Ii1111111111111111lll

Itl~l!
7 AD-A263 lll
618 - a

WL-TR-92-3098

CRACK FORMATION IN F-15 AIRCRAFT CANOPIES

Geoffrey J. Frank
Gregory J. Stenger

University of Dayton Research Institute

Dayton, Ohio 45469-0110 D TIC
ELECTE

E U
MAY0
S
5 1993
=
October 1990

Interim Report for Period May 1990 - September 1990

Approved for public release; distribution is unlimited.

FLIGHT DYNAMICS DIRECTORATE

WRIGHT LABORATORY
AIR FORCE SYSTEMS COMMAND
WRIGHT-PATTERSON AIR FORCE BASE, OHIO 45433-6553

93-09608
 ,, ,,, I , I I I I

NOTICE

When Government drawings, specifications, or other data are used for any purpose other
than in connection with a definitely Government-related procurement, the United States
Government incurs no responsibility or any obligation whatsoever. The fact that the government
may have formulated or in any way supplied the said drawings, specifications, or other data, is
not to be regarded by implication, or otherwise in any manner construed, as licensing the holder,
or any other person or corporation; or as conveying any rights or permission to manufacture, use,
or sell any patented invention that may in any way be related thereto.

This report is releasable to the National Technical Information Service (NTIS). At NTIS,
it will be available to the general public, including foreign nations.

This technical report has been reviewed and is approved for publica.ion.

Guy Graening, ILt, USAF Duncan A. Dversdall, Capt, USAF

Project Engineer Supv, A/C & Windshield System
Program Office

Richard F. Colcloug
Chief, Vehicle Subsystems Division
Flight Dynamics Directorate
Wright Laboratury

If your address has changed, if you wish to be removed from our mailing list, or if the
addressee is no longer employed by your organization please notify WL/FIVR, WPAFB, OH
45433-6553 to help us maintain a current mailing list.

Copies of this report should not be returned unless return is required by security considerations,
contractual obligations, or notice on a specific document.
WL-TR-92-3098

CRACK FORMATION IN F-15 AIRCRAFT CANOPIES

Geoffrey J. Frank
Gregory J. Stenger

University of Dayton Research Institute

Dayton, Ohio 45469-0110

October 1990

Interim Report for Period May 1990 - September 1990

Approved for public release; distribution is unlimited.

FLIGHT DYNAMICS DIRECTORATE

WRIGHT LABORATORY
AIR FORCE SYSTEMS COMMAND
WRIGHT-PATTERSON AIR FORCE BASE, OHIO 45433-6553
 NOTICE

When Government drawings, specifications, or other data are used for any purpose other
than in connection with a definitely Government-related procurement, the United States
Government incurs no responsibility or any obligation whatsoever. The fact that the government
may have formulated or in any way supplied the said drawings, specifications, or other data, is
not to be regarded by implication, or otherwise in any manner constru-,d, as licensing the holder,
or any other person or corporation; or as conveying any rights or permission to manufacture, use,
or sell any patented invention that may in any way be related thereto.

This report is releasable to the National Technical Information Service (NTIS). At NTIS,
it will be available to the general public, including foreign nations.

This technical report has been reviewed and ýs approved for publication.

J%
Guy Graening, ILt, USAF Duncan A. Dversdall, Capt, ['SA\
Project Engineer Supv. A/C & Windshiejd System
Program off ice

Richard E. Colciou-
Chief, Vehicle Subsystems• Division
Flight Dynamics Directorcte
Wright Laboratory

If your address has changed, if you wish to be removud from cur m:Yiling list, or if the
addressee is no longer employed by your organization please notify WL/FIVR, WPAFB, OH
45433-6553 to help us maintain a current mailing list.

Copies of this report should not be returned unless return is required by security considerations,
contractual obligations, or notice on a specific document.
 Form Approved
REPORT DOCUMENTATION PAGE OMB No. 0704-0188
jOolhcreporting burden for this collection of information is estimated to iverage 1 nour oer response, including the time 0or reviewing instructions, searching existi•g data •ources.
gatherting and maintaining the data needed. and comL.aeting and reven.o, zte colletiron of information Send comments regaroing this burden estimate or any other asoect of this
collection of information, including suggestions for reducing tins burden, to Washington Headouarters Services. Directorate for information Operations and Reports. 121S ,efterson
Davis Highway, Suite 1204, Arlington. VA 22202.4302, and to the Office of Management and Budget. Pilennork Reduction Project (0704.0i8s), Washington. DC 20503

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
October 1990 I nterim - May 1990 to September 1990
4. TITLE AND SUBTITLE S. FUNDING NUMBERS
CRACK FORMATION IN F-15 AIRCRAFT CANOPIES C F33615-84-C-3404
PE 64212F
6. AUTHOR(S) PR 1926
TA 01
Geoffrey J. Frank and Gregory J. Stenger WU 10

7. PERFORMING ORGANIZATION NAME(S) AND ADORESS(ES) 8. PERFORMING ORGANIZATION

University of Dayton Research Institute REPORT NUMBER
300 College Park
Dayton, Ohio 45469-0110 UDR-TR-90-110

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING, MONITORING

Flight Dynamics Directorate AGENCY REPORT NUMBER
Wright Laboratory
Wright-Patterson Air Force Base, Ohio 45433-6553 WL-TR-92-3098

11. SUPPLEMENTARY NOTES

12a. DISTRIBUTION ,' AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Approved for public release; distribution is unlimited.

13. ABSTRACT (Maximum 200 words)

Cracks have been reported in F-15 transparencies. The cracks extend from the free
edge of the transparency into the acrylic parallel to the transparency surface.
Similar cracks were reported on F-4 transparencies in the early 1970's. Investi-
gations at that time indicated that the crack growth had two phases: crack initia-
tion and crack growth. Crack initiation was attributed to absorption and desorptior
of moisture. Crack growth was caused by thermal stresses due to the difference in
coefficients of thermal expansion between the stretched acrylic transparency and thE
fiberglass edge attachment. Tests have been performed on samples from F-15 canopie!
to validate the crack growth mechanism. Research on the F-4 cracks showed that
crack growth does not occur if nylon-acrylic composite, which has a coefficient of
thermal expansion similar to that of stretched acrylic, is used as an edge
reinforcement material.

14. SUBJECT TERMS 15. NUMBER OF PAGES

F-15, transparency, acrylic, crack growth, absorption, 27
thermal expansion coefficient, thermal stress, nylon 16. PRICE CODE

17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19 SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT
OF REPORT OF THIS PAGE OF ABSTRACT
UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED UL
NSN 7540-01-280-5500 Standard Form 298 (Rev 2-89)
Prmcred by ANSI Std 139S-I
298-102
 FOREWORD

The efforts reported herein were performed by the Aerospace

Mechanics Division of the University of Dayton Research
Institute, Dayton, Ohio, under Air Force Contract F33615-92-C-
3400. This work was sponsored by the Flight Dynamics Laboratory,
Wright Laboratories, Wright-Patterson Air Force Base, Ohio. Air
Force administrative direction was provided by Capt. Paul J.
Kolodziejski, WRDC/FIVR, and Russell E. Urzi, WL/FIVR, Project
Engineers.

The work described herein was conducted during the period

May 1990 through September 1990. University of Dayton project
supervision was provided by Mr. Dale H. Whitford, Supervisor,
Aerospace Mechanics Division, and Mr. Blaine S. West, Head,
Structures Group.

Accesion For
NTIS CRA&I
DTIC TAB
Unannounced 0
Justification

Distribution I
Availability Codes
I Avail and Io
Dist special

i-5
 TABLE OF CONTENTS

SECTION PAGE
1 INTRODUCTION 1

2 PHYSICAL APPEARANCE OF CRACKS 3

3 SUMMARY OF F-4 INVESTIGATION 5

3.1 Bearing Strength Evaluation 5
3.2 Crack Initiation Studies 6
3.3 Crack Propagation Studies 6

4 CRACKrP OpAeATION VERIFICATION 8

5 SUM IY AND CONCLUSIONS 11

REFERENCES 13

APPENDIX:g D,'R. Mulvilte, I. Wolock, and R. J.

Thoma 1, "Crack Formation in F-4 Aircraft
Canopies" 14

iv
 LIST OF ILLUSTRATIONS

FIGURE PAGE
1 Edge Attachments for F-4 and F-15 Canopies 2
2 Typical Edge Crack Patterns in F-15 Canopy 4
3 Stress in Thickness Direction at F-15 Edge 9
4 Stress Intensity at the Tip of 0.05 in.
Crack in F-15 9

LIST OF TABLES

TABLE PAGE
1 Crack Propagation Tests on F-15 Edge
Attachment 10

v
 SECTION 1
INTRODUCTION

The formation of cracks was observed in the stretched

acrylic of F-15 canopies and windshields during service. The
cracks start at the free edge of the transparencies and extend
into the acrylic parallel to the surface of the transparency.
The cracks generally occur near the mid-surface of the acrylic.
and were not observed to extend as far as the line of bolt holes
near the edge of the transparencies. The majority of the cracks
are tightly closed, but the cracks on some windshields are
visibly open.

Similar edge cracks were previously investigated for F-4

canopies [1]. The F-4 canopies were made of stretched acrylic
with fiberglass edqe attachments similar to those of the F-15.
The cracks on F-4 canopies occurred during storage and extended
the entire width of the edge attachment. The results of the
investigation of the F-4 cracks are summarized below, and the
entire article (Reference 1) is included as an appendix. The
geometry and materials for the F-4 and F-15 edge attachments are
shown in Figure 1. Although the canopy edge configurations are
not identical, they are similar enough to indicate that the
results of the F-4 investigation are applicable to the F-15. A
limited number of experiments was performed to verify that the
crack propagation mechanism presented in the earlier work applies
for the F-15 transparencies.
 Fiberglass-acryiic cloth

• 090 , -
LT

.090'- 2.625
Stretched Acrylic

F-4 Canopy Edge Attachmen- (Ref )

Fiberglass-phenoh c cloth

.060
I-- ! 0

.060 _ 1.375 Stretched Acryiic--\

F-15 Canopy Edge Attachment

Figure 1. Edge Attachments for F-4 and F-15 Canopies.

 SECTION 2
PHYSICAL APPEARANCE OF CRACKS

Specimens were cut from the edges of both forward and aft
canopies. The cracks appeared to be typical of those examined oi
several canopies. The cracks on one sample are shown in Figure
2, where the edge attachment has been removed. Microscopic
evaluation of the edge cracks showed a large number of
microcracks originating &.t tool marks and abrasions on the edge.
Most of these cracks are less than 0.05 in. deep. A single
larger crack usually occurs near the mid-surface of the canopy.
The larger cracks have alternating regions of cyclic marks,
indicating fatigue, and river pc-tterns, indicating a tensile
overload. In none of the samples examined does the crack front
continue beyond the bolt holes. Also, cracks which originate at
the bolt holes do not propagate outside of a region near the bolt
hole. This indicates that compression from the bolts arrests the
crack growth.

3
Figure 2. Typical Edge Crack Patterns in F-15 Canopy.

4
 SECTION 3
SUMMARY OF F-4 INVESTIGATION

An investigation of edge cracks which formed in F-4 aircraft

canopies during storage was divided into 3 parts: bearing
strength evaluation, determination of conditions for crack
initiation, and determination of conditions for crack
propagation. The F-4 edge crack investigation is presented in
the appendix. A summary of the results is presented below.

3.1 Bearing Strength Evaluation

Tensile tests were conducted to evaluate the change in

bearing strength of the edge attachment due to the presence of
in-plane cracks (Appendix, p.14). Test specimens were
manufactured from samples of canopies and from new material as a
control. Tests were conducted on samples with and without
fiberglass edge attachments and with nylon edge attachments. The
results indicated:

1. The bearing strength oL specimens with edge attachments was

not decreased by the presence of in-plane cracks.

Stretched acrylic specimens with fiberglass edge attachments

actually tailed in tension at the tapered junction of the
edge attachment and acrylic due to stress concentrations.

iCanopy specimens with fiberglass edge attachments failed at

loads 1/3 less than those for the control specimens without
edge attachments.
Stretched acrylic specimens without edge attachments failed
in bearing at a higher loads than specimens with edge
attachments.

17)
3.2 Crack Initiation Studies

Tests were conducted to determine the effect of temperature

and moisture on stretched acrylic (Appendix, p. 15). The edge
attachment specimens used had no initial cracks. Results were:

1. No cracks were initiated by heating the specimens to 160IF

for extended lengths of time, or after 64 3-hour cycles
between room temperature and 160'F.
2. Preconditioning specimens at 95% relative humidity (RH) for

16 days at room temperature followed by heating to 160°F

produced cracks in some specimens. Longer times or higher
temperatures for preconditioning increased the percentage of
samples which cracked.

3. Preconditioning to a moisture content of 1.6% (achieved

after 7 days at 16 0 'F and 95% RH) produced cracks in all
specimens when subsequently exposed to low RH.
Preconditioning to 1.1-1.3% moisture content produced cracks
in some specimens when exposed to low RH. Cracks were not
initiated in specimens with lower moisture contents.

4. Increasing the smoothness of acrylic sample edges by

polishing appeared to decrease the time for crack initiation.
However, insufficient tests were conducted to quantify any
changes in crack initiation time.

3.3 Crack Propagation Studies

A fracture mechanics model was developed to determine the

temperature at which crack propagation would begin due to
stresses caused by the mismatch in coefficients of thermal
expansion (CTE) between fiberglass and acrylic (Appendix, p. 16).
The model treated one-half of the thickness of the edge
attachment as a cantilever beam under uniform bending. The
temperature T at which a crack would propagate was found to be:

6
 (GbEI) 2 h
T = 2(atI- 2 )EI + To

where G is the fracture energy, b is the specimen width, EI is

the flexural rigidity of the beam, h is the beam thickness, To is
a reference temperature, and al-U 2 is the difference between the
CTE's of the acrylic and fiberglass. Using temperature dependent
values for a, and E found in [2] and a room temperature fracture
2
energy 0.188 in.lb/in from [3], the critical temperature for
fracture of the F-4 canopy acrylic was estimated to be 170F.
Using this method, the critical temperature for fracture of the
F-15 canopy acrylic was calculated to be 160'F.
Tests were conducted to experimentally determine conditions
for crack propagation. Results of these experiments were:

1. Milling 600 grooves in the edge of the samples with

fiberglass edge attachments and heating to 160'F did not
cause crack propagation.

2. Initiating a crack at the base of the groove and heating to

160'F did cause crack propagation.

3. Cracks propagated after 3 or more hours at temperatures

between 110'F and 120'F.

4. No cracks propagated in samples where nylon was used in

place of the fiberglass edge attachment.

7
 SECTION 4
CRACK PROPAGATION VERIFICATION

The ability of thermal stresses to propagate cracks in F-1;

canopies was verified using finite element analysis and
experiments.

A non-linear finite element (FE) model of the F-15 canopy

edge was developed using the COSMOS/M program. Plain-strain
conditions and uniform temperature throughout the material were
assumed. Temperature dependent moduli and CTE's were used.
Acrylic properties from [2] and fiberglass-phenolic properties
from [4] were used. Stress at the edge of the acrylic on the
centerline is shown in Figure 3. The stress is directed along
the edge. The values are much lower than the 3000 psi tensile
strength measured through the thickness of specimens cut from F-
15 windshields.

An FE model like that described above was developed which

included an 0.05 inch long crack extending in from the edge at
the centerline. The stress intensity factor KI at the crack tip
is shown in Figure 4. Ref. [5] gives a room temperature fracture
toughness (critical stress intensity factor) for stretched
acrylic of KI = 600 lb-in3/2 Based on this value the model
predicts crack propagation at 125 F. Thus, the models indicate
that thermal stresses are not sufficient to cause fracture but
can propagate existing cracks, in agreement with the results of
the F-4 investigation.

To further validate the results from the F-4 investigation,

five samples cut from an F-15 canopy (Swedlow Inc., F1608-86-G-
0627-RJOS) were heated to observe crack growth. Samples were
tested with both microscopic (less than 0.01 in) and macroscopic
(:0.20 in) cracks. Fiberglass edge attachments were removed from
some of the samples. Nominal sample dimensions were 3x5 in. The
samples were heated using an infra-red heat source on one side to

8
 L4
L3
L2
SU
V) Ii
Lo)
LU 0.9
a-.-

LU 0.7

O.4
0.3

02

0 .
0 0 K0 20 3040 5060 706090 K)0

TEMPERATURE CHANGE (T-TO)(deg F)

Figure 3. Stress in Thickness Direction at F-15 Edge.

OLT

0.
0.7
0.5

0.40.8
0---

i-- 11.4
Li~.

i--

m 0.1

-- 02
0 K) 20 30 40 50 80 70 60 90 100
TEMPERATURE CHANGE (T-TO) (deg F)
Figure 4. Stress Intensity at the Tip of 0.05 in. Crack in F-15.

9
approximate aerodynamic heating. The specimens were heated at
4°F/min as measured by a thermocouple mounted on the side closer

to the heat source. Five tests were performed. The specimen

configurations and test results are listed in Table 1.

Table 1 - Crack Propagation Tests on F-15 Edge Attachment

Edge Initial
Sample Attachment Crack Length Results

1 X micro No effect after 2 hr @ 180'F

2 micro Cracks after 12 min @ 180'F
3 0.20" Crack growth @ 168°F
4 0.22" Crack growth @ 163°F
5 X 0.25" No effect after 2 hr @ 180'F

(X indicates fiberglass edge attachment was removed)

Crack growth occurred only on samples which had fiberglass edge

attachments. Cracking initiated at the corners where the
resultant stress through acrylic would be the greatest. On
specimens 3 and 4 the fracture extended the existing crack at the
mid-surface. On sample 2, fracture occurred near the heated
surface due to the asymmetric heating. Since most of the
transparency crack-, wete observed to lie nearly in the middle of
the acrylic, it is likely that fracture occurs due to uniform
heating, such as occurs in storage or on the flight line.

10
 SECTION 5
SUMMARY AND CONCLUSIONS

Similarities in the geometry and materials of the F-4 and F-

15 edge attachments indicate that the results of edge crack
studies for the F-4 are applicable to the F-15. Simple
experiments conducted on F-15 canopy materials verified the
reported crack growth mechanism. Principle results of the
investigation are:

1. Stresses due to a mismatch in coefficient of thermal

expansion (CTE) between fiberglass/epoxy edge attachment and
stretched acrylic propagate cracks at the edge of the
acrylic. The stresses are too low to cause fracture in
uncracked specimens, but do propagate cracks.

2. Cracks will not propagate if an edge attachment material

which closely matches the CTE of acrylic is used. A
suitable material is a nylon-acrylic composite.

3. Visual inspection shows that many microcracks exist at the

edge of the acrylic. Absorption of moisture followed by
exposure to low humidity or prolonged exposure to high
temperature and humidity can initiate cracks. Other
processes may also occur which initiate the cracks.

4. Bearing strength was not significantly affected in F-4

samples cracked the entire width of of the edge attachment.
Similar results can be expected for F-15 samples which have
a similar geometry. Tensile tests showed that the presence
of cracks does not significantly reduce the bearing strength
of the edge. However, a reduction in bending strength may
occur.

11
5. None of the cracks on the test specimens or inspected
canopies were observed to remain open at room temperature,
as has been reported on some F-15 windshields. Operating
temperatures of F-15 windshields are in a range where stress
relaxation and thermal relaxation may cause permanent
deformation which results in the cracks remaining open at
room temperature. These effects have not been investigated.

12
 REFERENCES

1. Mulville, D. R., Wolock, I., and Thomas, R. J., "Crack

Formation in F-4 Aircraft Canopies", Proceedings of the
Conference on Transparent Aircraft Enclosures, Las Vegas, 1973
(AFML-TR-73-126), pp. 669-682.

2. MIL-HNBK-17A, Plastics for Aerospace Vehicles, Part II.

Transparent Glazing Materials, June, 1977.

3. Broutman, L. J. and McGarry, F. J., "Fracture Surface Work

Measurements on Glassy Polymers by a Cleavage Technique. II.
Effects of Crosslinking and Preorientation", J. Applied Polymer
Science Vol. 9, (1965) pp. 609-626.

4. Modern Plastics Encyclopedia, 1985-1986 ed., (New York:

McGraw-Hill, 1985).

5. Williams, K., "Anisotropic Fracture of a Hot-Stretched

Acrylic Thermoplastic-for Aircraft", Proceedings of the Fourth
International Conference on Fracture Vol. 3B, Waterloo, Ontario,
Canada, (A78-33601 13-23), pp. 1135-1143.

13
 APPENDIX
(Reference 1)

CRACK FORMATION IN F-4 AIRCRAFT CANOPIES

by

D. R. Mulville, I. Wolock and R. J. Thomas

Naval Research Laboratory

Washington, D. C. 20375

Abstract

The initiation of cracks in stretched acrylic plastic has been

reported for F-4 canopies under storage conditions. The cracks started
at the edge and propagated parallel to the plane of the sheet between
the fiberglass-acrylic edge attachments. However, laboratory tests in-
dicated that the presence of these cracks has negligible effect on the
bearing strength of the edge attachment.

A limited investigation was undertaken to determine the conditions

under which the crack formed in the canopy. There appears to be two
separate processes: crack initiation and crack propagation.

The former is the more complex phenomenon. It was shown that

cra.ks will initiate in stretched acrylic plastic when it contains over
1'.6% moisture and is then exposed to low humidity. The exact conditions
for crack initiation were not defined.

Cracks will propagate in stretched acrylic with a fiberglass

edge attachment if cracks are present and the assembly is heated to
120*F or higher. They do not propagate with a nylon edge attachment.
Using a fracture mechanics approach, in which the difference in thermal
expansion produces an opening mode force, it is possible to predict the
temperature at which an existing crack will propagate. This was cal-
culated to be approximately 170*F. which was substantially verified in
experimental tests. However, existing crack do not propagate at ele-
vated temperatures when a nylon reinforced edge attachment is used
because there is little difference in the coefficients of thermal ex-
pansion of the acrylic and the nylon-acrylic laminate.

Introduction

The formation of cracks was observed in the stretched acrylic of

F-4 aircraft canopies while in storage prior to service. The cracks
initiated at the end of the edge attachment and propagated parallel to
the plane of the sheet into the interior of the canopy as far as the end

14
of the edge attachment. These canopies are formed of 3/8-inch thick
stretched acrylic and have a fiberglass-acrylic edge attachment. An
investigation was undertaken of the origin of these cracks and of
their effect on the performance of the canopy.

The investigation is divided into three parts: bearing tests,

crack initiation studies and crack propagation studies. The effect of
in-plane cracks on the performance of the edge attachment was evaluated
by determining the bearing strength of specimens modeling an edge
attachment assembly. In studying the problem of the origin of the in-
plane cracks, attention was directed to the effects of variations in
temperature and in humidity, since those are the two parameters that
would vary during the storage of the canopies. This part of the study
was subsequently divided into two separate problems; crack initiation
and crack propagation, since it became apparent during the investigation
that these processes occurred under different conditions.
Materials

Two F-4 aircraft canopies that had formed edge cracks during
storage were furnished by Mr. R. E. Wittman of the Air Force Flight
Dynamics Laboratory at Wright-Patterson Air Force Base. The stock
number was 1560 788 6502. Uncracked specimens were obtained from
samples of the typical edge attachment, purchased from Swedlow Inc.
(Fig. 1). In addition, to provide a comparison, edge attachment mater-
ial of the same configuration was obtained using nylon fabric as the
reinforcement instead of fiberglass. En some cases, tests were also
conducted on control stretched acrylic and on stretched acrylic taken
from the canopy, both without an edge attachment.

Bearing Studies

A limited investigation was conducted of the effect of in-plane

cracks in the acrylic on the bearing strength of the edge attachment
at room temperature. Specimens were prepared using the geometry which
approximated the configuration used in the aircraft. 1/4-inch diameter
holes were drilled in specimens 1/2-inch from the end. The specimens
were 1-inch wide, since the bolt holes on the canopy were that distance
apart. The holes were drilled through the fiberglass edge attachment
and the acrylic, and a metal face plate applied to each surface and the
assembly bolted together using a typical service bolt tightened to the
aircraft manufacturer's specified torque of 40 in.-lb. Tensile loading
was applied to the acrylic plastic at one end of the specimen and to
the edge attachment assembly at the other end through the bolt.

Tests were conducted on the control edge attachment material as

well as on specimens taken from F-4 canopies and on specimens of stretched
acrylic with no edge attachment. Some of the specimens were tested
containing cracks parallel to the surface whereas others were uncracked.

15
Most of the specimens were tested with the bolts torqued to 40 inch-lbs.,
but several were tested with the bolts just finger tight.

The results are presented in Table I. They indicate the

following:

1. Edge attachment specimens with cracks fail at approximately the

same loads as those specimens without cracks. The presence of a
crack parallel to the surface in the stretched acrylic does not reduce
the bearing strength of the material.

2. Actually, stretched acrylic specimens with a fiberglass edge attach-

ment do not fail in bearing. They fail in tension at the tapered
junction of the edge attachment and the stretched acrylic due to the
stress concentration caused by the difference in stiffness.

3. The canopy specimens with edge attachment failed at loads approxi-

mately 1/3 less than those for control edge attachment material. This
reduction is probably caused by degradation of the resin matrix in the
edge attachment composite, increasing the stress concentration at the
tapered end of the edge attachment.

4. Stretched acrylic, tested in bearing, fails at the same or slightly

higher load than control material with a fiberglass edge attachment.
However, the acrylic does not decrease in bearing strength due to aging,
and as a result its strength is almost double that of the aged canopy
edge attachment material. The need for edge attachment reinforcement
in combination with stretched acrylic glazing must again be questioned.
5. The data are not definiLive with regard to the effect of bolt torque
on the load at failure.

Crack Initiation Studies

In studying the initiation of cracks in the edge attachment assembly

a series of tests were conducted to determine the effect of moisture and
temperature on the behavior of the acrylic plastic with no initial cracks.

The results are summarized in Table II. It was found that heating
the control fiberglass edge attachment assembly at 160*F for extended
period of time did not result in crack initiation, nor did cycling from
room temperature to 160'F for sixty-four cycles result in crack initiation.
The possible role of moisture was introduced by conditioning the speci-
mens at high humidity at various temperatures before heating to 1600F.
Conditioning at 75 0 F and 95% RH followed by heating to 160°F did not
result in crack formation. However preconditioning at 140OF and 95% RH
did result in crack initiation in a number of specimens. Preconditioning
at 160*F and 95% RH also resulted in some crack initiation, but when the
preconditioning period was extended to 7 days, all of the specimens

16
cracked except for the stretched acrylic with no edge attachment, taken
from the canopy. Making the exposed acrylic edge extremely smooth by
polishing decreased the time for crack initiation, whereas roughening
the surface with a coarse file increased the time required for a crack
to form. Finally, preconditioning the sperimen in water at 160*F for
7 days led to crack initiation in all of the specimens tested, including
the stretched acrylic with no edge attachment exposed to 75*F and low RH.

To determine how the preconditioning affected the moisture content

of the stretched acrylic, water absorption measurements were made on test
specimens for various conditions for 7 days. The results are presented
in Figure 2. It appears that if the moisture content of the stretched
acrylic is 1.6%, which is achieved in water at 160*F in 7 days, then
cracks will initiate in all specimens when exposed to low RH. From 1.4
to 1.6%, which is achieved at 160°F and 95% RH in 7 days, cracks will
initiate in specimens with edge attachment when heated to 160 F. If the
moisture content is approximately 1.1-1.3%, which is reached at 160°F and
95% RH in 4 days or 140°F and 95% RH in 7 days, then some specimens will
crack. At lower moisture contents, stretched acrylic specimens will not
form cracks when heated to 160 0 F.

The acrylic plastic edge of the as-received edge attachment assembly

was fairly rough and tool marks were visible. However the limited tests
conducted indicated that polishing would not retard the formation of cracks
and instead might accelerate their initiation.

In several of the tests, acrylic plastic from the aged canopy did
not crack whereas the remaining test specimens did. The reason for this
behavior is not apparent.

There are a number of questions regarding crack initiation that

were not answered by this limited investigation. Conditions under which
cracks may initiate have been determined but it was not demonstrated that
conditions of this type were encountered in the storage of the canopies
that cracked. However, due to the limiteQ scope of this investigation,
further studies to clarify these questions were not pursued.

Crack Propagation Studies

The difference in coefficient of thermal expansion between acrylic

plastic and fiberglass-acrylic laminate is well known and Is presented
in Figure 3 (Ref. 1). An analysis was conducted to determine under what
conditions the stresses induced in the stretched acrylic plastic due to
this thermal mismatch could produce crack propagation in the acrylic.

If the edge attachment segment is analyzed as a cantilever bean

under uniform bending (Figure 4), a failure criterion can be developed
to predict the temperature at which the initial crack will propagate.

17
 At the acrylic-fiberglass interface the strain is

(La AT) - (a1 -a 2 ) (T-To)

where Aa is the difference in coefficients of thermal expansion, (aI -a 2 ),

of the acrylic and the fiberglass, respectively, and AT is the differ-
ence in temperature (T-To). Using this value of strain, the corresponding
bending moment is,

M - 2(AaAT) El
h
where h is one-half the thickness of the acrylic, and El is the flexural
rigidity of the beam.

For a cantilever beam under uniform bending, the strain energy

release rate, ., can be expressed as

bEI

where b is the width of the acrylic specimen. Substituting the expression

for the thermally induced moment, M, yieldsa relation between, 4, and
temperature,
S. (2(l -a2) (T-To) Eli a
L h bEt
This expression can then be solved for the temperature as follows,
T (JbEI)h + To
2(aI -a 2 )EI

Since both a and E are Leaperature dependent, these quantities must be

represented as function of temperature in order to calculate the critical
temperature at which crack propagation occurs, T.

In addition the strain energy release rate for this mode of crack
propagation must be known.. Broutman and McGarry (Ref. 2) have reported
values of the surface work (or one-half the fracture energy) for crack
propagation parallel to the plane 01 a multiaxially stretched (55%)
acrylic sheet of 1.65 x 104 ergs/cm (0.094 in.-lb/in.-) at room tempera-
ture.

Using this value to compute the strain energy release rate, the
critical temperature is estimated to be approximately 170°F. It should
be noted that this is a two-dimensional analysis, and that the experi-
mental value of fracture energy was reported for room temperature and not
at the critical temperature. At corners of the canopy the bending moment
acts in two directions, and hence initiation and propagation of cracks
are more likely to occur here at lower temperatures.

18
 Tests were conducted to determine experimentally under what
conditions cracks would propagate in the stretched acrylic at the edge
attachment. The results are summarized in Table III. In the first set
of tests, it was found that when specimens with 60* grooves machined
in the acrylic at the end of the edge attachment were heated to 160"F,
cracks did not propagate from the tip of the groove. In the next set of
tests, an actual crack was initiated in each specimen by machining a
notch in the end and then driving a wedge into the notch until a uniform
crack formed along the length of the specimen. When specimens of this
type with a fiberglass-acrylic edge attachment were heated to 1600F,
the crack propagated. However, with a nylon-acrylic edge attachment,
the cracks did not propagate.

A third group of tests were conducted to determine the minimum

temperature at which cracks will propagate. This was found to be between
110" and 120*F for stretched acrylic with a fiberglass-acrylic attachment.

Finally, tests were conducted to determine if there is a minimum

length crack that will propagate when the assembly is heated. Cracks
approximately 1/16-, 1/8-, and 1/4-inch long were initiated in specimens
with a fiberglass edge attachment. Cracks propagated in all of the
s,.ecimens when heated to 160*F.

Thus when a crack is initiated in stretched acrylic with a fiber-

glass edge attachment, the crack will propagate if the assembly is heated
to 120*F or higher. Cracks do not propagate with a nylon edge attachment
even at 160*F.

Sumary and Conclusions

Studies were conducted to determine the effects of in-plane cracks

in the stretched acrylic on the performance of a fiberglass edge attach-
ment assembly and to determine the origin of cracks that form in storage
in canopies using this material.

Bearin! tests on the edge attachment assembly indicated that:

1. The presence of cracks parallel to the surface in the stretched

acrylic does not reduce the bearing strength of the material.

2. This assembly does not fail in bearing but instead in tension at the
tapered end of the fiberglass reinforcement due to the stress concentra-
tion caused by the change in stiffness.

3. The aged edge attachment material fails at loads approximately one-

third less than that for unaged material.

4. When the stretched acrylic alone is tested in bearing, it fails in

tension at the bolt hole at loads that are the same or slightly higher

19
than the control material with fiberglass reinforcement and approximately
double that of the aged edge attachment. This again raises the question
of the justification for fabric reinforcement when at. tched acrylic is
used in aircraft canopies.

Studies of the initiation of cracks in stretched acrylic indicate

that if the moisture content is 1.6% or above, cracks will initiate when
the material is exposed to low humidity. At lover moisture contents,
crack initiation is not consistent. The exact conditions at which
cracks will form after this preconditioning vem not determined.

Once a crack has initiated in stretched acrylic with a fiber-

glass edge attachment, it will propagate if the assembly is heated to
120F or higher. However, cracks will not propagate with a nylon edge
attachment even at 160*F. Using a fracture mechanics approach, in which
the difference in thermal expansion produces an opening mode force, it is
possible to predict the temperature at which an existing crack will propa-
gate. This analysis predicts the behavior of the fiberglass edge attach-
ment as well as the nylon edge attachment.

This limited investigation has not established that F-4 canopies

can attain a moisture content of approximately 1.6% in storage. However,
it is not unreasonable that this condition is attained and that sub-
sequent environmental conditions could lead to the formation and propa-
gation of cracks that have been observed in the canopies. The use of a
nylon edge attachment would eliminate the propagation of the cracks.
However, it has not been demonstrated that these in-plane cracks pose a
danger to the structural integrity of the canopy.

Acknowledgement

This investigation was sponsored by the Naval Air Systems Comnand

with Mr. Charles F. Bersch as project monitor. The authors wish to
acknowledge many fruitful discussions held with Mr. Bersch. Cracked F-4
canopies were kindly furnished by Mr. R. E. Wittman, Air Force Flight
Dynamics Laboratory, Wright-Patterson Air Force Base. The cooperation of
Swedlow, Inc. is also acknowledged, from whom special edge attachment
material was purchased.

References

1. Plastics for Flight Vehicles. Part 1I. Transparent Glazing Materials.

MIL-Handbook 17 (Aug. 1961).

2. Broutman, L. J. and McGarry, F. J. "Fracture Surface Work Measurements

on Glassy Polymers by a Clevage Technique. II. Effects of Crosslinking and
Preorientation". J. Applied Polymer Science 9, 609 (1965).

20
TABLE 1. BEARING TESTS OF F-4 EDGE ATTACHMENTS

EDGE b BOLT C FAILURE d FAILURE

CRACKEDa ATTACHMENTb TORQUE LOCATIONd LOAD
lb.
CONTROL MATERIAL X 40 EAB 1700
X F EAB 1750

X X 40 EAB 1990

X X F EAB 1840

40 HOLE 2150

F HOLE 1710

CANOPY MATERIAL X 40 EAB 1180

X 40 EAB 1245

X 40 EAB 1210

X X 40 EAB 1160

X X 40 EAB 1070

X X 40 EAB 1260

40 HOLE 2250

40 HOLE 2300

a. X indicates crack in stretched acrylic parallel to the plane of the sheet.

b. Specimens without X were stretched acrylic with no reinforcement.
c. 40 inch-pounds or finger tight.
d. EAB indicates tensile failure in the acrylic at the end of the edge attachment.
Hole indicates tensile failure initiating at the bolt hole.

21
 .uu * *cn1n
en 41 C41 6 nC
I~ -- = C

U U; U uH
0d
to 0 w l-
>~~ ~ U U
v:

0*

0 4 14 5

-S.~c -. 4 *
VS m-4 0 0 1VS 4

M~* u 'r.uUA

414i -, .4 = 41 -
o014n 0 0.-'0

04 0

o ugo
0 0 0
a4 o .
00

41 0 'U
0U 000U oU
A0 A 1. 41 41 N4 41 r
0 rh IU U) 1
co4 0 0 o0. MU uH
U U* C. 01- . >
o
H~~~~
Cfl-.~~0
~~~
4
4
-4l 1
14)044
.. ~V 0
- 0
% 0 -. H V 0 0 Vi

441

0 0 S. en

0- 0 00
4 00
c ,.)
0 00 0 0>

0ý 04 W0 0

S... 0 0 0
22
TABLE III: CRACK PROPAGATION STUDIES OF F-4 EDGE ATTACHMENT

CONTROL

FIBERGLASS E/Aa NYLON E/A

1. HEAT AT 1600F NO EFFECT

(600 GROOVES IN END)b
2. HEAT AT 1600F PROPAGATE- DO NOT PROPAGATE
(CRACKS INITIATED IN EIID)

3. HEAT AT'VARIOUS TEMP. PROPAGATES @ 130 0 F/l HR. NO PROPAGATION UP

(CRACKS INITIATED IN END) PROPAGATES @ 120'F/3 HR. TO 130°F
NO PROPAGATION @ 110OF

4. HEAT AT 160°F 1/16S.-/8-,1/4-INCH

(CRACKS INITIATED IN END) CRACKS PROPAGATE

a. E/A = edge attachment.

b. End refers to stretched acrylic at free edge of canopy, in between fiberglass

reinforcement. (See Fig. 1). Grooves or cracks were parallel to the plane
of the sheet.

c. Length of cracks initiated in end.

23
I')

I,.-

C) 41

LUU

0 0

244
 ~0

0 C4

0 CL d: a

wwwa'

) 0 0
boo boOLw

00 Q (DN C (Q
0 0 0
UN303d)NOI
d8OEIV 310

25'
0LL 6' ' "' 'I
NI
2i A-ACRYLIC

z•5- B-NYLON-ACRYLIC
- C- FIBERGLASS--ACRYLIC

U4

Xa 3 -A-
z 2

75
GB

zw I C
0

S-80 0 80 160
TEMPERATURE, OF

Figure 3 Effect of Temperature on Linear Coefficient of Thermal

Expansion for Various Materials (Ref. 1)

26
6c

-~40
41
0

.44
PW

27OUS
G ERMNTPRNIN OFCE 19, 7- 1'S05