, butrequires a combined materials/ structures/fabricationapproach.

Presented at the American Helicopter Society 57th Annual Forum, Washington, DC, May 9-11, 2001.

-4-NLR-TP-2001-108

The design for crash survivability has become ofincreasing importance during the last decades, not inthe least because so many crashes have beendemonstrated to be potentially survivable. Crashes ontake-off and landing around the airfields were shownto be the most common survivable crash scenarios [1].The load cases under crash conditions are wellestablished, hence, the design of metallic aircraftstructure with "crashworthiness" capabilities, i.e.,capabilities to protect the passengers up to a certainlimit, is quite manageable. A Brite-Euram projectentitled "Design for Crash Survivability" was dedicatedto this subject in 1992 [2]. A subsequent Brite-Euramprogramme was started in September 1996, andcompleted in February 2000, entitled "CommercialAircraft - Design for Crash Survivability" with acronymCRASURV [3]. The objective of this project was todevelop the technology for the design of composite airframes with maximum safety with respect to potentiallysurvivable crash scenarios. The main activities withinthis project were the development of computercapabilities to simulate the crash behaviour ofcomposite structures, and the design, fabrication andtest of representative test articles for the validation ofthe computer codes. These articles consisted of anumber of generic sub-floor "box" structures, typicalfor helicopters and small commuter aircraft [4, 5], andof larger sub-floor structures, typical for transportaircraft such as the Airbus A-320 and the ATR-42/72.The development of the latter structure and itsnumerical analysis is described in the present paper.

APPROACH

The project team was assembled from representativesof the European aircraft industry, aerospace researchestablishments, universities, computer codedevelopers and several small businesses. The specifictask objectives were to develop appropriate materialmodels, to design and build representative compositeaircraft fuselage components, to test thesecomponents, to develop and apply computer codes tosimulate crash behaviour, to determine the effect ofthe crash loading on the occupants, and to assess thenewly developed methodology, and propose designguidelines.

The ultimate goal of the project was to developcomputer codes with which crash behaviour ofcomposite fuselage structures can be simulated. In theprevious Brite-Euram project, the crash behaviour ofmetallic fuselages was simulated and validated bycomparing numerical results with experimental resultsobtained from a drop test of a section of an AirbusA320 fuselage. For Brite-Euram project CRASURV,drop tests on sections of current air frames were notfeasible, because aircraft with composite fuselages donot yet exist (other than small aircraft and helicopters).

Hence, a significant part of the project consisted of thedesign, fabrication and drop-testing of tworepresentative composite fuselage sections, in orderto generate the experimental data needed for thevalidation of the new code developments.

One fuselage section, representative of a commuteraircraft like the ATR-42/72, was defined by the Italianaerospace company ALENIA. The second fuselagesection, representative of a larger airliner like theAirbus A320, was defined by EADS France. Theprincipal difference between the two aircraft (withrespect to crashworthiness) is, that the ATR-42/72 hasa very small space between the cabin floor and theouter shell, while the A320 contains a cargo holdbetween the cabin floor and the outer shell, hencecontains more space to allow deformations in a crash.The two fuselage sections developed for CRASURVwere limited to the sub-floor structures, including twoframes. The sub-floor structure is the part of thefuselage below the cabin floor, which is the major areathat will be crushed during a potentially survivablecrash. The present paper gives an overview of thedesign, fabrication, test and numerical analysis of thecommuter fuselage section, carried out in co-operationbetween ALENIA, the Dutch and German aerospaceresearch centers NLR and DLR, the French testcenter "Centre d'Essais Aeronautique de Toulouse"CEAT, and the Spanish aerospace company EADS-CASA.

DESIGN RATIONALE

Crashworthy structures are structures which aredesigned to sufficiently withstand survivable impacts,so the occupants will survive. For a crashworthydesign, two important aspects have to be covered: forthe passenger space a "livable" volume needs to bemaintained, with sufficient potential for evacuation,and the impact accelerations submitted to thepassengers need to be limited to well specified values,for impacts at various angles. So far, crashworthyaircraft structures made of composite materials havebeen developed and implemented for helicopterstructures and small aircraft. In this field, designrequirements were formulated, and design solutionswere developed and demonstrated in tests. On aresearch scale, NASA has investigated the behaviourof fuselage frames for fixed-wing aircraft in impacttests. The following lessons were learned:

MaterialsMetals absorb energy by plastic deformation. Formetal structures, this applies to all failure modes:tension, compression, bending, etc. Compositesabsorb energy mainly by fiber breakage. It is importantto break a fiber many times over, rather than onceonly. For composite structures, this seems possible

-5-NLR-TP-2001-108

only in a compression mode without buckling, i.e., in astabilized configuration. Failure in tension or bendingusually doesn't lead to multiple fiber fractures.Helicopter technology has indicated that carbon-fibersare best for energy absorption, and aramid fibers arebest to provide stability and integrity of a crushingstructure. Solutions are therefore sought in mixedlaminates of both fiber types, either as a laminate withdifferent material plies or as hybrid fabrics.

LoadingCompression loading is prevalent in a crash, and afailure mode in compression is a suitable mode forenergy absorption in a composite component.However, stability needs to be provided during thecrush loading. Also, the initial failure load is usuallymuch higher than the subsequent sustained crushingforce. This high load leads to correspondingundesirable high acceleration levels. Therefore,"trigger" mechanisms in the form of local weak spotshave to be integrated in the components which arededicated to absorb energy. However, such weakspots may undermine the structural stability of thecomponent. Trigger mechanisms also play a dominantrole in the type of failure mode following the failure ofthe trigger, hence are very critical to the actual energyabsorbed.

Structural conceptsThe cylinder is the optimum configuration with inherentstability when loaded axially in compression, but notoften a practical configuration. Sine-wave orcorrugated beams are the next best component withinherent stability. However, after failure of the trigger,this inherent stability may be significantly reduced, andadditional stability needs to be supplied by adjacentcomponents in the lateral direction. Tension-loadedcomponents made of composite materials, such asaircraft skins impacted on soft soil or water, are brittleand break directly. To provide resilience for suchloading cases, the tensor-skin concept was invented,which provides a large pseudo-plastic deformationcapacity to the skin [6]. Points of concentrated load-introduction in composite structures can be mademore resilient, and even energy-absorbing, byintroducing a large amount of aramid fibers in theparticular area, leading to the loading pin to slidethrough the composite component.

The ring-frame configuration as commonly used infixed-wing aircraft, is a difficult component forcrashworthiness. NASA studies have shown, that the"point-load" applied by the ground leads to immediatefracture at this point, followed by severe bending, andfurther breakages of the frame higher up [7]. This mayresult in an early disintegration of the structure, andthe bypassing of the dedicated energy-absorbingmeasures. The recommendations followed from these

studies were, to separate the livable volume "on top"from the energy absorbing components "below".Hence, a sufficiently strong, closed ring-frame, meantto survive the impact, should be positioned on top ofan expandable energy absorbing structure. Thisconcept was adopted by both the commuter and theairliner configurations developed in CRASURV.

COMMUTER FUSELAGE DESIGN

Based on the lessons learned from the proventechnology summarised above, the following choiceswere made. For energy absorption, two longitudinalsine-wave beams were selected, using the materialconfiguration and trigger mechanism as used in theboxes (see Fig. 1), but adapted for the requireddimensions (height) and load. Closed frames, placedon top of the sine-wave beams, were obtained bycontinuing the super-floor frame sections into the floorbeams, and connecting the sub-floor frame sections tothe super-floor sections by means of discrete hinges,rather than (more complicated and unpredictable)pseudo-plastic hinges, see Fig. 2. This would preventthe transfer of bending moments from the sub-floorframe sections to the super-floor sections. The sub-floor frame sections had to be maintained to withstandthe cabin pressure.

The sub-floor frame sections were expected to breakat the ground impact point, and were configured with abreakable splice plate, see Fig. 3, to control thefailure. These frame sections were to be loaded bybending, due to the impact force, and the reactions ofthe hinges and the sine-wave beams, and possibly tobuckle. However, with the crushing of the sine-wavebeams by the dummy passenger mass placed above,and the ground reaction from below, the sine-waveforce on the sub-floor frame sections would diminish,and the frames might rotate around the hinges.However, this would require the skin to break near thehinges, which was expected to occur without the needto trigger this failure.

Originally, the sub-floor frame sections were thoughtto be connected to the sine-wave beams, alike thecruciform joints in the composite box structures.Hence, the sine-wave beams were placed at a slightangle, in order to follow the expected rotation of theframes. Sine-wave beams, even when placed at aslight angle, and tested as a singular, laterallysupported component, had been shown to absorb asignificant amount of energy [8, 9]. Moreover, acommuter aircraft should be able to crash at a modestbank angle without loosing its crashworthiness. In alater stage of the design, it was decided to eliminatethe connection between sine-waves beams andframes, see Fig. 4, to prevent the deforming sub-floorframe sections to destabilize the sine-wave beams.

-6-NLR-TP-2001-108

This design concept was subsequently evaluated atthe component level: the sine-wave beam (Fig. 5) andthe hinge-frame configuration (Fig. 6) with splice plate.NLR fabricated two specimens of each component,and tested one of each in static compression, whileDLR tested the other two specimens "dynamically".

In the first of the tests carried out by DLR, the functionof the frame assembly with a collapsible splice platecould be proved. The splice plate between the two stiffframe sections broke very early after the impact andthen the frames started to rotate around the hinges,exactly in the way they were designed. The only pointof concern that could be found in this dynamic testwas the bonded connection between the skin and theflange of the frame. This bonded joint failed totally bypeeling stresses, and caused disintegration of theframe and the skin. As an improvement of thefuselage design, a number of rivets were used in thesection which was eventually drop tested, in additionto the bonding connection, to avoid this type of failure.

The excellent crushing behaviour of composite sine-wave beams in combination with a ply drop-off triggerwas demonstrated in the second dynamic test carriedout at DLR (Fig. 5). The 350 mm high sine-wave beam(rather high compared to typical helicopter beams)failed along the trigger line and showed now damagefar away from the trigger area. A continuous crushingprocess with a nearly constant load level of about 50kN was observed, from the point where the web,broken at the trigger line and translated downwards,contacts the lower flange again. Compared to thequasi-static compression test carried out at NLR, thecrushing load is about 25% lower (Fig. 7). Based onthe results of the component tests, the design of thecommuter section was finalized by Alenia.

FABRICATION

The fabrication and final assembly of the commutersub-floor structure, carried out at NLR, consisted of anumber of pre-cured parts, which were bondedtogether, while some parts were also bolted together.The composite parts were the skin (two plates joinedtogether by adhesive bonding along the impact line),with secondarily bonded I-shaped stringers, four C-section frame halves with bonded and bolted spliceplates, two sine-wave beams, two lateral floor I-beams, each consisting of four parts, and two railtracks on top of the sine-waves, of similar crosssection as the floor beams. Aluminium brackets wereused for the hinges, the connection between the endsof the floor beams and the skin, and the connectionsbetween the floor beams and the seat tracks. Theconnections between sine-waves and skin, andbetween frames and skin were bonded and alsobolted (Fig. 8).

DROP TEST

Test configurationBecause of the structural geometry and therequirement that the fuselage section must impact theground in a vertical position, a guidance system wasdevised by CEAT which was able to lift the specimento the predetermined height, to release the structureby a pyrotechnic system, and to guide the structuredownward during its fall (Fig. 9). In order to take thefriction due to the guidance system into account,various tests were performed to determine the heightcorrection. Consequently, the structure was lifted to2.6 m above ground level in order to reach the verticalspeed of 7 m/s on impact, common to the speed atwhich the metal A320 section was tested in thepreceding Brite Euram programme. The selection ofthe dummy mass to represent the passengers andsuper-structure was controversial. For a successfultest result for code-validation, which was the primaryobjective, a lower mass was suggested by severalpartners, but to validate the (first-of-its-kind) structurefor the ultimate design requirements, a higher masswas proposed by other partners. A compromise wasfound: the impact energy was therefore 20 KJ, bycoincidence equal to that for the boxes, and twice theenergy applied later to the airliner sub-floor structure.

To simulate the masses of occupants, seats, and theupper part of the fuselage, the structure was loadedwith 40 kg at a position near the skin, and with 300 kgabove each sine-wave beam. Part of the guidancesystem also contributed to the load above the hinge.To avoid tilting of the inner and outer masses, two stiffbeams were used to connect the two masses. The300 kg mass located above the sine-wave beams wasdistributed along the full 800 mm length of the beams.The structural mass was 720 kg, including the 30 kgmass of the specimen. The loading principle is shownin Fig. 10.

InstrumentationThe instrumentation was configured, taking intoaccount the suggestions made by partners involved inthe pre-test structure and occupant simulation. Thus,different types of sensors were set up at various partsof the commuter structure in order to provide thepartners with test data to correlate with their analysisresults. Furthermore, the measurements wereintended to support the analysis of the structuralbehaviour and to estimate the structural energyabsorption which is needed to limit the load andacceleration of the occupants. The total number ofmeasurement channels was 39: 18 strain gauges atthe composite structure including the half frames andthe cross beams, 4 strain gauges and 6 displacementcells at the left sine wave beam, 4 strain gauges at theseat floor attachment and also at the load masses,

-7-NLR-TP-2001-108

and 3 load cells at the reaction platform. Various highspeed camera's were set up around the test facility:one video at 25 frames/s, two high speed motioncamera's at 500 frames/s, and one high speed motionvideo (1000 frames/s)

Failure modeThe actual impact velocity was 7.09 m/s, the energy atimpact 19866J. The commuter test resulted in anunexpected (and by the codes not fully predicted)failure mode, and failed without absorbing the requiredamount of energy. Upon impact, the splice-platesbroke correctly, and the sine-waves triggered correctlytoo. However, the now destabilized sine-wave websslid across the fuselage skin in the outward direction,rather than staying in plane to absorb energy bycompression. The out-of-plane movement of the websinduced significant unexpected torsion moments onthe (I-section) seat tracks, which failed unintendedly.As the skin did not fracture near the hinges, theframes were prevented to rotate, so that they bent,buckled and failed. The overall energy absorption wasvery low, and not by the mechanisms intended.However, the experimental data were useful to modifynumerical models, and to validate the codedevelopments. The failed structure, after being pulledup (it had been crushed flat), is shown in Fig. 11.

Lessons learnedThe major fault of the design was the fact that thesine-wave beams had no lateral support, for instanceby lateral sine-wave beams such as in the boxes. Animproved solution might have been found byconnecting the frames to the sine-wave beams, asoriginally intended. However, these frames cannotsupport the sine-wave beams over the full height,because they are much lower. Another solution mighthave been to provide a "stopper" for the sliding sine-wave beam web, but that is probably a less efficientmechanism. The trigger of the sine-wave beam wasstill providing a rather high peak force. A trigger withreduced strength might have improved the failuremode. The skin did not break near the hinges. Atrigger line, for instance provided by a rivet row, mighthave initiated failure of the skin, thereby improving thechance that the frame section would rotate asintended. The failure of the seat tracks was notimportant, because it was consequential to theundesirable failure mode.

NUMERICAL ANALYSIS

Several partners participated in the numerical analysisof the crash behaviour of the commuter structure. Thediscussion that follows corresponds to the pre- andpost-test analyses of EADS-CASA: pre-testsimulations with the PAM-CRASH code and post-test

simulations with the PAM-CRASH and LS-DYNA3Dcodes.

The mesh of the finite element model, used in allsimulations, is shown in Fig. 12. The structure wasmodelled with shell elements, with an average lengthof 10 mm. The various components of the model wereconnected by rivets and/or adhesive, which weremodelled as rigid body multi-point constraints withrupture criteria. The masses, representingpassengers, seats and the test guiding system weremodelled with solid elements, attached to the top ofthe model. To avoid penetration between of thecomponents during the collapse of the model, a self-contact algorithm was used. The ground was modelledas a rigid wall with a sliding surface, with a frictioncoefficient of 0.6. The vertical velocity of 7 m/s wasapplied to all nodes of the model, while a gravity fieldof 9.81 m/s was considered. Summarising, the modelhad approximately 75000 nodes, 71000 elements and12000 rigid bodies.

Several pre-test simulations were performed withPAM-CRASH, with the objectives to prove theeffectiveness of the energy absorbing concept, inorder to minimize the test risks, and to providepredictions for comparison with the experimentalresults. Sensitivity analyses were carried out withregard to material properties, total mass, rivet andadhesive failure and other aspects. The compositematerial model used was a multi-layered unidirectional"bi-phase" model (material 130 in PAM-CRASH),which allows to specify separately the ply propertiesfor fiber and matrix. The input data were taken fromreports generated by partners EADS-DBA, DLR andESI. For the metallic behaviour, elastic-plastic modelswith isotropic hardening were used (material 102 inPAM-CRASH). The results of the pre-test simulationsshowed that during the collapse sequence, thesinewave beams did not hit the ground perpendicularly(with the corresponding risk of bending and sliding),and the frames almost did not rotate around thehinges as expected by the designers. The contactbetween frames and sine-wave beam webs did notallow the trigger mechanism to work properly,especially near the frames. Figs. 13-14 show thedeformed shape at 60 ms, when the vertical speedhas decreased to zero, and the model starts torebound.

The sequence of events observed in the pre-testsimulations was partially in agreement with the testresults. The splice plate failure, the impact angle of thesinewave beams, and the functioning of the triggermechanism occurred at about the same time as in thetest. However, the friction coefficient between thesinewave beam and the skin was too high, whichprevented the web of the sinewave beams to slide,

-8-NLR-TP-2001-108

and the web of the seat track did not break as in thetest. As a result, the pre-test simulations predicted afar too large amount of energy absorption by the sine-wave beams, compared to what was observed in thetest.

For the post-test simulation with PAM-CRASH, themodel was slightly modified. In order to get a morerealistic failure (rupture), i.e., less ductile behaviour,the properties of the composite materials of theframes, the sine-wave beams and the seat trackswere embrittled. The ultimate failure loads of rivetsand adhesive, friction coefficients and elementelimination criteria (maximum strain) were decreased.In addition the actual test conditions were imposed:7.09 m/s and a total mass of 719 kg. As a result ofthese modifications, the analysis predicted a collapsesequence which was well in agreement with theobserved failure mode (see Figs. 13-14).

Subsequently, the post-test analysis model of PAM-CRASH was translated to a LS-DYNA3D model. Inthis case, the composite material model was adamage model with the Chang matrix failure criterion,combined with strain limiters (material 54 in LS-DYNA3D), which is able to perform a progressivesoftening (reduction of material strength and Young'smodules) of those elements, of which theneighbouring elements have already failed. The inputdata, generated by partners in the consortium, wereembrittled as for the PAM-CRASH post-testsimulations. The analysis results showed a similarcrash sequence as observed when using PAM-CRASH.

CONCLUSIONS

A design concept was developed for a sub-floorstructure of a fuselage, representative of an "ATR-type" commuter aircraft, made largely out of fiberreinforced composite material, with the requirement forthe structure to be crashworthy. As compositefuselages for such large aircraft, satisfying thisrequirement have never been built before, theexercise can be considered to be the first of its kind,and of a highly explorative nature. The structure wasbuilt and tested by dropping it at a vertical velocity of 7m/s, while loaded with dummy masses to representthe passengers and the structural weight of the superstructure. The test data was successfully used tovalidate the computer software developed for thesimulation of crashing composite aircraft structure.

The design concept was not successful, in that theenergy absorption capability was significantly lessthan foreseen. This capability was to be provided bythe controlled crushing of longitudinal sine-wavebeams. However, due to the lack of lateral support,

these beams slipped sideways, and escaped most ofthe compressive loading. In box type structures testedearlier, this phenomenon did not occur, whichindicates the importance of the presence of rigidlyconnected lateral support structure to stabilize thosecomponents which are dedicated to absorb energy ina compressive mode. It is believed, that given suchadditional lateral support, the energy absorbingconcept will work for commuter aircraft made ofcomposite materials. However, alternative designconcepts can be contemplated, possibly by developingentirely new and innovative energy absorbing frameconcepts.Based on the simulations of the commuter drop testsperformed by CASA, it can be concluded that PAM-CRASH and LS-DYNA3D have proved to be robusttools for this application. A major inconvenience wasthe large CPU time required, that in some cases wasclose to 30 days for a 90 ms simulation, which is aproblem that will be overcome with time. A majorproblem was the use of material data generated indetail tests, representative of failure modes whichwere not actually encountered in the drop test of thecommuter structure.The objective of the research programme was veryambitious: new computing capabilities were to bedeveloped and validated, by comparison withexperimental results obtained for novel compositestructures which had to be developed simultaneously.The experimental outcome of the design effortpresented here points at the need to dedicate a futureresearch programme entirely to the development ofsuccessful design concepts for crashworthy compositefuselage structures, now making use of the computersimulation capabilities developed within the Brite-Euram CRASURV programme.

ACKNOWLEDGEMENT

CRASURV – Design for Crash Survivability is an RTDproject, partially funded by the European Union underthe Aeronautics Area of the programme on Industrialand Materials Technology (BRITE/EURAM).

REFERENCES

1. Frijns, R.H.W.M., Wiggenraad, J.F.M., Schaefers,A.T.M., "Assessment of Survivable CrashScenarios During the Period 1980-1992”, NLR CR94012, 1994.

2. Brite-Euram project IMT Area 3 Aeronautics:"Crashworthiness for Commercial Aircraft" AERO-CT92-0030, 1993-1995.

3. Brite-Euram project CRASURV "CommercialAircraft - Design for Crash Survivability", CT96-0207, 1996-1999.

-9-NLR-TP-2001-108

4. Johnson, A.F., Kindervater, C.M., Thuis, H.G.S.J.,Wiggenraad, J.F.M., "Crash Resistant CompositeSubfloor Structures for Helicopters", AGARD FVP-Symposium: Advances in Rotorcraft Technology,Ontario, Canada, 1996.

5. McCarthy, M.A., Wiggenraad, J.F.M., "NumericalInvestigation of a Crash Test of a CompositeHelicopter Subfloor Structure", CompositeStructures 51, 2001.

6. Michielsen, A.L.P.J., et al, "Design, Test andAnalysis of Tensor Skin Panels for ImprovedCrashworthiness in Case of Water Impact",American Helicopter Society, Specialists' Meeting"Crash Safety Challenges and InnovativeSolutions", Phoenix, Az, USA, 1998.

7. Jackson, K.E., Kellas, S., Kindervater, C.M.,Luetzenburger, M., "Experimental and SimulatedCrash Responses of Composite AirframeStructures", American Helicopter Society 50thAnnual Forum, Washington, USA, 1994.

8. Michielsen, A.L.P.J., Wiggenraad, J.F.M., "Reviewof Crashworthiness research of CompositeStructures, Part 1: NLR-contribution", NLR CR97046 L, 1997.

9. McCarthy, M.A., Harte, C.G., Wiggenraad, J.F.M,et al, "Computational Analysis of Energy-Absorbing Composite Aerospace Structures underCrash Loading", International Symposium onComposite Materials, International Conference onComputational Engineering Science ICES '98,Atlanta, Ga, USA, 1998 .

Fig. 1 Sub-floor box structures (NLR)

Fig. 2 Design concept composite commuter

Fig. 3 Splice plate

Fig. 4 Frame-beam intersection

-10-NLR-TP-2001-108

Fig. 5 Test set-up for the dynamic sine-wavebeam test (DLR)

Fig. 6 Frame components with hinges

Fig. 7 Comparison of static and dynamic sine-wave beam tests (DLR)

Fig. 8 Final assembley of commuter sub-flor struct(NLR)

Fig. 9 Test set-up CEAT

Fig. 10 Loading principle (CEAT)

-11-NLR-TP-2001-108

Fig. 11 Post-test configuration (CEAT)

Fig. 12 Commuter FEM model for numerical simulations (CASA)

-12-NLR-TP-2001-108

Fig. 13 Test/simulations comparisons of defomred shape at total vertical speed equal zero (CASA)

Vz(c.g.) ≈ 0 m/s

Vz(c.g.) ≈ 0 m/s

Vz(c.g.) ≈ 0 m/s

(a)

(b)

(d)

(c)

-13-NLR-TP-2001-108

Fig. 14 Detail of SWB deformed shape at total vertical speed equal aero (CASA)

Vz(c.g.) ≈ 0 m/s

Vz(c.g.) ≈ 0 m/s

Vz(c.g.) ≈ 0 m/s

(a)

(b)

(c)