UNLIMITED C nd

UNCLASSIFIED

Ln
U)
'-U

Ln
Nq A REVIEW OF
CRASHWORTHINESS OF
COMPOSITE AIRCRAFT
STRUCTURES

by

C. Poon

National Aeronautical Establishment

et

AERONAUTICAL NOTE
OTTAWA NAE-AN-63
FEBRUARY 1990 NRC NO. 31276

I . ,fcr for ptu-'

1

SNational Research Conseil national

Council Canada de recherches Canada
 NATIONAL AERONAUTICAL ESTABLISHMENT

SCIENTIFIC AND TECHNICAL PUBLICATIONS

AERONAUTICAL REPORTS

Aeronautical Reports (LR): Scientific and technical information pertaining to aeronautics

considered important, complete, and a lasting contribution to existing knowledge.

Mechanical Engineering Reports (MS): Scientific and technical information pertaining to

investigations outside aeronautics considered important, complete, and a lasting contribution to
existing knowledge.

AERONAUTICAL NOTES (AN): Information less broad in scope but nevertheless of importance
as a contribution to existing knowledge.

LABORATORY TECHNICAL REPORTS (LTR): Information receiving limited distribution

because of preliminary data, security classification, proprietary, or other reaons.

Details on the availability of these publications may be obtained from:

Graphics Section,
National Research Council Canada,
National Aeronautical Establishment,
Bldg. M-16, Room 204,
Montreal Road,
Ottawa, Ontario
K1A OR6

ETABLISSEMENT NATIONAL D'AERONAUTIQUE

PUBLICATIONS SCIENTIFIQUES ET TECHNIQUES

RAPPORTS D'AERONAUTIQUE

Rapports d'aeronautique (LR): Informations scientifiques et techniques touchant

l'a6ronautique jugdes importantes, completes et durables en termes de contribution aux
connaissances actuelles.

Rapports de genie mecanique (MS): Informations scientifiques et techniques sur la

recherche externe A 'adronautique jugdes importantes, compltes et durables en termes de
contribution aux connaissances actuelles.

CAHIERS D'AERONAUTIQUE (AN): Informations de moindre portde mais importantes en

termes d'accroissement des connaissances.

RAPPORTS TECHNIQUES DE LABORATOIRE (LTR): Informations peu dissdmindes pour

des raisons d'usage secret, de droit de propridt( ou autres ou parce qu'elles constituent des donn6es
prdliminaires.
Les publications ci-dessus peuvent ftre obtenues 4 radresse suivante:
Section des graphiques,
Qonseil national de recherches Canada,
Etablissement national d'adronautique,
Im. M-16, pibce 204,
Chemin de Montr6al,
Ottawa (Ontario)
KIA OR6
 UNLIMITED
UNCLASSIFIED

A REVIEW OF CRASHWORTHINESS OF COMPOSITE

AIRCRAFT STRUCTURES

ETUDE SUR LA RESISTANCE A L'ECRASEMENT DES

STRUCTURES D'AERONEF EN MATERIAUX COMPOSITES

by/par

C. Poon

National Aeronautical Establishment/

Etablissement national d'adronautique

AERONAUTICAL NOTE
OTTAWA NAE-AN-63
FEBRUARY 1990 NRC NO. 31276

W. Wallace, Head/Chef
Structures and Materials Laboratory/ G.F. Marsters
Laboratoire des structures et mathriaux Director/directeur
 SUMMARY

A review of major research activities in North America with respect

to the crashworthiness of composite aircraft structures was performed with
the goal of identifying potential Canadian contribution to R&D areas where
effort would be required to complement the on-going programs in the United
States. e areas reviewed included the crashworthiness design criteria of
the U.S. y; major experimental programs undertaken by the FAA, the
U.S. Army, ASA, and Bell Helicopter Textron Inc. in the U.S., as well as
the UniversitX of Toronto in Canada in developing crashworthiness design
concepts for c posite structures; and the capabilities of crash dynamics
computer codes.Recommendations include a study on the effect of aircraft
size on crashworthiness design requirements; the implementation of the
KRASH code in Canada to establish commonality in analytical methods
with major U.S. and European users; an investigation on the energy
absorption capabilities of the design features for small aircraft containing
composite and/or composite/hybrid structures; and a parametric study
on the crashworthiness design of composite-to-composite and
composite-to-metal joints..j'-

RtSUM]k

Une dtude des principales activitds de recherche menses en Am~rique

du Nord sur la resistance A l16crasement des structures d'a~ronef en
materiaux composites a 06 effectude dans le but d'identifier les 6lments de
participation possible du Canada dans les domaines de recherches et
d6veloppement ou un effort serait requis pour appuyer les programmes
r6guliers des Atats-Unis. Les domaines 6tudids comprennent les critbres de
conception de r6sistance A l'dcrasement de I'U.S. Army; d'importants
programmes d'expdrimentation entrepris par la FAA, I'U.S. Army, la
NASA et Bell Helicopter Textron Inc. aux Etats-Unis, de m~me que par
l'University of Toronto au Canada visant A mettre au point des concepts de
conception de r6sistance A l'dcrasement pour les structures d'adronef; et les
possibilit6s d'exploitation des codes machines pour le calcul des forces
dynamiques lors de l'crasement. Les recommandations comprennent
une 6tude de l'effet de la dimension des a~ro-efs sur les exigences de
conception relatives A la resistance A l-cras( , e,_: l'implantation du code
KRASH au Canada pour uniformiser les me-"s analytiques avec les
principaux utilisateurs am6ricains et europdent., ine 6tude des capacitds
d'absorption d'dnergie des caractdristiques de conception des petits sion For
adronefs qui contiennent des structures en matdriaux composites et (ou) en
composites/hybrides; et une analyse param6trique des conceptions de GRA&I 1 "
rdsistance A l'6crasement des joints composite sur composite et composite TAB El
sur metal. iounoed
'fIcatio

/ ~.ia'\ Distribution/

(iii)
K ~
Availability Codes
Dis
t
Avail and/or
Special
/~IniL
 TABLE OF CONTENTS

Page

SUMMARY (iii)

LIST OF TABLES (v)

LIST OF FIGURES (v)

LIST OF ABBREVIATIONS (vi)

1.0 INTRODUCTION 1
1.1 Background 1
1.2 Report Organization 2

2.0 CRASHWORTHINESS DESIGN CRITERIA 2

3.0 ENERGY-ABSORBING COMPOSITE STRUCTURES 3

3.1 FAA Crash Dynamics Program 4

3.2 Bell Helicopter Textron Inc. (BHTI) Research Programs 5
3.3 U.S. Army / NASA Research Programs 6

4.0 CRASH SIMULATION OF COMPOSITE STRUCTURES 10

4.1 Computer Code KRASH 10

4.2 Computer Code DYCAST 11
4.3 Crash Dynamics Computer Code Developed at UTIAS 13

5.0 CONCLUSIONS AND RECOMMENDATIONS 14

6.0 ACKNOWLEDGEMENT 16

7.0 REFERENCES 16

(iv)
 LIST OF TABLES

1. U.S. Army Crash Impact Design Conditions .................................................... 21

2. Mechanical Properties of Fiber and Matrix Materials Used

in R eference 15 ....................................................................................................... 22

3. Mechanical Properties of Fiber and Matrix Materials Used

in R eference 16 ....................................................................................................... 22

4. Composite Frame Test Conditions ...................................................................... 23

5. Comparison of Pertinent Features of the Three KRASH Versions ............ 24

6. KRASH Experimental Verifications ................................................................... 25

LIST OF FIGURES

1. Key Parameters for Comparison of Crashworthiness Performance ............ 27

2. Load-Deflection Characteristics of Crushable Subfloor Structure

w ith "Triggering Device". ....................................................................................... 28

3. Effect of Matrix Failure Strain on the Energy Absorption of

Graphite Reinforced Composite Materials ...................................................... 29

4. Energy Absorption of AS4/5245C, AS6/F185 and AS6/HST-7

C om posite M aterial ............................................................................................... 30

5. Graphite-Epoxy Z-Shaped Frame Cross Section ............................................... 31

6. Failure Locations for Frames 4 and 5 ................................................................ 32

7. Floor Level Acceleration Pulse for Frames 2 and 5 ......................................... 33

8. Circumferential Flange Strains for Frames 1 and 2 ......................................... 34

9. KRASH and NASTRAN Models of Helicopter Composite Cabin ............... 35

KEYWORDS: Acceleration forces, composite structures, crash impacts,

crashworthiness, DYCAST, energy absorption, failure modes,
helicopter, KRASH, nonlinearity, post-crash integrity, small aircraft,
structural joints, thermoplastics.

(it)
 LIST OF ABBREVIATIONS

ACAP Advanced Composite Airframe Program

AVRADCOM Army Aviation Research and Development Command

BHTI Bell Helicopter Textron Inc.

CFE Self-Consistent Finite Element
D/t Inside Diameter to Wall Thickness Ratio
DYCAST Dynamic Crash Analysis of Structures

FAA Federal Aviation Administration

GL/E Glass/Epoxy
GR/E Graphite/Epoxy
K/E Kevlar/Epoxy
LMFE Lump Mass Finite Element
NAE National Aeronautical Establishment
NASA National Aeronautical and Space Administration
NRC National Research Council (Canada)
PEEK Polyetheretherketone
SMI Sypher:Mueller International Inc.

TC Transport Canada
TDC Transportation Development Centre
UTIAS University of Toronto Institute of Aerospace Studies

(vi)
A REVIEW OF CRASHWORTHINESS OF COMPOSITE AIRCRAFT STRUCTURES

1.0 INTRODUCTION

1.1 Background

The Transportation Development Centre (TDC) and the Director of Airworthiness,

Transport Canada (TC) commissioned the Sypher:Mueller Inc. (SMI) to review the status
of North American research and development in small aircraft crashworthiness. The object
was to identify the potential for Canada to play a role in developing crashworthiness
requirements for small aircraft and to develop a comprehensive R&D program which would
contribute to the enhancement of the safety of small aircraft. The study was completed in
July 1987 with the issue of Report TP8655E entitled "Small Aircraft Crashworthiness" by
the TDC [1].

Following he research study, a seminar was held on March 9, 1988 under the joint
sponsorship of the TDC and TC. The purpose was to explore the future focus of Canadian
R&D activity in small aircraft crashworthiness, to develop a coordinated R&D program
between government, industry, and the universities, and to enhance the joint R&D activities
between Canada and the U. S. The seminar was attended by all sectors of the Canadian
Aviation Community, including Transport Canada, National Research Council (NRC),
Aerospace Industries Association of Canada, Canadian Aviation Safety Board, University
of Toronto Institute of Aerospace Studies (UTIAS), Canadian Owners and Pilots
Association, plus representatives of the United States Federal Aviation Administration
(FAA), National Transportation Safety Board, and the General Aviation Manufacturer's
Association. Upon the recommendation of the Canadian participants, a Small Aeroplane
Crashworthiness R&D Committee was established under the initil's sponsorship of TC and
the National Aeronautical Establishment (NAE)/NRC. The role of the Committee is to
coordinate and advise TC on the conduct of R&D for the improvement of crashworthiness
of small aeroplanes which include normal, utility and aerobatic aeroplanes in Canada.

The first meeting of the Committee on June 1, 1988 established a task for SMI to
develop a small aeroplane crashworthiness R&D plan based on the seminar proceedings
and the conclusions and recommendations of the report TP8655E. In the second meeting
on August 29, 1989, the Committee approved the draft report "Crashworthiness R&D Plan"
prepared by SMI [2]. The report produced a list of five program areas. The program area
with the highest priority dealt with Seats, Harnesses, and Floor Sub-Structures while the
program area next in priority concerned Composite Materials and Structures. The
Committee recommended the establishment of a working team consisting of technical staffs
from TC and NAE to produce work statements in these two program areas. To respond to
the Committee's recommendation, the Structures and Materials Laboratory of NAE
initiated two in-house reviews of the research work performed in these two program areas.
This report presents the review of the work done to address the crashworthiness aspects of
composite structures.
 2

1.2 Report Organization

Most of the work covered in this review was performed in the U.S. under the
sponsorships of the FAA, U.S. Army and NASA. This review excluded the historical
evolution of regulatory requirements and development of accident analysis in small aircraft
crashworthiness since these two topics are well covered in Report TP8655E [1].

A general, comprehensive overview of crashworthiness research activities in both the

U.S. and Canada is provided in Report TP8655E [1]. This present review attempts to avoid
duplications and focuses on technical details of the research activities with respect to
crashworthiness of composite airframe structures. A description of the development of the
Aircraft Crash Survival Design Guide [3] by Simula Inc. for the U.S. Army and highlights
of crashworthiness design criteria of the MIL-STD-1290A [4] are given in Section 2.
Although the criteria discussed are considered significantly more demanding than those
required in the Federal Aviation Regulations, they are regarded as extremely valuable in
developing new crashworthiness standards including those for new materials. Section 3
presents an overview of major research activities addressing the energy-absorbing
characteristics of composite structures. It includes programs undertaken by the FAA, Bell
Helicopter Textron Inc. (BHTI), U. S. Army and NASA that range from testing of simple
cylinders to complex crash testing of full-scale structures. State-of-the-art computer codes
that are capable of crash simulation of composite structures are discussed in Section 4. The
last section is devoted to conclusions and recommendations.

2.0 CRASHWORTHINESS DESIGN CRITERIA

The Aircraft Crash Survival Design Guide [3] was prepared for the Safety and
Survivability Technical Area of the Applied Technology Laboratory, U.S. Army Research
and Technology Laboratories of the U.S. Army Aviation Research and Development
Command (AVRADCOM) by Simula Inc. A major portion of the data contained in the
Design Guide was taken from the U.S. Army sponsored research in aircraft crashworthiness
conducted from 1960 to 1979. Since its initial publication in 1967, the design guide has been
revised several times to incorporate the results of continuing research in crashworthiness
technology. The fourth revision of the Design Guide contains a comprehensive treatment
of all aspects of aircraft crash survival documented up to 1979 and its design criteria were
used to formulate the latest revision of the U.S. Army's aircraft crashworthiness military
standard MIL-STD-1290A(AV), published in September 26, 1988. Highlights of the design
criteria contained in MIL-STD-1290A(AV) [4] are presented in the following paragraphs.

The goal of a crashworthiness design of an airframe structure is to prevent occupant

fataiities and minimize the severity of injuries in survivable crash impacts. Survivable crash
impacts were defined as those in which the forces transmitted to the occupant through the
seat and restraint system do not exceed the limits of human tolerance to abrupt
accelerations and in which the structure in the occupant's immediate environment remains
substantially intact to the extent that a livable volume is provided for the occupants
throughout the crash sequence. Typical Army survivable crash impact conditions are shown
in Table 1. To achieve this goal, the airframe must be designed to maintain structural
 3
integrity and a livable space for the occupants. The airframe should also be designed to
possess high roll-over strength and high retention strength for large mass components. In
order to minimize the effects of post-crash hazards, the structure should be designed to
maintain the integrity of normal exits for emergency egress and provide protection against
flammable fluid systems.

Energy-absorbing features are required to prevent excessive and injurious

acceleration (G) forces from being transmitted to the occupants under the crash impact
conditions given in Table 1. The limits of human tolerance to abrupt accelerations can be
found in Volume II of the Aircraft Crash Survival Design Guide [3]. The forward fuselage
structure should be designed to absorb energy during longitudinal impacts and to minimize
soil plowing. In addition, the floor structure must be designed to maintain sufficient
strength througthout the crash to carry the loads applied by occupants, seats and cargo
restraint systems. A crushable sub-floor structure should be incorporated to absorb energy
in vertical crash impacts. Energy-absorbing seats and proper occupant restraint systems
should be provided to improve survivability and minimize the severity of injuries. The wings
should be designed to possess frangible characteristics such that they break free from the
fuselage under high impact loads. The landing gear should be designed to provide energy
absorption to reduce the vertical velocity of the fuselage.

In order to satisfy the crashworthiness design criteria discussed, a total systems

approach which includes a strong protective shell to protect the occupants from crushing
as well as energy-absorbing components to prevent injurious G forces must be adopted. It
is important to demonstrate that in replacing metals by composite materials in aircraft
structures, the capability to absorb energy and to maintain post-crash integrity is not
compromised. In the following section, the review is focused on the research activities in
the development of energy-absorbing composite structures which demonstrate compliance
with the crashworthiness design criteria.

3.0 ENERGY-ABSORBING COMPOSITE STRUCTURES

Polymer-based fiber composite materials offer significant benefits over metallic

structures in reducing weight and cost as well as improving fatigue and corrosion resistance.
However, these materials exhibit low strain-to-fai',ire ranges compared to such metals as
2024 aluminium, a ductile metal that can tolerate rather large strains, deform plastically,
and absorb a considerable amount of energy in the nonlinear region without fracture.
Because of this difference between metals and composites, crash energy absorption with
composites must come from innovative design to enhance stress-strain behaviour. Innovative
design concepts involving notched corners, corrugated and tapered edge joints, and less stiff
laminate layups were found to be effective in improving load-deflection characteristics of
composite structures.

Assembly operations such as mechanical joining, adhesive bonding, co-curing, and

other kinds of attachments contribute to crashworthiness. A further enhancement in energy
absorption and post-crash integrity can be achieved by optimizing the design of joints. The
design of composite joints is well known to be influenced by many parameters which affect
 4

not only the strengths of the joint hut also the failure modes [51. Test data presented in the
form of design curves that show the relationships between these parameters and joint
strengths are commonly applied in empirical or semi-empir:cal methods for the optimization
of joint designs. For crashworthiness considerations, the empirical relation.,,hips between
these parameters and energy absorption are also required for the optimization of joint
designs. A parametric investigation of adhesively bonded aluminiu m joints tinder impact-
tension loading was undertaken jointly by the Alcan International Ltd. and NAE [6]. In the
Alcan-NAE joint investigation, a conventional drop weight impact machine was modified
to test adhesively-bonded CouponsLinder tension loading. The effects of various parameters
including curing conditions. surface treatment procedures, adhesive systems, and joint
configurations on joint strength and energy absorption were investigated. The present review
showed that a similar parametric investigation involving composite joints had not been
undertaken.

Major research activities to develop design concepts for energy-absorbi ng helicopter

and transport aircraft composite structures have been undertaken in the U.S. and other
countries. Design concepts were evaluated through crash testing of composite test articles
that ranged from simple cylinders to full-scale structures and the results were compared to
those generated using metallic stri tures as benchmarks. The present review covered
significant research activities undertaken by the FAA, BIITI, U. S. Army and NASA.
3.1 FAA Crash Dynamics Program

One major area of the FAA sponsored research in crash dynamics addressed
crashworthiness considerations of composite materials [7]. The FAA's approach in assessing
crashworthiness designs of composite structures is that the comb)ination of composite
materials' behaviour and designs must exhibit equal or better energy absorption than the
metals which they replace. Key parameters which can be used as measures of performance
for comparing composite materi:als with metals have been established (Figure 1). The
parameters are specific energy, which considers energy and weight; load uniformity, which
indicates the relationship between peak arid average impact force: stroke ratio, which
provides a measure of the effective use of materials: energy dissipation density, which
indicates the degree of effectiveness in absorbing energy; ratio of dynamic to static forces,
which indicates strain rate effects; and crash stress, which is related to the forces and area
involved in the loading. The results of a study undertaken by NASA w,%ith respect to
application of composite materials to transport category airframe structures [81 illustrated
the manner in which designs and material behaviour affected these parameters.

In collaboration with NASA, the FAA performed drop tests of sections of narrow-
and wide-body airplane frames [9], a full-scale drop test of a complete test article
('Uaurinburg') [10], arid a remotely piloted airplane crash test (Controlled Impact
Demonstration test) [II]. Although these tests were on metallic structures, they generated
test data which could be used as a benchmark against those obtained from similar tests on
composite structures. Also in collaboration wvith NASA, the FAA sponsored the
development of the computer codes KRASI I and 1)YCAST for crash dynamics analysis (e"
both metallic arid composite structures. The characteristics of these codes will be discussed
in Section 4.
3.2 Bell Helicopter Textron Inc. (BHTI) Research Programs

In a BttTI in-house research program, initial investigations of the crash impact

behaviour of composite materials were focused on the testing of simple composite cylinders.
Crush testing of cylinders made of graphite/epoxy (GR/E), Kevlar/epoxy (K/E), and
fiberglass/epoxy (GL/E) was performed by Cronkhite [12] to determine basic compression
failure modes and energy absorption of these composites. He found that both the GR/E
and GL/E cylinders failed in a brittle fracturing mode while the K/E cylinders failed in a
progressive folding mode much like a ductile metal. The GR/E material showed the highest
specific energy absorption (energy absorbed per pound weight).

The BHTI in-house program continued to investigate various energy-absorbing

su)floor design concepts [13]. Of the concepts investigated, a K/E sandwich structure was
selected for the design and fabrication of crushable subfloors of composite airframes. This
decision was based on the results obtained from design-support testing of small keel beam
sections which indicated that the K/E sandwich construction demonstrated excellent stroke-
to-length ratios of up to 80% before bottoming out, good energy-absorbing and load-limiting
ability, good post-crushing structural integrity and no significant load rate sensitivity.

In a contracted R&D program conducted by BHTI for the Applied Technology

Laboratory of the U.S. Army Research and Technology Laboratories (AVRADCOM), the
energy-absorbing K/E sandwich subfloor structure was incorporated into two full-scale
composite cabin test sections and drop tested [14J. The floor structure utilized a crush
initiator or "triggering device" to prevent the occurrence of excessive peak loading prior
to the onset of controlled-load crushing (see Figure 2). The composite cabins were derived
from BHTI's Advanced Composite Airframe Program (ACAP) and were fabricated with
K/E, GR/E and GL/E materials. The KRASH computer code was used to design the
ACAP composite airframe to meet the crashworthiness requirements of the U.S. Army
MIL-STD-1290 [15]. The requirement for vertical crash impact velocity specified in MIL-
STD-1290 is 42 ft/s. Both cabin sections were actually drop tested at 30 ft/s. Since both
cabin sections were not equipped with landing gear, it was justified to adopt a lower drop
velocity by assuming that the landing gear had slowed the aircraft from 42 ft/s to 30 ft/s
prior to fuselage contact.

The first cabin section was drop tested at a 30 ft/s vertical impact velocity and at a
flat attitude with no pitch or roll. Floor and bulkhead mounted stroking seats and
anthropomorphic dummies were installed to study the effects of the crash on a human
occupant. The second cabin section was impacted at a 20 deg. roll attitude and at the same
velocity as the first test. The dummies and stroking seats were replaced with 177-pound
rigid lumped masses in the second drop test. Accelerations of the large masses were
measured in both drop tests and displacements of the stroking seats were recorded in the
flat drop test. In additior, high-speed motion pictures were used to record the cabin
structure response.

Both cabin drop tests successfully demonstrated that the U.S. Army's crashworthiness
requirements could be met by a composite airframe structure based on the excellent post-
 6
test condition of the cabins' protective shell structure and the performance of the energy-
absorbing components. The U.S. Army's Aircraft Crash Surviva1 Design Guide [3]
recommends that the vertical acceleration of the pelvis should not be above 23G for a
duration of longer than 6 msec in order to minimize the probability of injury to the
occupants. Both floor and bulkhead mounted seats stroked properly and attenuated the
decelerative loads on the dummies to the noninjurious levels recommended by the Design
Guide. However, the vertical acceleration of the floor exceeded the 23G level for about 30
msec. Therefore, without the stroking seats, the occupant would have probably experienced
injurious decelerative loads during the crash impact. This demonstrated the effectness of
adopting a systems approach in crashworthiness design in protecting the occupants: The
fuselage structure served to maintain a protective shell, provide energy absorption, and
retain the seats and large masses while the seat restrained the occupants, provided
additional energy absorption, and controlled decelerative loads on the occupant to tolerable
levels.

The computer codes KRASLI and DYCAST were used to simulate the crash
response of the composite cabin sections. Details of the KRASH and the DYCAST codes,
which are widely used in crash dynamics analysis, and a brief assessment of the simulation
capability of these two codes will be given in Section 4.

3.3 U.S. Army / NASA Research Programs

Farley of the U.S. Army Research & Technology Laboratories (AVRADCOM)

investigated the energy absorption characteristics of selected composite material systems and
compared the results with alurninium [16]. Cylindrical specimens were fabricated with both
tape and woven fabric prepregs using GR/E, K/E and GL/E materials and 6061
aluminium. Both static compression tests and drop-weight impact tests were conducted. The
cylinders were subjected to static or dynamic loading that was applied parallel to the axis
of the cylinders. Chamfering one end of the composite cylinders was adopted as a crush
initiator to prevent the excesssive buildup of peak loads at the onset of cylinder collapse
and to facilitate the occurrence of subsequent controlled-load crushing of the cylinder. The
results varied significantly as a function of material type and ply orientation. In general, the
GR/E cylinders absorbed more energy than the GL/E or K/E cylinders for the same ply
orientation. The [00/_+150 ] GR/E cylinders absorbed more energy than the aluminium
cylinders. GR/E and GL/E cylinders failed in a brittle mode and had negligible post-
crushing integrity, whereas the K/E cylinders failed in an accordian (folding) buckling mode
similar to the aluminium cylinders. The post-crushing integrity of hybrid composite
cylinders was also investigated and found not to be significantly better that of the single
material cylinders. Composite material systems investigated were found to be impact speed
insensitive up to 23 ft/s. Farley concluded that quasi-static test results were applicable to
low velocity impacts.

In a subsequent study, Farley [171 conducted static crushing tests on GR/E and K/E
cylinders to examine the influence of specimen geometry on the energy absorption capability
of composite materials. Cylinders with inside diameters ranging from 0.5 to 4 inches were
subjected to end loading that was applied parallel to the axis of the cylinder. One end of
each cylinder tested was chamfered to facilitate the initiation of the crushing process. The
 7
inside diameter to wall thickness ratio (D/t) was determined to significantly affect the
energy absorption capability. The GR/E and K/E cylinders exhibited a nonlinear and bi-
linear relation, respectively., hetween energy absorption and D/t ratio. The increase in
energy absorption as D/t ratio decreased was found to relate to a reduction in interlaminar
cracking. The energy absorption of K/E cylinders was determined to be geometrically
scalable whereas that of GR/E cylinders was not scalable.

Farley [18] conducted further static crushing tests on cylinders to evaluate the
influence of fiber and matrix maximum strain at failure on the energy absorption capability
of graphite reinforced composite material systems. Composite material systems investigated
were:

(1) Thornel 300/Fiberite 934 -- A low strain fiber in a low strain matrix.
(2) Thornel 300/Narmco 5245C -- A low strain fiber in a high strain matrix.
(3) Hercules AS4/Fiberite 934 -- A high strain fiber in a low strain matrix.
(4) Hercules AS4/Narmco 5245C -- A high strain fiber in a high strain matrix.

The material properties of the fiber and matrix materials used are presented in
Table 2. The T300 and AS4 graphite fibers have similar extensional moduli although the
AS4 fiber has approximately a 20% greater strain at failure than the T300 fiber. The 5245C
matrix material has approximately a 100% greater strain at failure than the 934 matrix
material. Test results shown in Figure 3 indicated that the fiber and matrix maximum strain
at failure significantly affected energy absorption. The higher strain-at-failure system,
AS4/5245C, exhibited the highest energy absorption capability among the four composite
systems tested. All material systems tested exhibited typical failure modes of brittle fiber
reinforced composite material systems. Two predominant failure modes were observed: the
bending crushing mode and the fracturing crushing mode. The former mode is a lower
energy-absorbing mode. For T300/5245C and AS4/5245C, the crushing mode of was
primarily a fracturing crushing mode. The crushing characteristics demonstrated by these
four material systems suggested that to achieve the maximum energy absorption from a
particular fiber, the matrix material must have a greater strain at failure than the fiber.

Farley's work described earlier did not investigate very high ultimate strain material
systems. In an extensive collaborative effort with NASA, Farley, et al. [19] conducted static
crushing tests on cylindrical specimens to evaluate the effects of very high ultimate failure
strain fibers (> .015 strain) and toughened matrices relative to energy absorption. In
addition, the effects of fiber stiffness, fiber volume fraction, lamina stacking sequence, and
hybrid material systems were assessed. The materials investigated and their mechanical
properties are shown in Table 3.

Static tests were conducted on cylinders fabricated from AS6/F185 and AS6/HST-
7. The ultimate failure strain of the AS6 fiber is 17% higher than that of the AS4 fiber used
in Reference 18. The F185 matrix has approximately 600% greater failure strain than the
5245C matrix used in Reference 18. The HST-7 matrix is an interleafed system where the
base matrix has similar properties as the 934 matrix while the interleafing material is a high
strain adhesive. Test results shown in Figure 4 indicate that the higher failure strain
material systems, AS6/F185 and AS6/HST-7, have lower energy absorption than the
 8
AS4/5245C used in Reference 18. The cylinders fabricated from AS6/FI85 with ply
orientations of [+45] and [+75], were found to exhibit a ductile folding mode similar to
that of Kelvar reinforced composites and have post-crushing integrity. The folding mode
for these two ply orientations was due to the suppression of the formation of interlaminar
cracks in the high strain matrix. The AS6/HST-7 material exhibited better energy
absorption capability than the AS6/F185 material but it exhibited less energy absorption
capability than the AS4/5245C material. The diameter of the AS6 fiber is smaller than that
of the AS4 fiber. This factor was considered to be the primary cause for the lower
compression strength of the two composite systems reinforced with the AS6 fibers when
compared with the AS4/5245C system.

The investigation of Farley, et al. [19] also found that energy absorption was
inversely proportional to fiber stiffness, the effect of fiber volume fraction on energy
absorption was less than 10% for both GR/E and K/E material systems, lamina stacking
sequence could influence the energy absorption of the material tested by a factor of 3, and
hybrid composite systems composed of graphite and Kevlar reinforcements resulted in
better energy absorption capability than aluminium while retaining post-crushing integrity.

The work carried out by Farely, et al. on cylinders has established that composite
material systems possess good energy absorption capability under compressive loadings.
However, under tensile or bending conditions, structural integrity may be lost at initial
fracture and energy absorption can be low. Under crash conditions, aircraft structural
elements experience complex loadings which are not always compressive. Therefore, a need
existed to examine generic composite structures under crash loadings.

Recognizing the limitations of compression testing on small cylinders and the need
to generate test data under complex crash loadings, NASA and the U.S. Army initiated a
joint reseach program to investigate the crashworthiness potential of composite materials
by performing full-scale drop tests on generic composite structural elements. In this
program, Boitnott, et al. [20] conducted drop tests on GR/E frames with a Z-shaped cross
section (see Figure 5). The dimensions of the cross section shown in Figure 5 were typically
used in designs proposed for fuselage structure of advanced composite transports. A
diameter of six feet for the frames was chosen to reduce specimen fabrication costs and to
facilitate testing. The complete circular frame was assembled from four 90 sections by using
mechanical fasteners and splice plates.

These frames were drop tested onto a concrete floor to simulate the crash conditions
shown in Table 4. Five tests were performed. In the first test, a 3/16 inch steel cable was
attached across the horizontal diameter of the frame to represent the constraint of a floor
on the lateral expansion of an impacted frame. In the second and the third tests, a 1-inch
diameter aluminium tube was used to represent this constraint more accurately. In the first
three tests, ten-pound masses were attached to the left and right sides of the frame at the
frame-floor intersection to represent structural and/or seat/occupant loads on the frame.
In the fourth test, the applied mass was increased to 100 pounds by replacing the aluminium
bar with an 80 pound steel bar. In the fifth test, a semi-circular frame was constructed using
the upper half of the fourth frame since no visual damage was observed and the strains
measured in the upper halves of the frames used in previous tests were generally very low.
 9
A steel bar similar to the one used in the fourth test but slightly longer was used to
simulate the floor constraint and to provide seat/occupant mass. In the first two tests, the
frames were oriented so that impact occurred in the lower half of the frames at a splice
plate. In subsequent tests, the frames were rotated 450 so that impact occurred in the lower
half of the frames at a location midway between splice plates. Accelerometer, strain gauge,
and high-speed photographic measurements were made and these data were used to
characterize the impact behaviour of frames with different masses to represent structural
or seat/occupant masses.

Failures of all GR/E frames tested involved complete separations through the cross
section in the lower half of the frames. In the first two tests which involved impacts at a
splice plate, failures of the lightly loaded composite frames were confined to an area 150
to the right of the impact point. In the third test which involved impact at a location
midway between the splice plates, the lightly loaded frame suffered failure at the impact
point. In the last two tests which involved impacts at a location midway between the splice
plates of the more heavily loaded frames, the frames failed first near the impact point and
then subsequently failed once on the left side and twice on the right side of the impact
point (see Figure 6).

Typical filtered acceleration pulses measured at the frame-floor intersection for

frames 2 and 5 are shown in Figure 7. Peak floor level accelerations of 63 G's and 15 G's
were determined for frames 2 and 5 respectively. The maximum dynamic load at the floor
level on each of the frames may be determined by multiplying the peak floor level
acceleration by the floor mass. Using this method, the dynamic loads experienced by all the
frames were found to be nearly constant for the greatly varying test conditions applied in
this investigation.

All strain gauges indicated an initial pulse and initiation of frame failure occurred
at this peak. For frames 1 and 2, circumferential flange strains (see Figure 8) were found
to be highest near the impact point but were determined not to be high enough to cause
a material failure. Instead the frame failure was initiated by a local structural instability.
Strains measured by the gauges mounted on the flanges and web of frames 3 and 4 depicted
the complexity of the structural deformations as a result of the impact. Strain values
obtained near the impact point indicated that the strains were tensile on the inside flange
and compressive on the outside flange. High radial bending strains measured by strain
gauge rosettes on the web clearly showed that the original Z cross section was distorted.
High in-plane shear strains near the impact point were also measured with these rosettes.
In general, low strains were measured in the upper half of the frames.

The experimental floor level accelerations and strain values were used to calibrate
the results obtained from an analysis of the response of the frames using the finite element
code DYCAST. A review of the DYCAST code and its qualification against crash test
results of aircraft structures will be presented in Section 4.
 10

4.0 CRASH SIMULATION OF COMPOSITE STRUCTURES

An analytical tool is very important to the aircraft designers who are involved in
evaluating and optimizing the crashworthiness of aircraft structures. Using such a tool the
designer can simulate the complex behaviour of the structures under various impact loads
without resorting solely to expensive and time consuming scale model testing. In order to
be cost effective, the analysis must adopt the simplest feasible mathematical model
representation of the actual structural response under crash conditions, while maintaining
an acceptable level of accuracy. Because of the complexity of the problem, computer
techniques are required to adequately address large deflections, nonlinear material
response, local buckling and post-buckling behaviour, as well as component fractures. The
following section will present the review of the simulation capability of three computer
codes; two of the codes, the KRASH and DYCAST codes, have been developed in the U.S.,
while the third code has been developed at the UTIAS in Canada.

4.1 Computer Code KRASH

Under contract to the U.S. Army at the onset of the 1970's, the computer code
KRASH was developed by the Lockheed California Company to provide an analytical
capability to determine helicopter structural dynamic responses to multidirectional crash
impact forces [21]. This analytical capability was required to support crashworthiness design
trade-off studies. Subsequent to the U.S. Army's sponsored efforts, KRASH was upgraded
under an FAA contract and its capability was directed toward the analysis of light fixed
wing aircraft subjected to crash impact conditions [22]. The upgraded version is called
KRASH79 and contains many new features while retaining its original concept. The
completion of the FAA sponsored research in 1985 [23] has resulted in the current version,
denoted KRASH85, which is capable of simulating crash scenarios of transport category
aircraft. A comparison of the pertinent features of the three KRASH versions is shown in
Table 5.

The computer code KRASH [24] can be used to perform a nonlinear transient
response analysis to simulate the crash impact behaviour of any arbitrary three-dimensional
structure. The analysis includes both geometric and material nonlinear structural behaviour
capability. KRASH is often referred to as a "hybrid" analysis method because it generally
requires input data derived from other analyses or tests since the structure is represented
in a rather coarse manner using nonlinear beam and spring structural elements and lumped
masses. The code integrates the Euler equations of motion of the lumped masses connected
together by the beam members, each with six degrees of freedom, and computes the time
histories of accelerations, velocities, and displacements. In addition, using small deflection
linear analysis or large deflection plastic analysis, the internal beam forces, shears, moments
and torsions are computed. The loads and deflections of the external springs used to
simulate those portions of the aircraft coming into contact with the ground are also
determined.

For the design of crashworthiness structures, the appropriate KRASH code is often
used in conjunction with a conventional finite element program such as NASTRAN. For
example, in the design of the composite cabin sections in Reference 14, KRASH85 was used
 11
for the dynamic analysis of the cabin drop test conditions and NASTRAN was used for
determining internal loads required for strength analysis (see Figure 9). The NASTRAN
analysis was conducted using load factors from KRASH85 based on a "snapshot" of the
dynamic loads at points in time when the loads were critical. The strength analysis
performed using the NASTRAN internal loads was used for sizing the structure above the
floor of the cabin where the deformation behaviour was expected to be mainly elastic. The
sizing of the crushable subfloor structure was based on energy absorption and load
attenuation requirements from the KRASH85 analysis using load-deformation characteristics
of key energy absorbing components derived from design support test data.

During the course of the development of the KRASH codes, extensive comparisons
have been made between test and analysis results, both by Lockheed and various other
users [11, 21, 25-38]. A partial list of configurations and test conditions for which
correlations have been performed is shown in Table 6. Furthermore, KRASH has been
applied by numerous aircraft manufacturers to evaluate a wide range of design
configurations. The development of KRASH has been greatly benefited from, and been
accelerated by the positive cooperation of an international user community. For example,
as noted in Reference 38, the KRASH79 code was modified by Aerospatiale to adapt to
specific requirements and Lockheed, in collaboration with Aerospatiale, included several
of the changes in the KRASH85 code. The feedback from KRASH users continues to be
the best source of updating, improving and expanding the code as an analytical tool to
support design trade-off studies. While efforts are continuing to expand its capability and
applicability, the current usage of the KRASH code has indicated that a satisfactory level
of acceptance has been achieved.

4.2 Computer Code DYCAST

DYCAST (Dynamic Crash Analysis of Structures) is a nonlinear, dynamic, finite

element computer code developed by the Grumman Aerospace Corporation with principal
support from NASA and the FAA as part of the combined NASA/FAA program for
aircraft crashworthiness [39]. This code has the capability of modelling stringers, beams,
bulkheads, and structural surfaces such as skins fabricated from isotropic as well as
orthotropic materials. In order to provide the above modelling capability, the DYCAST
code maintains a basic element library consisting of stringers with axial stiffness; beams with
axial, two shear, torsion, and two bending stiffnesses; isotropic and orthotropic membrane
skin triangles with in-plane normal and shear stiffnesses; isotropic plate bending triangles
with membrane plus out-of-plane bending stiffnesses; and nonlinear springs with axial
stiffness and a user specified load-deflection curve. The last element type can be used as
either an elastic, dissipative or gap element. Under crash loadings, the changing stiffnesses
in the structure are accounted for by plasticity (material nonlinearities) and very large
deflections (geometric nonlinearities). The material nonlinearities are accommodated for
by one of the three elastic-plastic options available in the code. Geometric nonlinearities
are handled in an updated Lagrangian formulation. The stiffness variations due to structural
failures are computed based on a failure option which is imposed automatically whenever
a criterion for failure strain of the material is met or manually set by the user. The overall
accuracy and computational cost of the simulation depends on the number of elements and
degrees of freedom used. Refining the model may improve simulation accuracy, but always
 12
drives up the costs. Therefore, the analyst must have sufficient expertise in modelling the
structure for the nonlinear crash analysis in order to produce sufficiently accurate results
within acceptable time and cost restraints.

The total cost of an analysis is composed of the labour involved in creating the
model and evaluating the results and the cost of using the computer. Based on the
experience of Winter, et al. [40], a first time full vehicle finite element model could require
up to four person-months of effort to prepare and verify, depending on the model size and
complexity. However, after the first time model is complete, it can be modified easily and
at small cost, enabling the investigation of the effects of structural modifications. The
computational costs are also dependent on model size and complexity. For a linear stress
and vibration analysis, a finite element model could contain more than 20,000 degrees of
freedom. Such a model would probably have already existed for the conventional design and
analysis of the structure. However, the use of such an existing model, instead of creating
a new one especially prepared for crash analysis, could save much time and labour in model
preparation, but because of the limitations of current computers, a nonlinear dynamic
analysis of this type of model using the DYCAST code is currently not feasible because of
the computer time required. The optimal model size for a nonlinear finite element problem
is therefore greatly dependent on the speed and memory capacity of the computer. Again
based on the experience of Winter, et al. [40], a model having between 2000 and 3000
degrees of freedom and requiring several thousand time steps could complete the analysis
overnight on the vector computers, such as the Cray-1 and the Cyber 205, using an in-core
solution with a memory of 2 million words. Larger problems or smaller computer memory
capacity and speed will require restart runs and longer time to complete the analysis. It is
obvious that the finite element model must be designed to fit the available computer
resources and includes all the important details for an accurate crash simulation.

In order to model the structure within available computer resources and produce
accurate results, the DYCAST code allows the use of simple nonlinear spring elements to
model crushable, energy absorbing components which exhibit extremely nonlinear
behaviour. The advantage is that instead of requiring a large number of plate elements
involving several thousand degrees of freedom for each collapse zone, only one degree of
freedom is added for each nonlinear spring. However, this hybrid method requires the
analyst to specify the expected large deformation behaviour by providing an input curve of
load vs deflection or moment vs rotation generated by conducting component testing.

Applications of the DYCAST code include crash simulations of an aircraft fuselage

subfloor [41], composite helicopter cabin sections [14], composite fuselage frames [20], an
aircraft seat [42], and transport fuselage sections of the joint FAA/NASA Controlled Impact
Demonstration test [30]. Of relevant interest to the present review is the accuracy of the
crash simulation of composite structures using the DYCAST code. As mentioned earlier,
the DYCAST analytical results were compared with the test results of the crash testing of
the composite fuselage frames in the joint U.S. Army/NASA program 1201. As noted in
Reference 20, only fair correlation was obtained between predicted and experimental frame-
floor level accelerations. Frame response predicted by the DYCAST analysis showed
significant sensitivity to moderate changes in the failure strain. Additional analytical studies
are required to assess further the capabilities and possible improvements needed to analyze
 13

composite structural elements using the DYCAST computer code.

As mentioned earlier, both the KRASH and DYCAST computer codes were
evaluated in the BHTI research program on composite helicopter fuselage technology [141.
These two codes were used to simulate the crash responses of the full-scale composite
cabins under flat drop test and 20' roll drop test conditions. The analysis results obtained
from these codes were correlated with the drop test results. In addition, the KRASH code
was used to design the composite cabins for the drop tests in accordance with the
crashworthiness requirements of the MIL-STD-1290. Based on the good crash impact
performance of the composite cabins and satisfactory correlations of analytical and test
results, KRASH proved to be a useful and reasonably accurate analysis tool for the
crashworthiness design of composite structures. The accuracy of the DYCAST predictions
of damage, deformation, and accelerations was generally good for the flat drop test.
However, the agreement between DYCAST and test results for the 20 roll drop test was
mixed, with fairly good agreement in deformation and generally poor agreement in
acceleration predictions.

4.3 Crash Dynamics Computer Code Developed at UTIAS

Under the joint sponsorship of the TDC and the Natural Sciences and Engineering
Research Council of Canada, a comprehensive program was undertaken at the University
of Toronto Institute of Aerospace Studies (UTIAS) to investigate, both analytically and
experimentally, the dynamic behaviour of aircraft fuselage structures subjected to various
impact conditions [43]. Two analytical models were developed to derive the motion
equations, one based on a self-consistent finite element (CFE) technique [44] and the other
on a more approximate, lumped mass finite element approach (LMFE) [45]. Extensive
vertical drop and free-flight impact tests of scale model fuselage sections of stiffened
aluminium were conducted using a pendulum gantry developed at UTIAS for a wide range
of wing loads, angles of incidence and impact velocities [44]. Test data were obtained in
terms of structural strains, G loads and high speed photography of the modes of dynamic
collapse. Analytical results obtained by using the CFE and LMFE techniques demonstrated
good correlations with experimentally generated G loads, dynamic peak strains and transient
failure modes.

In a subsequent phase of the UTIAS program, partly funded by the Canadian

Defence Research Establishment Suffield and the Natural Sciences and Engineering
Research Council of Canada, a finite element computer model for analysing the crash
response of stiffened composite fuselage structures was developed [46]. As demonstrated
in Reference 46, the finite element computer model established by using a formulation
based on Reissner/Mindlin plate theories can treat stiffened laminated shell buckling, large
deflections, nonlinear material behaviour and element failure. Although experimental
comparisons were not yet available, numerical results presented [46] for several test cases
clearly showed that this code could determine transient displacements, velocities and
accelerations. As reported in Reference 47, a composite fuselage model one meter in
diameter by two meters in length was constructed and flight impact tests were to be
conducted following a numerical analysis of the model for given load conditions. It was also
stated that a non-destructive program was necessary to precede the final crash test to
 14

optimize the amount of test data that could be obtained from such an expensive test article.
However, a recent report indicated that flight impact tests of the composite fuselage had
not been conducted [46]. Reference 46 was the most recent report which describes the flight
impact test status of the composite fuselage at UTIAS.

5.0 CONCLUSIONS AND RECOMMENDATIONS

A review of major research activities in North America related to crashworthiness

of composite structures was conducted. The intent of the review was to identify potential
Canadian contributions to research on the crashworthiness of small aircraft containing
composite structures and to propose research to complement the on-going activities in the
U.S. This review effort encompassed the work undertaken or sponsored by the FAA, the
U.S. Army, NASA, Bell Helicopter Textron Inc., Lockheed California Company and
Grumman Aerospace Corporation in the U.S. and Transport Canada/Sypher:Mueller
International Inc. and the University of Toronto Institute of Aerospace Studies in Canada.

The conclusions of this review have led to specific recommendations for further
activities which will provide long and short term benefits in improving the crashworthiness
of small aircraft containing composite structure. These proposed activities are concentrated
on the dynamic impact testing of composite structures and on computational structural
techniques where expertise already exists in Canada. Consequently, it is believed that these
research activities have the potential to achieve their objectives. In order to maximize the
chance of achieving the research objectives, these activities should be conducted in a
coodinated manner involving several centers of expertise across Canada since the resources
of most Canadian organizations are very stretched and too limited to make any effective
contribution alone. The conclusions and the corresponding recommendations are provided
as follows:

1. The research activities in crashworthiness design of composite structures pursued

in the U.S. were highly focused on rotary wing and transport type aircraft. These
research activities have generated a data base which will have applicability towards
the establishment of crashworthiness design guidelines for small aircraft containing
composite and/or composite/metal hybrid structures. However, it is considered
extremely difficult if not impossible to directly relate crashworthiness design criteria
and regulatory requirements developed from research on these types of aircraft to
small aircraft containing composite and/or composite/metal hybrid structures
because of size effects. Research on composite cylinders indicated that GR/E
material sytems were not geometrically scalable. This review has shown that
insufficient research is being directed, both in the U.S. and Canada, into the impact
dynamics of small aircraft containing composite and/or composite/metal hybrid
structures. Therefore, size effects on the crashworthiness behaviour of composite
structures should be addressed before the application of the current crashworthiness
design guidelines to composite small aircraft. Before the concern on size effects is
resolved, full-scale crash testing of representative composite structures should be
conducted.
 15

2. This review indicated that a parametric study of the effects of the design variables
affecting composite joints on failure mode, strength and energy absorption under
dynamic loading had not been undertaken. Since a parametric study can provide
useful insight and data for the optimization of composite joint design with respect
to crashworthiness, it is highly recommended as an immediate research item.

3. This review has shown that the approach of using appropriate testing in conjunction
with supportive analyses to assess design concepts of crash dynamics is feasible and
could become a practical design methodology for composite small aircraft. The first
recommendation will generate a data base to calibrate the analysis methods. State-
of-the-art crash dynamics analysis codes reviewed include the KRASH, DYCAST
and the code developed at the UTIAS. All three codes are capable of simulating
crash behaviour of both metal and composite structures and have proved to be
valuable analytical tools for crashworthiness design trade-off studies. The KRASH
code is the most widely used code and has an international user base. Any R&D
program would require the KRASH code to be commissioned in Canada because
this will establish commonality in analytical tools with the U.S. as well as European
users. Subsequent to the implementation of the KRASH code, an effort must be
made to gain experience in using the tool for the crashworthiness design of small
aircraft containing composite and/or composite/metal hybrid structures and to
develop a suitable material data base which is of interest to the Canadian aviation
community. This effort is necessary because the KRASH model is rather coarse and
requires test data and considerable experience to simulate crash scenarios.

4. Since the KRASH code cannot provide the internal load distribution of a structure,
a complementary structural analysis tool is required for strength prediction. Both the
U [IAS and DYCAST finite element simulations allow a more detailed and complete
representation of the structure but require considerable modelling setup time and
computer run time. Further work is recommended to explore the possibility to use
the two simulations to complement each other, that is, the KRASH code could be
used for overall vehicle response and the UTIAS or DYCAST for modelling critical
areas of the structure where a more detailed structural representation and analysis
is required. This could lead to a more cost-effective analysis.

5. The capability of the UTIAS code to simulate the crash behaviour of composite
structures has not been subjected to experimental verification. A comparison between
the UTIAS code and other codes has not been conducted. In order to use the
UTIAS code or to use the UTIAS code with KRASH for the design of crashworthy
composite aircraft structures, the capability of the UTIAS code must be verified
against experimental results. Also, it is considered worthwhile to compare the UTIAS
code with DYCAST and KRASH.

6. This review clearly shows that the crashworthiness behaviour of an aircraft depends
on the materials used, as well as the construction and the shape of the structural
elements. Preferred materials in crash application must exhibit non-brittle behaviour
so that energy is absorbed through plastic deformation and maintain post-crash
structural integrity. Conventional GR/E and GL/E material systems were shown to
 16
be brittle under crush loading. However, K/E as well as composite systems with high
strain-to-failure graphite fiber and toughened thermoset matrices were found to
exhibit ductile folding mode similar to aluminium and to possess post-crushing
integrity. The energy absorption and failure mode were found to be affected by
hybrid construction, lamina stacking sequence, lay-up, fiber stiffness, matrix strain at
failure and fiber volume fraction. A crush initiator or "triggering device" resulting
from chamfering the end of composite cylinders or corrugated web design with
notched attached angles in crushable subfloor construction was found to be effective
in preventing excessive buildup of peak loads before the onset of controlled-load
crushing. Since small aircraft containing composite and/or composite/hybrid
structures may exhibit different dynamic responses under crash conditions when
compare with helicopters and transport type aircraft, it is important to catalogue the
detail design features for small composite/hybrid aircraft and to investigate their
energy absorption capabilities. This recommendation will lead to the establishment
of design data bases for input to analytical models.

7. The U.S. Army MIL-STD-1290A should be used as a basis for the development of
crashworthiness design guidelines for small aircraft containing composite structures.
The data to be generated from the recommended research items described in this
section would be useful for the modification of the design criteria contained in MIL-
STD-1290A to suit the requirements for composite aircraft structures.

6.0 ACKNOWLEDGEMENT

This work was carried out under the sponsorship of NAE Structures and Materials
Project 07336, Fatigue and Damage Tolerance of Structures.

7.0 REFERENCES

1. Monroe, R. L. and McLeish, W., "Small Aircraft Crashworthiness," Technical

Development Center Report TP8655E, July 1987.

2. Monroe, R. L. and McLeish, W., "Crashworthiness R&D Plan", Sypher:Mueller

International Inc., to be published as a Transport Development Center Report.

3. Desjardins, S.P., Lannanen, D.H. and Singley, D.T., III, "Aircraft Crash Survival
Design Guide," USARTL-TR-79-22A to -22E, Vol. I to V, 1980.

4. Military Standard, MIL-STD-1920A(AV), "Light Fixed- and Rotary-Wing Aircraft

Crash Resistance," Department of Defense, Washington, D. C., 26 September 1988.

5. Poon, C., "Literature Review on the Design of Composite Mechanically Fastened

Joints," NAE-AN-37, National Research Council Canada, Ottawa, February 1986.
 17

6. Poon, C. and Farahbakhsh, B., "Impact Testing of Adhesively-Bonded Aluminium

Joints," NAE-LTR-ST-1670, National Research Council Canada, Ottawa, June 1988.

7. Wittlin, G. and Caiafa, C., "Transport Airplane Crash Simulation, Validation and
Application to Crash Design Criteria," Proceedings of the AGARD Conference on
Energy Absorption of Aircraft Structures as an Aspect of Crashworthiness,
Luxembourg, AGARD-CP-443, May 1988.

8. Jackson, A. C., "Transport Composite Fuselage Transmission," NASA CR 4035,

December 1986.

9. "DC-10 Fuselage Drop Test Report," Report No. 7252-1, Arvin/Calspan Report
prepared for the FAA Technical Center, Atlantic City, N.J., March 1984.

10. Johnson, D. and Garodz, L., "Crashworthiness Experiment Summary, Full-Scale

Transport Controlled Impact Program," DOT/FAA/CT-85/20, June 1986.

11. Wittlin, G., "KRASH Analysis Correlation -- Transport Airplane Controlled Impact
Demonstration Test," DOT/FAA/CT-86/13, December 1986.

12. Cronkhite, J. D. and Berry, V. L., "Crashworthy Airframe Design Concepts --

Fabrication and Testing," NASA CR3603, Washington, D.C., September 1982.

13. Cronkhite, J. D., Berry, V. L., and Winter, R., "Investigation of the Crash Impact
Characteristics of Htelicopter Composite Structures," USAAVRADCOM TR-82-D-
14, Applied Technology Laboratory, U.S. Army Research and Technology
Laboratories, Fort Eustis, Virginia, February 1983.

14. Cronkhite, J. D., "Design of Airframe Structures for Crash Impact," Proceedings of
the American Helicopter Society Specialists' Meeting on Crashworthy Design of
Rotorcraft, Atlanta, Georgia, April 1986.

15. Carnell, B. L. and Pramanik, M., "ACAP Crashworthiness Analysis by KRASH,"

American Helicopter Society Journal, Vol. 29, 1984, pp. 38-42.

16. Farley, G. L., "Energy Absorption of Composite Materials," Journal of Composite

Materials, Vol. 17, May 1983, pp. 267-279.

17. Farley, G. L., "Effect of Specimen Geometry on the Energy Absorption Capability
of Composite Materials," Journal of Composite Materials, Vol. 20, July 1986, pp.
390-400.

18. Farley, G. L., "Effect of Fiber and Matrix Maximum Strain on the Energy Absorption
of Composite Materials," Journal of Composite Materials, Vol. 20, July 1986, pp. 322-
334.
 18
19. Farley, G. L., Bird, R. K. and Modlin, J. T., "The Role of Fiber and Matrix in Crash
Energy Absorption of Composite Materials," Proceedings of the American Helicopter
Society Specialists' Meeting on Crashworthy Design of Rotorcraft, Atlanta, Georgia,
April 7-9, 1986.

20. Boitnott, R. L., Fasanella, E. L., Calton, L. E. and Carden, H. D., "Impact Response
of Composite Fuselage Frames," SAE Technical Paper No. 871009, presented at the
General Aviation Aircraft Meeting and Exposition, Wichita, Kansas, April 28-30,
1987.

21. Wittlin, G. and Gamon, M. A., "Experimental Program for the Development of
Improved Helicopter Structural Crashworthiness Analytical and Design Techniques,"
Lockheed-California Company, USAAMRLDL-TR-72-72, Volumes I and II, May
1973.

22. Wittlin, G. and Gamon, M. A., "Full-Scale Crash Test Experimental Verification of
a Method of Analysis for General Aviation Airplane Structural Crashworthiness,"
Lockheed-California Company, FAA Report FAA-RD-77-188, February 1978.

23. Gamon, M. A., Wittlin, G. and LaBarge, W. L., "KRASH85 User's Guide --
Input/Output Format," Lockheed-California Company, DOT/FAA/CT-85/10, May
1985.

24. Wittlin, G.,"Program KRASH: Evolution of an Analy, -al Tool to Evaluate Aircraft
Structural Crash Dynamics Response," Proceedings of the American Helicopter
Society Specialists' Meeting on Crashworthy Design of Rotorcraft, Atlanta, Georgia,
April 1986.

25. Badrinath, Y. V., "Simulation Correlation and Analysis of Structural Response of a

CH-47-A to Crash Impact," USARTL-TR-78-24, August 1978.

26. Mens, J., "Suirvey of Crashworthiness Achievements on Aerospatiale Helicopters,"

41st Annual Forum, American Helicopter Society, Fort Worth, texas, May 1985.

27. Cronkhite, J. D. and Berry, V. L., "Investigation of the Crash Impact Characteristics
of Helicopter Composite Structures," USAAVRADCOM-TR-82-D-14, February
1983.

28. Hayduk, R. J., Thomson, R. G., Wittlin, G. and Kamat, M. P., "Nonlinear Structural
Crash Dynamics Analysis," SAE Paper 790588, April 1979.

29. Wittlin, G. and Lackey, D., "Analytical Modelling of Transport Aircraft Crash
Scenarios to Obtain Floor Pulses," FAA Report DOT/FAA/CT-83/23, NASA CR
1666089, April 1983.
 19

30. Wittlin, G. and Caiafa, C., "Transport Airplane Crash Simulation, Validation and
Application to Crash Design Criteria," Proceedings of the AGARD Conference on
Energy Absorption of Aircraft Structures as an Aspect of Crashworthiness,
Luxembourg, AGARD-CP-443, May 1988.

31. Och, F., "Crashworthiness Activities on MBB Helicopters," Proceedings of the

AGARD Conference on Energy Absorption of Aircraft Structures as an Aspect of
Crashworthiness, Luxembourg, AGARD-CP-443, May 1988.

32. Nitschke, D. and Frese, J., "Crushing Behaviour of Helicopter Subfloor Structures,"
Proceedings of the AGARD Conference on Energy Absorption of Aircraft Structures
as an Aspect of Crashworthiness, Luxembourg, AGARD-CP-443, May 1988.

33. Jarzab, W. and Schwarz, R., "Crashworthiness of Aircraft Structures," Proceedings

of the AGARD Conference on Energy Absorption of Aircraft Structures as an
Aspect of Crashwortxiness, Luxembourg, AGARD-CP-443, May 1988.

34. Mens, J. and Bianchini, J. C., "Computing Codes for Development of Helicopter
Crashworthy Structures and Test Substantiation," Proceedings of the American
Helicopter Society Specialists' Meeting on Crashworthy Design of Rotorcraft,
Atlanta, Georgia, April 1986.

35. Pramanik, M. B. and Carnell, B. L., "Landing Gear Performance Simulation by

KRASH Program," Proceedings of the American Helicopter Society Specialists'
Meeting on Crashworthy Design of Rotorcraft, Atlanta, Georgia, April 1986.

36. Wittlin, G., "Aircraft Crash Dynamics: Modelling, Verification and Application," in
Structural Crashworthiness, Edited by Jones, N. and Wierzbicki. T., Butterworth and
Co., London and Boston, 1983, pp. 259-282.

37. Lamborn, L. C., "Crew Seat Stroke Requirements for Helicopter Rolled Attitude
Impact Crashworthiness," Proceedings of the American Helicopter Society Specialists'
Meeting on Crashworthy Design of Rotorcraft, Atlanta, Georgia, April 1986.

38. Soltis, S., Caiafa, C., Wittlin, G., "FAA Structural Crash Dynamics Program Update -
- Transport Category Aircraft," SAE Paper 851887, Long Beach, CA., October 1985.

39. Pifko, A. B., Winter, R. and Ogilvie, P. L., "DYCAST -- A Finite Element Program
for the Crash Analysis of Structures," NASA CR 4040, Washington, D.C., January
1987.

40. Winter, R. and Pifko, A. B., "Modelling Strategies for Finite Element Crash
Simulation of Complete Vehicles," Proceedings of the American Helicopter Society
Specialists' Meeting on Crashworthy Design of Rotorcraft, Atlanta, Georgia, April
1986.
 20

41. Carden, H. D. and t-ayduk, R. J., "Aircraft Subfloor Response to Crash Loadings,"
SAE Paper 810614, 1981.

42. Alfaro-Bou, E., Williams, M. S. and Fasanella, E. L., "Determination of Crash Test
Pulses and their Application to Aircraft Seat Analysis," SAE Paper 810611, 1981.

43. Tennyson, R. C., Nanyaro, A. P., Teichman, H. C. and Hansen, J. S.,

"Crashworthiness of Light Aircraft Structures," Transport Canada Research and
Development Centre Report TP3927, 1981.

44. Tennyson, R. C. and Hansen, J. S., "Study of the Crash Behaviour of Aircraft
Fuselage Structures," in Structural Crashworthiness, Edited by Jones, N., and
Wierzbicki, T., Butterworth and Co., London and Boston, 1983, pp. 218-258.

45. Tennyson, R. C., Teichman, H., Nanyaro, A. P. and Mabson, G., "Study of the
Crashworthiness of Light Aircraft Structures," Proceedings of the Twenty Second
Structures, Structural Dynamics and Materials Conference,
AIAA/ASME/ASCE/AHS, Atlanta, Georgia, 1981.

46. Hansen, J. S. and Tennyson, R. C., "The Dynamic Behaviour of Stiffened Composite
Fuselage Shell Structures," Proceedings of the AGARD Conference on Energy
Absorption of Aircraft Structures as an Aspect of Crashworthiness, Luxembourg,
AGARD-CP-443, May 1988.

47. Johnston, J., "Design and Fabrication of a Scaled Composite Fuselage Section,"
University of Toronto, M.A.Sc. Thesis, 1984.
 21

LU
a,
z

o C> Nj U) 0 N r-0
Nm N cn) oN -

0<
-J (
>u z
0

z
0

0 a) )
I-i < ~ . -1 Mu 5 z
00 0 -0%-C j

-0 0

Ew a..

C~ M)
ou-~~
0 0 UC .

C CcCUL
0 0 - -

~50 0 :5) C CC

ot 0 0 ) Cc
CD
M- -0
W
- - -
~C -

00 c 00)

a) 0LLa 0 c
 22

EXTENSIONAL TENSILE
MATERIAL MODULUS, E DENSITY FAILURE
DESCRIPTION GPa (Msi) g/cm 3 (lb/in 3 ) STRAIN

T300 231.5 (33.5) 1.75 (0.063) 0.012

AS-4 235.0 (34.0) 1.80 (0.065) 0.015

934 4.0 (0.58) 1.40 (0.046) 0.010

5245 3.8 (0.55) 1.25 (0.045) 0.020

TABLE 2 MECHANICAL PROPERTIES OF FIBER AND MATRIX

MATERIALS USED IN REFERENCE 18

MATERIAL EXTENSIONAL DENSITY p, ULTIMATE

DESCRIPTION STIFFNESS, E FAILURE STRAIN
GPa (Msi) g/cm 3 (lb/in 3 ) cm/cm (in/in)

FIBER

T300 231.5 (33.5) 1.75 (0.063) 0.012

AS-6 235.0 (34.0) 1.80 (0.065) 0.018

P55 380.0 (55.0) 2.00 (0.072) 0.005

F75 520.0 (75.0) 2.00 (0.072) 0.004

Kevlar-49 124.0 (18.0) 1.44 (0.052) 0.028

MATRIX

934 4.00 (0.58) 1.40 (0.050) 0.010

974 3.90 (0.57) 1.27 (0.046) 0.020

F185 2.00 (0.29) 1.26 (0.045) 0.130

HST-7* 3.80 (0.55)** 1.40 (0.050)** 0.010"*

* Interleafed matrix material

Appropriate values

TABLE 3 MECHANICAL PROPERTIES OF FIBER AND MATRIX

MATERIALS USED IN REFERENCE 19
 23

TEST IMPACT ADDED CONFIGURATION CENTER

NO. VELOCITY MASS RESTRAINT
FPS LBS

1 26.1 20 CABLE

2 27.5 20 TUBE

3 27.5 20 TUBE

4 20 100 BAR

5 20 93 7 BAR

TABLE 4 COMPOSITE FRAME TEST CONDITIONS (REF. 20)

 24

FEATURES KRASH KRASH 79 KRASH 85

1. ENERGY DISTRIBUTION YES YES* YES**

2. ELEMENT RUPTURE YES YES* YES*
3. INJURY CRITERIA (DRI)a YES YES* YES
4. PLOT CAPABILITY/SUMMARIES YES YES YES**
5. VOLUME PENETRATION YES YES YES
6. PLASTIC HINGE ALGORITHM NO YES YES*
7. SHOCK STRUT NO YES YES
8. FLEXIBLE AND/OR SLOPED TERRAIN NO YES YES
9. ACCELERATION PULSE EXCITATION NO YES YES
10. UNSYMMETRICAL BEAM REPRESENTATION NO YES YES
11. STANDARD MATERIAL PROPERTIES NO YES YES
12. EXTERNAL SPRING DAMPING NO YES YES
13. MASS LOCATION PLOTS NO YES YES
14. PRE- AND POST-DATA PROCESSING NO YES YES
15. RESTART CAPABILITY NO YES YES
16. SYMMETRICAL MODEL CAPABILITY NO YES YES
17. CG FORCE MOTIONHISTORY NO YES YES*
18. VOLUME CHANGE CALCULATIONS NO YES YES
19. STANDARD NONLINEAR CURVES NO 5 6
20. STIFFNESS REDUCTION FEATURE (KR)b NO NO YES
APPPLICABLE TO DAMPING
21. COMBINED FAILURE LOAD (LIC)c NO NO YES
22. INITIAL BALANCE-NASTRAN NO NO YES
23. TIRE VERTICAL SPRING NO NO YES
24. ARBITRARY MASS NUMBERING NO NO YES
25. EXTERNAL FORCE LOADING NO NO YES
26. OLEO METERING PIN NO NO YES
27. ADDITION OF DESCRIPTIVE NAMES TO NO NO YES
IDENTIFY INPUT DATA

Enhanced one level, ** Enhanced two levels

a Dynamic response index b Stiffness reduction factor c Load interaction curve

TABLE 5 COMPARISON OF PERTINENT FEATURES OF THE

THREE KRASH VERSIONS (REF. 24)
 25

GROSS IMPACT VELOCITY (m/sec)

AIRCRAFT WEIGHT REFERENCE
(kg) VERTICAL LONGITUDINAL LATERAL

ROTARY WING
Utility type 3909 7.0 - 5.6 21

Cargo typo 11045 12.8 8.3 - 25

Multi-purpose 1727 6.0 6.0 - 26

Multi-purpose 1645 10.0 - - 26

Composite substructure 1605 9.1 - 27

Composite substructure 1605 8.6 - 3.1 27

LIGHT-FIXED-WING

Single-engine, high-wing 1091 14.0 21.3 - 22

Single-engine, high-wing 1091 6.7 21.7 - 22

Single-engine, high-wing 1091 14.9 21.3 - 22

Single-engine, high-wing* 1091 13.1 21.2 - 22

Twin-engine, low-wing
substructure 248 8.4 - - 28

TRANSPORT

Medium size* 72272 5.5 5.24 - 29

Medium size 88636 5.3 79.2 - 11

* Test performed on soil; all other tests on rigid surface.

TABLE 6 KRASH EXPERIMENTAL VERIFICATION (REF. 24)

 27

P.-
a CD C

Cd)
- 0 -

LU CDCD

cm U H CD C:,) CD.'

<- t C..
rr w

tU- U)
U./) C/)

C)/

wo

CL

CD 0

cc a- U

ED, CD 0u
CD M X==CD
;_-- C-L en En

MD
le .
U.' 2-U'C)

Im CD = =)
.. '~~~.. .~- - ... n =' L-3
a- LUL)JU

cm C C

SCD~~

IL
 28

FLOOR

_450 4NOMEX HONEYCOMB CORE

KEVLAR/EPOXY
FABRIC
CURVED WEB-FLANGE
A'TACHMENT TO
INITIATE CRUSHING

STRUCTURAL

BULKHEAD FLOOR

u -- CRUSH ZONE

CONTROLLED LOAD STRUCTURE

UNCONTROLLED HIGH PEAK LOAD

CAN DAMAGE STRUCTURAL SHELL

CONTROLLED LOAD STRUCTURE

(HIGH ENERGY ABSORPTION, CONTROLLED LOADS)
0
,-I

CONVENTIONAL STRUCTURE
(LOW ENERGY ABSORPTION, HIGH PEAK LOADS)

DEFLECTION

FIG. 2: LOAD-DEFLECTION CHARACTERISTICS OF CRUSHABLE SUBFLOOR

STRUCTURE WITH "TRIGGERING DEVICE" (REF. 13)
 29

x103
125(500)
100(400)-
75(300) -

SPECIFIC 50(200) -
SUSTAINED 0 T300/5245
CRUSHING 25(100) 0 T300/934
STRESS, 0 AS-4/5245
N-m (a) [0/15] 4 (b) [0/45] 4 (c) [0/75] 4 A AS-4/934
3
kg x10
125(500)
lbin 100(400)
lbm 75(300)
50(200)
25(100) T300
o -N 030 N ....AS_,t
...
1.0 2.0 1.0 2.0 1.0 2.0........AS-4
ULT
MATRIX
(d) [15] 6 (e) [45] 6 (f) [75] 6

FIG. 3: EFFECT OF MATRIX FAILURE STRAIN ON THE ENERGY ABSORPTION OF

GRAPHITE REINFORCED COMPOSITE MATERIALS(REF. 18)
 30

CCD

Co~

U-')

2.2

0 OO
-t C.O4
00

Sn
2a IflE
Lu
LZ

CLu
U)d
 31

. 1- 1.16"--

.250" R

.080" - '- 3.00"

FIG. 5: GRAPHITE-EPOXY Z-SHAPED FRAME CROSS SECTION (REF. 20)

 32

c"~j

Itto

ww

I0

z
0

0
IL

134

ww

CV LL

LnI
 33

2 U. V
'U
(0

U, z
w C
20
-I U)
0. L
LV

2U. U.0

U.
N

FZ
0
P

CM Lu
LuU

CLu
N 0U
LI.U

0. 0

00n
 34

L.)U

CL) CM4

V) z
E C.

E 2N

(V - LL
E )
o 0
LU C.

L C.,l _ C..j

L..L

*"I
CV -

(V aI
EE
 35

NOTE: FLOOR, SKINS, BULKHEADS, AND ROOF

SEAT 'BEAMS MODELED WITH BEAM ELEMENTS
ATTENUATOR

UPPER BODY
SPINAL BEAM ROOF BEAMS
PELVIS

BULKHEAD

ATTENUATING SEAT SIDE SKINS

AND DUMMY MODEL
USED IN 30-FT/SEC
FLAT DROP ANALYSIS

CRUSHABLE __
SUBFLOOR
SPRINGS
(LOAD-DEFLECTION
FROM TEST)

IKRASH DYNAMIC MODEL

ACTUAL STRUCTURE

NOTE: APPLIED INERTIA LOADS

AND GROUND REACTION
LOADS FROM KRASH
NASTRAN INTERNAL LOADS MODEL

FIG. 9: KRASH AND NASTRAN MODELS OF HELICOPTER

COMPOSITE CABIN (REF. 14)
 REPORT DOCUMENTATION PAGE / PAGE DE DOCUMENTATION DE RAPPORT

REPORT/RAPPORT REPORT/RAPPORT

NAE-AN-63 NRC No. 31276

la lb

REPORT SECURITY CLASSIFICATION DISTRIBUTION (LIMITArIONS)

CLASSIFICATION DE SECURITE DE RAPPORT

Unclassified Unlimited

TITLE/SUBTITLE/TITRE/SOUS-TITR E

A Review of Crashworthiness of Composite Aircraft Structures

4

AUTHOR (S)/AUTEUR(S)

C. Poon
5
SERIES/SERIE

Aeronautical Note
6

CORPORATE AUTHOR/PERFORMING AGENCY/AUTEUR D'ENTREPRISE/AGENCE D'EXECUTION

National Research Council Canada
National Aeern ucl ta nt Structures and Materials Laboratory
7 National Aeronautical Establishment

SPONSORING AGENCY/AGENCE DE SUBVENTION

DATE FILE/DOSSIER LAB. ORDER PAGES FIGS/DIAGRAMMES

COMMANDE DU LAB.

90-02 41 9
9 10 11 12a 12b

NOTES

13

DESCRIPTORS (KEY WORDS)/MOTS-CLES

1. Crashworthiness 3. Energy Absorption

14
2. Composite Structures

SUMMARY/SOMMAIRE

A review of major research activities in North America with respect to the

crashworthiness of composite aircraft structures was performed with the goal of identifying
potential Canadian contribution to R&D areas where effort would be required to complement the
on-going programs in the United States. The areas reviewed included the crashworthiness design
criteria of the U.S. Army; major experimental programs undertaken by the FAA, the U.S. Army,
NASA, and Bell Helicopter Textron Inc. in the U.S., as well as the University of Toronto in
Canada in developing crashworthiness design concepts for composite structures; and the
capabilities of crash dynamics computer codes. Recommendations include a study on the effect of
aircraft size on crashworthiness design requirements; the implementation of the KRASH code in
Canada to establish commonality in analytical methods with major U.S. and European users; an
investigation on the energy absorption capabilities of the design features for small aircraft
containing composite and/or composite/hybrid structures; and a parametric study on the
crashworthiness design of composite-to-composite and composite-to-metal joints.
15