AD-A251 767

WL-TR-91-3088

COUPON AND BIRDSTRIKE TESTING OF

F- II ADBIRT WINDSHIELDS WHICH HAVE
BEEN SUBJECTED TO SIMULATED PRESSURE/
THERMAL SERVICE LIFE

Daniel R. Bowman
University of Dayton
300 College Park Avenue
Dayton, Ohio 45469-0110

January 1992
JUN 19 19

Interim Report for Period February 1990 - October 1990

Approved for public release; distribution unlimited.

FLIGHT DYNAMICS DIRECTORATE 00

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

922 , l 4in1>
 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 publication.

T.J. CHOE, ILt, USAF RALPH J., SPEE

Project Engineer Chief, Aircrew Protection Branch
Windshield Sys Prog Office Vehicle Subsystems Division

FOR THE COMMANDER

Chief
Vehicle Subsystems Division

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.
 Unclassified
SECURITY CLASSIFICATION OF THIS PAGE
Form Approved
REPORT DOCUMENTATION PAGE OMBNo.0704-0188
la. REPORT SECURITY CLASSIFICATION lb. RESTRICTIVE MARKINGS
Unclassified
2a. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION /AVAILABILITY OF REPORT
Approved for public release; distribution
2b.' DECLASSIFICATION/DOWNGRADING SCHEDULE is unlimited

4. PERFORMING ORGANIZATION REPORT NUMBER(S) 5. MONITORING ORGANIZATION REPORT NUMBER(S)

UDR-TR-91-73 WL-TR-91-3088
'I

6a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION
University of Dayton (If
applicable) Flight Dynamics Directorate (WL/FIVR)
Research Institute Wright Laboratory
6c. ADDRESS (City, State, and ZIP Code) 7b. ADDRESS(City, State, and ZIP Code)
300 College Park Avenue Wright-Patterson AFB, OH 45433-6553
Dayton, Ohio 45469

8a. NAME OF FUNDING /SPONSOR'.':, 8b. OFFICE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
ORGANIZATION (If applicable)
I _F33615-84-C-3404
8c ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS
PROGRAM PROJECT TASK IWORK UNIT
ELEMENT NO. NO. NO ACCESSION NO.
64212F 1926 101 12
i1. TITLE (Include Security Classification)
COUPON AND BIRDSTRIKE TESTING OF F-111 ADBIRT WINDSHIELDS WHICH HAVE BEEN SUBJECTED TO
SIMULATED PRESSURE/THERMAL SERVICE LIFE
12. PERSONAL AUTHOR(S)
Bowman, Daniel R.
13a. TYPE OF REPORT 13b- TIME COVERED 14. DATE OF REPORT (Year, Month, Day) 15. PAGE COUNT
Interim FROM FEB90 TO OCT9O January 1992 60
16. SUPPLEMENTARY NOTATION

17. COSATI CODES 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number)
FIELD GROUP SUB-GROUP F-111 Aircraft Coupon Testing
ADBIRT Windshield Transparencies
Birdstrike Testing
19. ABSTRACT (Continue on reverse if necessary and identify by block number)
This program consisted of birdstrike testing, crack analysis, and coupon testing of
F-111 ADBIRT windshield transparencies which had been subjected to pressure/thermal
testing in the WPAFB Building 68 Transparency Durability Facility. Three pairs of F-111
ADBIRT windshield transparencies (left and right hand) were used in this program, one
pair each from Sierracin/Sylmar Corp., Swedlow, Inc., and PPG IndLstries, Inc. The
edges and bolt holes of all of the transparencies were examined for edge cracking.
The three rignt-hand windshields were birdstrike tested, and the three left hand
windshields were used for coupon testing. Dynamic mechanical analysis (DMA), gel
permeation chromatography (GPC), tensile, and edge attachment testing were conducted.
Simulated service life in the durability facility did not produce as much structural
degradation in terms of birdstrike resistance as in-service aging. A significant number of
cracks were found in the windshields in the vicinity of the edge attachments, similar to
from in-service aged windshields. Coupon testing revealed no bulk polycarbonate
crackinq inon.
deorada
20. DISTRIBUTION /AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION
[MUNCLASSIFIED/UNLIMITED 0 SAME AS RPT - DTIC USERS Unclassified
22a NAME OF RESPONSIBLE INDIVIDUAL 22b TELEPHONE (Include Area Code) 22c OFFICE SYMBOL
Mr. Ruise~i E. Urzi 1 (513) 225-6524 WL/FIVR
DD Form 1473, JUN 86 Previous editions are obsolete. - SECURITY CLASSIFICATION OF THIS PAGE
Unclassified
 PREFACE

The efforts reported herein were performed by the Aerospace Mechanics Division of
the University of Dayton Research Institute (UDRI), Dayton, Ohio, under Air Force Contract
F33615-84-C-3404, modification P00011. The program was sponsored by the Wright
Laboratory, Flight Dynamics Directorate, Wright-Patterson Air Force Base, Ohio. Air Force
administrative direction and technical support were provided by Mr. Russ Urzi, of WLIFIVR.
The work described herein was conducted during the period February 1990 to October
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.
Mr. Daniel R. Bowman was the principal investigator.

Acoesslon For

DTI, TA3 El
~Jii:
~~~ed El
.7!

IN -Y Sn_ G70A--&

'Dist
 TABLE OF CONTENTS

SECTION PAGE
I INTRODUCTION 1

1.1 Background 1
1.2 Objectives 2
1.3 Scope 2

2 TEST ARTICLE, COUPON TEST MATRIX, AND SPECIMEN PREPARATION 3

2.1 Test Article Procurement for Experimental Studies 3

2.2 Test Matrix 3
2.3 Specimen Layout and Fabrication 3

3 EXPERIMENTAL TESTS 7

3.1 Tensile Edge Attachment Tests 7

3.1.1 Test Objective 7
3.1.2 Specimen Configuration 7
3.1.3 Test Method 7
3.1.4 Test Data 7
3.1.5 Data Analysis/Correlation 7
3.2 Polycarbonate Tensile Test 11
3.2.1 Test Objective 11
3.2.2 Specimen Configuration 11
3.2.3 Test Method 11
3.2.4 Test Data 11
3.2.5 Data Analysis/Correlation 16
3.3 Dynamic Mechanical Analysis (DMA) Tests 16
3.3.1 Test Objective 16
3.3.2 Specimen Configuration 16
3.3.3 Test Method 16
3.3.4 Test Data 16
3.3.5 Data Analysis/Correlation 22
3.4 Gel Permeation Chromatography (GPC) Tests 23
3.4.1 Test Objective 23
3.4.2 Specimen Configuration 23
3.4.3 Test Method 23
3.4.4 Test Data 23
3.4.5 Data Analysis/Correlation 23

4 CRACK ANALYSIS 30

v
 TABLE OF CONTENTS (continued)
PAGE
SECTION
32
5 FULL SCALE BIRDSTRIKE TESTING
32
5.1 Test Setup and Test Facility
32
5.2 Birdstrike Test Results
46
6 CONCLUSIONS/RECOMMENDATIONS
49
REFERENCES

vi
 LIST OF ILLUSTRATIONS

FIGURE PAGE

2.1 F-Ill ADBIRT Transparency Nominal Cross-Section 4

3.1 Tensile Edge Attachment Specimen 8

3.2 Tensile Edge Attachment Test Setup 9

3.3 Tensile Test Specimen 12

3.4 Tensile Test Setup 13

3.5 Typical Tensile Load-Displacement Curve 15

3.6 DMA Scan of Acrylic Surface Plies 17

3.7 DMA Scan of PPG SN-H-04-20-2380 Polycarbonate Plies 18

3.8 DMA Scan of Swedlow SN 027 Polycarbonate Plies 19

3.9 DMA Scan of Sierracin SN 057 Polycarbonate Plies 20

3.10 Polycarbonate Molecular Wcght Distribution from PPG Windshield,

SN 87-H-04-20-2380 25

3.11 Polycarbonate Molecular Weight Distribution from Swedlow Windshield,

SN 027 26

3.12 Polycarbonate Molecular Weight Distribution from Sierracin Windshield,

SN 057 27

3.13 Polycarbonate Composite Molecular Weight Distribution from PPG,

Swedlow, and Sierracin Windshields 28

5.1 UDRI Impact Physics Test Range 5 33

5.2 Aft Arch/Center Beam Assembly 34

5.3 Comparison of Cross-Section Properties Between Flight Hardware

and UDRI Test Hardware 35

vii
 LIST OF ILLUSTRATIONS (continued)

FIGURE PAGE

5.4 Bird Impact Location 36

5.5 High Speed Camera Locations 37

5.6 Birdstrike Tested F- 11I ADBIRT Windshield and Test Summary,

Sierracin SN 082 38

5.7 Birdstrike Tested F-I II ADBIRT Windshield and Test Summary,

Swedlow 018 40

5.8 Birdstrike Tested F- 11I ADBIRT Windshield and Test Summary,

PPG SN 86-H-11-04-2010 42

5.9 Birdstrike Test Results of New, Service Aged, and Durability

Facility Aged F-Ill ADBIRT Windshields 45

viii
 LIST OF TABLES

TABLE PAGE

2.1 Test Transparencies 5

2.2 Coupon Test Matrix 6

3.1 Tensile Edge Attachment Test Results 10

3.2 Tensile Test Results 14

3.3 Glass Transition Temperatures from DMA 21

3.4 Gel Permeation Chromatography Results 24

ix
 SECTION 1
INTRODUCTION

1.1 BACKGROUND
The U.S. Air Force, recognizing the importance of maintaining bird impact resistance
protection for its pilots and recognizing that high performance aircraft transparencies are a
high cost item, is committed to continued monitoring, testing, and evaluation of aircraft
transparencies.

In 1984, under contract No. F33615-84-C-3403, project 1926, the Wright Laboratory
contracted with the UDRI to conduct a program to test service-aged F-I II transparencies.
The main objective of that program was to determine the effect of in-service* aging on bird
impact resistance capability. The program was conducted from May 1985 to December 1987
and included 22 full-scale birdstrike tests of baseline and in-service aged windshields.
Reference 1 gives a complete discussion of this subject.
The structural integrity of in-service aged F-1I I ADBIRT windshields was found to be
significantly reduced by in-service aging. Results of the bird impact tests indicated that the
windshield capability is reduced from a 470 knot baseline capability (as tested on simulated
flight hardware) to 360 knots after 2 years in-service aging (40% in terms of impact energy),
and reaches an asymptotic minimum value of 325 knots at an installed age of 5 years.
Birdstrike risk assessment of the windshield indicated that, given a birdstrike, degradation
causes the likelihood of penetration to increase significantly with installed age.
The reduction of bird impact resistance capability with installed age caused Air Force
concern. As a result, the Air Force contracted with UDRI (Contract F33615-84-C-3404,
modification P00011) to conduct additional research of the F-I II windshield structural
degradation problem. The program consisted of laboratory coupon tests of in-service aged
and baseline F-I 11 ADBIRT windshield coupons; research of polycarbonate degradation and
craze testing; fractography; and finite element analysis of the windshield edge attachment.

* In-service age is defined as the amount of time the transparency was on the aircraft, also
referred to as installed age.

I
Coupon testing indicated no polycarbonate degradation. Analysis of the edge attachment
revealed numerous fatigue cracks at the edges and in the vicinity of the bolt holes. We
believe that these fatigue cracks were the direct cause of the reduction in birdstrike resistance
of in-service aged windshields. Finite element analysis showed significant tensile stresses at
the edges for various pressure/thermal load cases. These stresses were high enough to
propagate existing cracks, and in several cases the stresses were high enough to initiate
cracks. Craze testing of the sealants, cleaners, and other chemicals used to install or clean
aircraft windshields indicated that many of the substances which are used in conjunction with
aircraft transparencies cause crazing of polycarbonate. This crazing, in conjunction with the
cyclic in-service pressure/thermal loads, was considered to be the most likely initiator of the
fatigue cracks in F- 111 ADBIRT windshields. A complete discussion of the results of that
program may be found in Reference 2.

1.2 OBJECTIVES
The objectives of this program are:
(1) To determine if F-Ill ADBIRT windshield transparencies subjected to simulated
service life testing in the WPAFB Full-Scale Durability Facility are experiencing structural
degradation similar to in-service aged windshields, and

(2) To gain additional insight into the cause of edge attachment cracking and
subsequent structural degradation of in-service aged F-Ill ADBIRT windshield
transparencies.

1.3 SCOPE
This program included an experimental coupon test phase, edge attachment crack
analysis, and full-scale birdstrike testing. Laboratory coupon tests were chosen to analyze
windshield degradation causes and effects. Sections 2 and 3 summarize laboratory coupon
tests of the F-Ill ADBIRT windshields. The edge attachment crack analysis is summarized
in Section 4. The birdstrike testing is summarized in Section 5. Conclusions and
recommendations are presented in Section 6.

2
 SECTION 2
TEST ARTICLE, COUPON TEST MATRIX, AND SPECIMEN PREPARATION

2.1 TEST ARTICLE PROCUREMENT FOR EXPERIMENTAL STUDIES

Six F-Ill ADBIRT windshields were used for this program. These windshields were
previously used for pressure/thermal testing in the WPAFB Full-Scale Durability Facility.
Figure 2.1 presents the nominal cross section for each transparency manufacturer. Table 2.1
is a list of the windshields used in this program and includes a brief summary of the test
history for each set of transparencies.

2.2 TEST MATRIX

This program was based on a coupon test matrix of 93 specimens, as shown in Table
2.2. The experimental coupon tests were chosen based on the results of the Reference 2
program.

2.3 SPECIMEN LAYOUT AND FABRICATION

The dynamic mechanical analysis (DMA), gel permeation chromatography (GPC), and
tensile coupon specimens were fabricated from material near the center of the windshields
away from the edges. The tensile edge attachment coupon specimens were fabricated from
material along the aft arch of the windshields. Only the left-hand windshields were used for
coupon testing- the right-hand windshields were used for birdstrike testing. All specimen
fabrication was accomplished in the UDRI machine shop. Specimens were first cut from the
full-size transparencies by jig-sawing and/or band-sawing. Selected edges of specimens, such
as edge attachment sides, were milled as necessary. Cutting temperature was controlled
during milling, when required, through the use of cooling air. Edges were machined dry in a
vertical mill using a four-flute 1-inch diameter cutter at 900 RPM and a table feed of 6-1/2
inches/minute. Care was takeih to minimize heat-up by removing less than 0.030-inch of
material per cut.

3
LO

;o
(0
0

100

000

C) 0l

000

-4

-4

44
 z
o u u u

00

I - ~ z -

IL Is' . .-
N 0 a\~ 0.N0 - EW

C/ , 00 00

u
z
C'4
< 0 0 0 00 0

Z -4 C400Q
cz
<- C)

0 0

00 -

CC)

u . 4C) ) C0)0C

.o 0-C

C4I 0 00 )C

5
 TABLE 2.2
COUPON TEST MATRIX

NUMBER OF COUPONS

TEST PPG SWEDLOW SIERRACIN

SN 87-H-84-20-2380 SN 027 SN 057

Tensile Edge 6 6 6
Attachment
Tensile 20 20 20

Dynamic Mechanical 3 3 3
Analysis (DMA)
Gel Permeation 2 2 2
Chromatography
(GPC)

TOTAL 93 COUPONS

6
 SECTION 3
EXPERIMENTAL TESTS

3.1 TENSILE EDGE ATTACHMENT TESTS

3.1.1 Test Objective
The objective of this test was to evaluate tensile edge attachment strength.
Edge attachment structural integrity is necessary for optimum bird impact resistance.
3.1.2 Specimen Configuration
The edge attachment beams were cut from the transparencies at the aft arch.
The specimen geometry is shown in Figure 3.1.
3.1.3 Test Method
The edge attachment specimens were tested with the fastener end mounted in
a fixture which simulated the attachments of the actual transparency design; the other end
included a 0.9" diameter hole which was pin-loaded. The test fixture is shown in Figure 3.2.
The test specimens were loaded in tension at a displacement rate of 500 in/min. using a
electrohydraulic closed-loop MTS test machine. For all tests, load versus displacement data
was stored in the digital memory of a transient recorder and played back at reduced speed on
an X-Y recorder.
3.1.4 Test Data
Table 3.1 is a summary of the edge-attachment test data.
3.1.5 Data Analysis/Correlation
Swedlow coupons had much higher edge strength values than PPG or
Sierracin. Those PPG and Sierracin specimens which exhibited brittle failure modes had
pre-existing fatigue cracks before the test (some of which were not detected in the crack
analysis of Section 4). Typically, those specimens with cracks had lower energy to failure
values. In a birdstrike test, this would translate to lower birdstrike capability. Overall, these
test results compare favorably with the results of previous testing conducted by UDRI
(Reference 2) of coupons cut from in-service aged F-111 ADBIRT windshields. Failure
energies for these tests ranged from 1519 to 2462 in.-lbs. Failure energies from the coupons
cut from in-service aged windshields (tested previously) ranged from 146 to 2795 in.-lbs.

7
 4J

0 4-)

4-)

E'-4
LOa

NN

0-

-'-
Figure 3.2 Tensile Edge Attachment Test Setup.

9
 TABLE 3.1
TENSILE EDGE ATTACHMENT TEST RESULTS

Manuf./ Serlaif/
Spec. Time in Peak Load Average Energy Average
ID Durability Facility (ib) (ib) (in-lb) (in-lb) Failure Mode

1 Sierracin 3300 1656 DF, BH, CFMR

2 SN 057 3375 1538 DF/BF, BH, CFMR
3 1225 flight hours 3475 3471 1600 1519 DF, BH, CFMR, AFGPR/CFGPR
10 3400 1558 DF, BH, AFMR
17 3650 1302 DF/BF, AFGPR, AFMR
18 3625 1462 DF, AFGPR, AFMR

7 Swedlow 5075 2426 DF, BH, CFMR

8 SN 027 5175 2923 DF, BH, DMR
9 847 flight hours 4900 5108 2694 2462 DF, BH, CFMR/DFMR, DMR
11 5350 2532 DF, BH, CFMR/AFMR, DMR
15 5150 2252 DF, BS/BH, DMR
16 5000 1948 DF, BS/BH, DMR

4 PPG 3525 1139 BF, BH, CFMR/AFMR

5 SN 87-H-04-20-2380 3700 2008 DF, BH, AFMR
6 156 flight hours 3525 3824 1757 1767 DF/BF, BH, CFMR/DFMR
12 5120 2154 DF/BF, BH, CFMR, DMR
13 3525 1832 DF, BH, AFMR
14 3550 1709 DF, BH, CFMR/DFMR

Legend:

DF - Ductile failure of polycarbonate plies

BF - Brittle failure of polycarbonate plies
BS Bolt shear
BH Bolt head pulled through glass phenolic retainer
AFGPR - Adhesive failure of bond line between glass phenolic retainer and polycarbonate
CFGPR - Cohesive failure of bond line between glass phenolic retainer and polycarbonate
AFMR - Adhesive failure of bond line between metal retainer and polycarbonate
CFMR - Cohesive failure of bond line between metal retainer and polycarbonate
DFMR - Ductile failure of metal retainer
DMR - Ductile deformation of metal retainer

10
3.2 POLYCARBONATE TENSILE TESTS

3.2.1 Test Obiective

The objective of this testing was to measure tensile properties of the bulk
polycarbonate (structural ply). The tensile test is one of the most common measures of
material performance; elongation and toughness are extremely important material
characteristics for applications involving impact (such as birdstrike).
3.2.2 Specimen Configuration
The tensile test specimen is shown in Figure 3.3. These specimens were
fabricated on a lathe. The 1/4-20 threads were machined with a radiused root to preclude the
possibility of failure initiating at the thread root. Ten specimens were fabricated from each of
the two polycarbonate structural plies from each windshield.
3.2.3 Test Method
The tensile testing was conducted using a MTS electrohydraulic closed-loop
test machine. The tensile test setup is shown in Figure 3.4. The tensile tests were conducted
using stroke control at a nominal actuator displacement rate of 5000 in/min. The
corresponding nominal strain rate was 200 in/in/sec. The actual engineering strain rate
achieved in the elastic portion of the tests was 131 in/in/sec. Actual engineering strain rate
after yield was 232 in/in/sec. Load and crosshead displacement were measured. High-speed
photography (5000 frame/sec) was used to obtain displacement data, providing more exact
displacement data than had been obtained previously. Consequently, tensile modulus, strain
rate, and elongation are also more exact than values generated from previous testing.
3.2.4 Test Data
Test data for the tensile tests are presented in Table 3.2. A typical
load-displacement curve is shown in Figure 3.5. No linear portion of the load-displacement
curve was observed during initial loading (at low strains) from which an initial elastic
modulus could be obtained. However, the load displacement curves were fairly linear from
test initiation up to 50% of the first peak load. Thus, the tensile modulus was calculated from
that portion of the load displacement curve. Toughness (area under the load displacement
curve) is presented in inch-pounds. Yield stress is the same as the tensile yield strength, and
ultimate stress is the same as tensile strength at break, as defined in ASTM D 638.

11
 D= 0 125 +ooo2DIA
+0.002
D +o.oo, DIA
0. 125±o~oiOR
UNC -7
1/4-20

0. 175 --
- - 0 . 400
"se --- REF.

- 0. 350 +0.005

I . 150 + °.°10

DIMENSIONS SHOWN ARE INCHES

Figure 3.3 Tensile Test Specimen.

12
Aw.

40K

Figure 3.4. Tensile Test Setup.

13
 TABLE 3.2
TENSILE TEST RESULTS

,auf,/ Serial#/ Tensile Yield Ultimate

Time in Diameter Modulus Average Stress Average Stress Average Eneryj Average Elgtjim Average
Dwai I it Foci Ii!tly P DIj (in) (psi) 0pi (pi (pi ps) (i) (n-Ib) (i-1b) () ()
PPG INNER 1 0.125 961685 13596 12427 60.5 121.9
SN87-N-04-20-2390 2 0.1235 978253 13617 12626 60.8 133.8
156 fligM hors 3 0.124 857426 13249 12266 59.5 131.9
4 0.123 1426965 13516 12729 62.1 137.6
5 0.124 1379696 1066155 13249 13478 12626 1243 65.4 59.5 145.8 133.0
6 0.123 1456714 13623 13D45 *** ****
7 0.1245 1431979 13297 11603 51.2 117.4
6 0.1235 1125929 13409 12676 64.1 142.0
9 0.124 631340 13663 12473 60.5 133.2
90 0.124 631340 13560 11903 51.3
NIOLE 1 0.1215 562597 14016 13692 56.4 127.1
2 0.123 557416 13255 1236 **** a..**
3 0.123 726357 13571 12867 62.6 136.5
4 0.1245 644264 13554 12576 57.2 122.6
5 0.124 469424 622263 13456 13451 11075 12455 64.3 59.6 70.6 120.7
6 0.124 704379 13249 11903 53.8 118.6
7 0.107 - -
8 0.124 --..
9 0.123 617369 13255 12361 57.6 126.3
10 0.1235 674306 13252 12635 65.9 140.9
Siedloo I1NER I 0.1255 616421 13743 12631 61.1 123.4
Si 027 2 0.125 536481 13904 13140 *** a....
847 fIi i, hoars 3 0.1245 730566 13656 12116 52.6 112.5
4 0.124 713354 13922 1273 61.6 130.8
5 0.124 844631 676494 13974 13784 1317 12212 65.1 54.6 137.1 116.9
6 O.1255 454044 13692 12277 55.3 119.2
7 0.124 135665 13870 10247 53.6 995.0
6 0.1245 1204935 13554 11141 36.0 81.5
9 0.125 1119422 13547 11714 46.0 101.1
10 0.1235 1992225 13963 12887 61.6 132.7

NIOLE I 0.123 686736 13510 12306 65.3 123.3

2 0.124 630442 13456 11903 52.9 115.0
3 0.124 511897 13456 12318 58.4 126.3
4 0.1235 626633 93565 12731 62.6 136.5
5 0.124 546024 615760 13456 13436 12635 12241 66.5 58.4 144.1 126.9
6 0.1245 764454 13400 11551 47.7 106.2
7 0.1235 550454 13409 12522 60.4 132.7
8 0.124 605243 13456 12214 56.6 125.1
9 0.1245 552599 13459 11911 55.1 122.6
90 0.1245 663119 13194 12116 58.6 155.4
Sierracin INER 1 0.124 563087 13560 12939 62.6 131.7
SN 057 2 0.123 715671 13791 13571 66.6 141.7
1225 flight hntrs 3 0.124 56913 93560 13094 67.0 141.7
4 0.1235 467481 13565 12104 48.0 105.6
5 0.1245 60572 S942 13451 13511 12635 1269O 63.2 62.4 134.7 133.9
6 0.1245 5166 13451 13246 66.7 141.0
7 0.123 534774 13676 12939 60.2 130.3
8 0.123 671297 13791 12361 64.9 137.2
9 0.124 610162 1350 12525 58.4 125.2
10 0.1235 61004 12783 12470 66.6 141.7
I.OLE 1 0.124 57M36 13249 13094 70.9 131.4
2 0.124 597214 13197 12990 69.7 150.4
3 0.123 652663 13360 12676 60.9 147.9
4 0.123 656154 13413 12939 65.3 141.5
5 0.1235 606397 593836 13461 13273 11478 12736 59.9 65.3 134.6 142.5
6 0.124 492701 13146 12525 62.5 139.0
7 0.123 51631 13150 13202 69.6 92.9
8 0.1235 595443 13252 12783 63.4 140.3
9 0.124 542009 13146 12635 67.4 147.2
to 0.123 624306 13360 12834 63.5 140.3

14
 fl idwl 2 ..
ITIM
I WItIlH iiii t I ifW111111 MIfillf! I
i Ifli illi Ili IT11"i Ili 4 1
9 NIN
Rif",R 19111 -41 ITO
IRT-1, lsli
: ill 911 !:,!:1
Ill if"It,;;:: -ll!, 1;I
AN 111 Iliij I . R11111"
IMHE
W I Rif
1114 1111141 4p M IN,lmlilfl;
it' O N!Ill
Ufft
I Rf U P 1111 I lilt
if fill
it!
H'I 111!i 111
Eli l::; : :
ti twp fill Ill 111TIOoff lIi Mfill'.
[1111104111141 VUH i W til. fill If 111M .- ii l. 11 '34 4011111 !W.
'1 lilt
IIIIIIIIIN 101t I t I -.. T
, , LTC ! :iit Hit It IT'Ift 11
flit

I I MOM IT li M;L
f;!! Ntit
lilt MOR11 it IOIL
14 Iff ill
it t
...... O fIfillull -11111 11
Will oil 10 1111
11
tit
TIMM I-ItM 1111
Iit:I i:lilt I lin illitli
T11ill till fill I!
illIIi 111 it Ill, Itli ltlt if! 11;B11T .
,fill;;:;, if11 WiNilIV
ANN
IM:
it liq l! !N1111111111111111111
I

Ill It 111RI1111 it, iil:::::

_111111M U M I
!it If
I., 111111
LTli
i1ii'll illif I -Ili fill tiilfii 44 1.

CNLn 11JU ...

TITII '141111ifil it ..i 111111111111 1 fil I I WtiR: I 111A
Ilif .1114,11:11ii
i it
I r
r... IT l;.!:
Ilion

I MEE
0.1 DISPLACEMENT, INCHES

Figure 3.5. Typical Tensile Load-Displacement Curve.

15
 3.2.5 Data Analysis/Correlation
The strain rates achieved in the high rate tensile testing are reasonably close
to strain rates measured during birdstrike testing. These tests indicate that the bulk
polycarbonate properties are not changing significantly. There are no obvious trends of
increased yield strength with increased service aging in the durability facility (which is
typically a characteristic of thermal aging); in addition, there does not appear to be any
significant reduction in elongation or toughness. The initial elastic tensile modulus varies
from 600,000 to 1,000,000 psi. This variation is due to the fact that the initial elastic tensile
modulus is difficult to measure accurately, because of the rate of testing and fixture slack
which is taken up at the initiation of the test.

3.3 DYNAMIC MECHANICAL ANALYSIS (DMA) TESTS

3.3.1 Test Objective
The objective of this testing was to determine if service aging in the
durability facility was affecting the DMA curves. Changes in polymer structure are reflected
by changes in the DMA curves. Dynamic mechanical analysis has been used by many
researchers to evaluate the effects of aging, solvent attack, and UV degradation of polymers.
3.3.2 Specimen Configuration
Specimens were cut from the bulk windshield material away from the edges.
The specimens were rectangular beams, 2.5 inch x 0.5 inch x 0.1 inch nominal. One
specimen was fabricated from each of the two polycarbonate plies and from the acrylic ply
from each windshield.
3.3.3 Test Method
Each ply was analyzed in flexure from -125' to 175°C using a 20 C/min
heating rate at a fixed frequency of 1.0 Hz using a Dupont Dynamic Mechanical Analyzer.
3.3.4 Test Data
DMA scans of the transparency specimens are shown in Figures 3.6-3.9.
Table 3.3 summarizes the glass transition temperatures (Tg) for each specimen which were
determined from the maximum of the loss modulus curve, E".

16
 C(Od )..*3360-1 ----

aIn 0 In a In >

a I OXI

0
-0
N

CLL

0
r- r- 0

0 ('4 ~u

zooU
c-4 E

>17
 S i fCod ).33 Ba-I
r, -j
a Q
in >,
I 0

in N
c'5
d

'a a
aa
* ---

4)

U--4
~ 0 0

a4
0
c'Jo

Un 0

CL i

- E a,
00
N 44'

I I

ITI

Inn

18
 a Zn r cad )3 B 1 r--- 2 J

E In

Inn

0. 0
0W

0. 4-..4-a

\ L
CL

.001,. Il Q

C))

0 0
1/"
3t
Vl

19
 o r~cod )113j160 - --
00

9 n
Ann

E I
4c'
-u
0.0

-4

'II 0 0
1 0 44

0
n 1/1 L

IV-
0 0 .0

-4 "-4-0

'-4 -H-
U) 7,I

-. 4

20
 TABLE 3.3

GLASS TRANSITION TEMPERATURES FROM DMA

MANUFACTURER/ PLY Tg ( 0 C)
SERIAL NUMBER
PPG Acrylic 119
SN 87-H-04-20-2380
Middle Polycarbonate 150
Inner Polycarbonate 155

Swedlow Acrylic 118

SN 027
Middle Polycarbonate 147
Inner Polycarbonate 156.5

Sierracin Acrylic 126

SN 057
Middle Polycarbonate 156
Inner Polycarbonate 160

21
 3.3.5 Data Analysis/Correlation
Nomenclature used in this report to describe transitions and relaxations in
Dynamic Mechanical Analysis of polycarbonate are as follows: a=Tg (glass transition
temperature) = 150'C, 61 - 25*C, 82 - 80 0 C, and y- -100 0 C, as used in Reference 2. The a,
l3, 62, and y general regions are shown in Figure 3.7.
There is evidence of physical aging/thermal history (annealing effects) in the
-35 to 135 temperature range; however, these changes do not appear to be significantly
affecting the polycarbonate mechanical properties as evidenced by the tensile testing. There
are no major changes in the shape or location of the loss peak which is associated with the
glass transition temperature (the so called alpha transition) and this is evidence that there is
no significant chain scission or cross-linking taking place, and therefore no changes in
molecular weight would be expected. Large changes in molecular weight would be evidenced
by significant locational changes of the loss peak. For all manufacturers, the alpha transition
occurs at a temperature of 5C to 100 C lower for the middle (outboard) polycarbonate ply
than for the inner polycarbonate ply. Except for the PPG specimens, the gamma peak is also
fairly stable in shape and location. The beta two peak is recognizable for the Swedlow and
Sierracin specimens; however, it is very subtle and difficult to recognize for the PPG
specimens. In all cases, the height of the beta two peak is smaller for the middle (outboard)
polycarbonate ply (the height of the beta two peak has been reported to decrease as a
specimen experiences thermal aging). This is consistent with the actual thermal environment
experienced by the windshield. The middle (outboard) polycarbonate ply experiences more
time at higher temperatures in the Durability Facility. The cockpit interior is maintained at a
nominal temperature of 75'F and the inner (inboard) polycarbonate ply stays relatively cool.
In actual field conditions, however, the inboard polycarbonate ply does experience significant
elevated temperatures on the flight line. Cockpit temperatures have been reported to be as
high as 200'F in the summertime. The Durability Facility does not currently simulate this
flightline condition. There was no significant difference between inboard and outboard
polycarbonate ply dynamic mechanical analysis plots from the testing of baseline and
in-service aged F-Ill windshields as reported in Reference 2. This would indicate that both
plies are experiencing similar thermal aging in the field. The thermal aging detected here and
in the Reference 2 program is not significantly affecting polycarbonate mechanical properties.

22
With time thermal aging would be expected to cause degradation of the mechanical
properties. The results of the Reference 2 program indicated no significant bulk
polycarbonate degradation with up to 5 years of in-service aging.

3.4 GEL PERMEATION CHROMATOGRAPHY (GPC) TESTS

3.4.1 Test Obiective
The objective of the GPC testing was to determine if service aging in the
durability facility produced a molecular weight reduction of the polycarbonate.
3.4.2 Specimen Configuration
Shavings of polycarbonate were removed from the center of the individual
plies. They were dissolved in tetrahydrofuran (THF). Many of the samples were difficult or
impossible to completely dissolve even after a week in the solvent and several hours in an
ultrasonic bath. The undissolved residue was assumed to be additives.
3.4.3 Test Method
GPC measurements were conducted utilizing an HP 1090A liquid
chromatagraph GPC system. Polystyrene standards were used and the column used was a
Phenomenex Ultracarb 5 micron particle size linear column, 30 cm long, mixed porosity for
covering the greatest range of molecular weights, approximately 400 to 1,000,000.
Polycarbonate molecular weight calibration tables were used to obtain molecular weights.
3.4.4 Test Data
A summary of the test data is presented in Table 3.4 and Figures 3.10-3.13
are plots of molecular weight distribution.
3.4.5 Data Analysis/Correlation
The PPG specimens had higher molecular weights than those from Swedlow
or Sierracin. This in itself is not any indication of degradation, because the companies do not
all use the same supplier of polycarbonate. The middle (outboard) polycarbonate ply for
Sierracin and Swedlow had a slightly lower molecular weight than the inner (inboard)
polycarbonate ply. This may be an indication of thermally induced molecular weight
degradation. The middle (outboard) polycarbonate ply experiences greater temperatures than
the inboard ply. However, the results of the tensile testing indicate that this reduction is not
causing a structural problem.

23
 TABLE 3.4

GEL PERMEATION CHROMATOGRAPHY RESULTS

MANUFACTURER/ PLY Mn Mw Mz
SERIAL NUMBER
PPG Middle Polycarbonate 10890 31100 57860
SN 87-H-04-20-2380
Inner Polycarbonate 10980 30180 55250
SWEDLOW Middle Polycarbonate 7744 23190 46730
SN 027
Inner Polycarbonate 7160 23670 48280
SIERRACIN Middle Polycarbonate 6810 21330 44300
SN 057
Inner Polycarbonate 7254 22870 47460

NOTES:
MN = number average molecular weight
Mw = weight average molecular weight
Mz = Z average molecular weight

24
 LO
oL

0
4-4

LI
-J

0 -4

QI)

La.]

0) Q)
0 r'.

LUJ

0 C

0 0

00
0

25
 CN
C0

O 0
-'-4

- -4

E/)

0
4-4

0
F-4

-4
LUI m
00

C) 0)

0LJ0 0

00 0-

ol 0

>-1

0
0 ) 0 to) 0 LO ))L 0 C' Lo
4 C 6 03

26
 cn

LO -4

44

.-I

-4

(.D 4-3

L. 4J

H4-

-4

LLii
00

LIJ 0
4-)

00 0

P4W

Ln V) -4
0

o00
0 LO 0 LO 0 1O 0 1O 010

iIO Id IHO11M
27
 0
'-4

0
4

+
t U

41)

cF-

-4J

LLJ r.

z aD cUJ4-I

LLJ C4
LLJ a ::
wq 04
ZOOOOO 0
- < i i _3 E

0C1

':' 0

0) '0 0 10 0 10 0 10) 010O

28
 The number average (M,) and weight average (Mw) molecular weights for
Sierracin and Swedlow from this program were lower than the corresponding molecular
weights for Sierracin and PPG specimens from the Reference 2 program. This may indicate
that the thermal profiles used at the durability facility are more severe than those experienced
in service. Additional testing would be required to confirm this.
Overall, there is no evidence of significant molecular weight reduction for the
specimens tested, which included the bulk polycarbonate. It is possible that there is
molecular weight reduction in the top 5-10 microns of the surfaces at the transparency edges
and/or at the bolt holes which is not detected by the testing.

29
 SECTION 4
CRACK ANALYSIS

Prior to birdsirike testing and coupon testing, the sealant at the windshield edges was
removed to determine if cracks existed at the edges and in the vicinity of the bolt holes.
Similar studies conducted by UDRI in the Reference 2 program revealed numerous fatigue
cracks in windshields which had been removed from service. The results of this crack study
are shown below.

MANUFACTURER/ WINDSHIELD SIMULATED POLY. CRACKS

SERIAL NO. SERVICE LIFE INBOARD/OUTBOARD
PPG Left 6.25 mo. no visible cracks
SN 87-H-04-20-2380
PPG Right 6.25 mo. 74/68
SN 86-H-11-04-2010
Swedlow Left 2 y. 9.9 mo. 3/2
SN 027
Swedlow Right 2 y. 9.9 mo. none/4
SN 018
Sierracin Left 4 y. I mo. none/51
SN 057
Sierracin Right 4 y. 1 mo. none/6
SN 082

Analysis of the crack documentation was completed to determine patterns in crack

location (sill, center beam, aft arch, forward arch, inboard polycarbonate ply, outboard
polycarbonate ply), direction of travel, and length. Only one of the transparencies (PPG SN
87-H-04-20-2380) was crack free. However, observation of the failure surfaces from the
tensile edge attachment testing, reported in Section 3.1, revealed pre-existing fatigue cracks in
PPG SN 87-H-04-20-2380 at two bolt holes (six specimens were tested with two bolt holes
per specimen). These cracks were not detected by eye during the crack study reported in this
section. It is likely that a number of fatigue cracks exist in each of the windshields which
were not detectable by eye. In contrast, the companion right hand windshield (PPG SN
86-H- 11-04-2010) had 142 cracks. If it can be assumed that both windshields experienced the

30
same exact test history (and there is no reason to believe that they did not), it is possible that
an inconsistency in manufacturing caused these differences in cracking. The same comment
can be made for the two Sierracin windshields; however, the disparity between crack numbers
is not as great, and the Sierracin windshields had much greater simulated service life.
We did not see any correlation between simulated flight hours and total number of
cracks. The crack analysis conducted in the Reference 2 program indicated a correlation
between total number of cracks and in-service age with a general trend of increasing crack
length and number with increasing service life. A possible explanation for this disparity
between cracks produced by simulated life and actual service life is that the temperature
cycling that occurs on the flight line (hot summer days and cold winter nights) may produce
relatively uniform crack growth, while the in-flight pressure/thermal loading conditions, which
are of shorter duration and are more random, may produce random growth. Also, crack
initiation is a somewhat random event and this effort represents only a limited number of data
points. Another difference between simulated service life and in-service aged windshield
cracking is crack location by ply. In the Reference 2 program, the majority (on the order of
90%) of the cracks occurred in the inboard ply, while for the windshields with simulated
service life, the outboard polycarbonate ply tended to have more cracks. This may also be
attributed to thermal conditions not simulated at the Durability Facility, or to saturation of the
edges of the windshields during windshield and overall aircraft cleaning in the field. The
inboard ply would be exposed to solutions that collect at the edges between the windshield
and the frame more so than the outboard ply, and would be more susceptible to chemical
crazing followed by crack growth for windshields in the field. The Durability Facility installs
the windshields according to the F-Ill T.O., which requires cleaning of the frame with
isopropyl alcohol prior to sealing, application of a primer (EC1945B) to the frame, and then
sealing with Pro-Seal 899B-2. Both the primer and the sealant have been found to cause
crazing of polycarbonate. Very few cracks initiated at the transparency edges for the
windshields from the Durability Facility; most cracks initiated at the bolt holes (except for
Sierracin SN 057 which had extensive crazing/cracking along the edges). The windshields
from the Reference 2 program had roughly equivalent numbers of cracks at the bolt holes and
the edges for Sierracin windshields, while there were iany more cracks at the edges for PPG
windshields.

31
 SECTION 5
FULL-SCALE BIRDSTRIKE TESTING

5.1 TEST SETUP AND TEST FACILITY

The birdstrike testing was conducted at the University of Dayton Impact Physics Range
5, see Figure 5.1. An F- I11E crew escape module, Air Force serial number 68-024, was used
as a test stand. The center beam and the windshield aft arch were fabricated using high
strength steel as designed by UDRI to simulate the dynamic structural behavior of the actual
system. The UDRI simulated arch/center beam assembly is shown in Figure 5.2, and a
comparison of cross-section properties for the UDRI test hardware and flight hardware is
shown in Figure 5.3. A more complete description of the development of this test hardware
is documented in Reference 1. Three right hand aft arches were manufactured for this effort
to allow for expected structural damage. A new arch was used for each of the three birdstrike
tests.
Artificial 4-pound gelatin birds were impacted at the most critical location of the
windshield, the upper inboard corner (see Figure 5.4). Right hand windshields were tested
with the left hand windshield, canopy framework, and canopy transparencies installed. Three
high-speed movie cameras (5000 frames/sec) were utilized to record each bird impact event.
Camera locations are shown in Figure 5.5.

5.2 BIRDSTRIKE TEST RESULTS

The technical data from each of the birdstrike tests are summarized in the following
standard data forms, and still photographs of each tested windshield are shown in Figures
5.6-5.8.
The Sierracin windshield had 122 mulated flight hours in the WRDC Building 68
ful!-scale durabili y facility, which is equivalent to approximately 4 years and I month of
in-service aging. It was birdstrike tested at 373 knots and passed. The Swedlow windshield
had 847 simulated flight hours, which is equivalent to approximately 2 years 10 months of
in-service aging. It was birdstrike tested at 405 knots and passed. The PPG windshield had
156 simulated flight hours, which is equivalent to approximately 6 months of in-service aging.
It was birdstrike tested at 475 knots and failed catastrophically. These test points are plotted

32
 V))

4-))

P-4
41

C. V)

(A (A

0-

33
 z

<I-

0'z

Uj

UU

U)

U)

a)
-P

a)
4-,

44i

341
 N

FA- 1A %0 (A * (A .

' qw % 0 ON
-~
of it - - 'ic-

9. 3- a c4 4

CJ ro
ii ii
'.0~
B
-
c
**E ~
- -z0W
* ~~ I . -0I

C;4

O.
c -

cc 00c
39U 0n
.3
cr. - ' U

AS 2 n-i 0 -

OC a)

- 14
q 0
~ - C
.- ~ N
t.~ d ~.44

L 5EE

ar -c a 4

CF . x 10 IM 0
-E CD

F4

35aK
Impact
Point 12

Right-Hand Windshield

TOP VIEW
All dimensions in inches as
measured from the edge of the
transparency along the trans.
parency surface.

Figure 5.4. Bird Impact Location.

36
 Bird Flight Path

Top View

Bird Flight Path

Side View

22

Bird Flight Path

Figure 5.5. High-Speed Camera Locations.

37
 C'4
00

C)

Ci)

-4

rz4
ai)
4J)

ci)

4J
Ua
'-I

44

38
 UDRI
F-111 RIGHT-HAND TRANSPARENCY
BIRD IMPACT TEST

TEST SUMMARY

I. BASIC TEST DATA

Date of Test: 9/11/90 Test No.: 5-0620

Impact Pt.: 8" down from aft edge, 12" over from center beam (standard) impact
point
Planned Impact Vel.: 365 kts (617 fps) Actual Impact Vel.: 373 kts (631 fps)
Bird Wt.: 4.055 Ambient Temperature: 71F

II. TEST HARDWARE

Crew Module Ident. AFSN 68-024 MFGR. SN 227

R/H Windshield:
Manufacturer: Sierracin
Serial Number: SN 082
Date of Manufacture: 8-84
Test History: 1225 simulated flight hours
Weight: 49.3#
L/H Windshield: PPG 015-057 DOM 9-23-80
R/H Canopy: PPG 504973 FSPP DOM 3-26-75
L/H Canopy: Sierracin SN 013 DOM 9-77

Aft Arch Configuration: UDRI simulated aft arch

Fasteners: Screws Nuts Washers Torque

Aft Arch: NAS1203-17 MS 21043-3 #10 SAE 40 in-lb
Center Beam: NAS1204-15 1/4# Grade B 1/4" SAE 25 in-lb
Sill: NAS1204-15 1/4# Grade B 1/4" SAE 25 in-lb
Forward Arch: NAS1203-17 10-32 Machine #10 SAE 25 in-lb

IIl. PRE-TEST OBSERVATIONS

Some surface scratches, small delamination at forward corner by center beam.
Distortion of acrylic at forward corner by the sill.

IV. POST-TEST OBSERVATIONS

Impact point dead on, surface acrylic cracking, evidence of ductility behind the
the impact point. Small permanent pocket, very small permanent radial
deformation of aft arch and forward flange rolled over slightly. Bolts behind
impact point deformed.

V. SIGNIFICANCE OF TEST
Pass at 373 knots. Transparency is not degraded as bad as in-service aging
would indicate. Transparency looks capable of withstanding higher velocity.

39
 -4

'-A

U)

rtb

ro

-4

-4

-)
rn.
0)

.4-
En

40
 UDRI
F-111 RIGHT-HAND TRANSPARENCY
BIRD IMPACT TEST

TEST SUMMARY

I. BASIC TEST DATA

Date of Test: 9/13/90 Test No.: 5-0621

Impact Pt.: standard

Planned Impact Vel.: 400 kts (676 fps) Actual Impact Vel.:404.6 kts (683 fps)
Bird Wt.: 4.054 Ambient Temperature: 71°F

II. TEST HARDWARE

Crew Module Ident.

R/H Windshield:
Manufacturer: Swedlow (86175)
Serial Number: 018
Date of Manufacture: 8-86
Test History: 847 simulated flight hours
Weight: 48.5#
L/H Windshield: PPG 015-057 DOM 9-23-80
R/H Canopy: PPG 504973 FSPP DOM 3-26-75
L/H Canopy: Sierracin SN 013 DOM 9-77

Aft Arch Configuration: UDRI simulated aft arch

Fasteners: Screws Nuts Washers Torque

Aft Arch: NAS1203-17 MS 21043-3 #10 SAE 40 in-lb
Center Beam: NAS1204-15 1/4# Grade B 1/4" SAE 25 in-lb
Sill: NAS1204-15 1/4# Grade B 1/4" SAE 25 in-lb
Forward Arch: NAS1203-17 10-32 Machine #10 SAE 25 in-lb

III. PRE-TEST OBSERVATIONS

Some scratches, outer retainer loose in places, sealant between acrylic and
retainer separated in places. Some delamination, acrylic distorted (wavy along
edges).

IV. POST-TEST OBSERVATIONS

Target dead on, surface acrylic and sublayer cracking. Permanent pocket formed
behind impact point. Bolts bent behind impact point on aft arch, aft arch
deformed radially, and windshield support flange rolled over at impact point.

V. SIGNIFICANCE OF TEST
Pass at 405 knots. Transparency appears to be capable of withstanding higher
velocity.

41
 C>J

04

-4
I-4

rI'

44

41

-4

'4
-

4J

42
 UDRI
F-111 RIGHT-HAND TRANSPARENCY
BIRD IMPACT TEST

TEST SUMMARY

I. BASIC TEST DATA

Date of Test: 9/14/90 Test No.: 5-0622

Impact Pt.: standard

Planned Impact Vel.: 460 kts (777.4 fps) Actual Impact Vel.: 475 kts (803 fps)
Bird Wt.: 4.046 Ambient Temperature: 71°F

II. TEST HARDWARE

Crew Module Ident.

R/H Windshield:
Manufacturer: PPG
Serial Number: LBSN 86-H-11-04-2010
Date of Manufacture: 11/86
Test History: 156 simulated flight hours
Weight: 48.3#
L/H Windshield: PPG 015-057 DOM 9-23-80
R/H Canopy: PPG 504973 FSPP DOM 3-26-75
L/H Canopy: Sierracin SN 013 DOM 9-77

Aft Arch Configuration: UDRI simulated aft arch

Fasteners: Screws Nuts Washers Torque

Aft Arch: NAS1203-17 MS 21043-3 #10 SAE 40 in-lb
Center Beam: NAS1204-15 1/4# Grade B 1/4" SAE 25 in-lb
Sill: NAS1204-15 1/4# Grade B 1/4" SAE 25 in-lb
Forward Arch: NAS1203-17 10-32 Machine #10 SAE 25 in-lb

III. PRE-TEST OBSERVATIONS

Delamination around edges, also acrylic bubbled at edges. Large bubble at aft
arch. Large bubble at sill. Interior coating somewhat milky and dirty.

IV. POST-TEST OBSERVATIONS

Extensive poly and acrylic cracking; large 15x15" piece of laminate broken out
at impact point. Estensive delamination, 2 bolts sheared off. No apparent
damage to arch.
Catastrophic failure.

V. SIGNIFICANCE OF TEST
Failure at 475 knots.

43
in Figure 5.9 on top of the test data from the birdstrike testing program conducted by UDRI
of in-service aged F-Ill windshields, Reference 1.
The Sierracin and Swedlow windshields passed birdstrike tests when shot at 50 knots
above the capability curves which resulted from birdstrike testing of the in-service aged
windshields. The PPG windshield failed at 45 knots above the capability curve (note there
were a significant number of fatigue cracks in this windshield), however 475 knots is right at
the limit of the capability for new windshields using the UDRI hardware. Tht full-scale
durability facility currently simulates pressure/thermal profiles from the flight environment
and cleaning operations, but does not include UV light, moisture, flightline cockpit heating, or
ambient thermal effects (although several of these are being considered for incorporation into
the facility). The differences in birdstrike degradation between simulated and actual service
life are most likely a result of those environmental factors which are not simulated in the
durability facility.

44
 a-(0

0 Nz

00 Cie

wz u

LU 4

04 94 U)-

4J) t-

M- .4)

Z Z (AM$ -4
I ICt4 co
LONJ 0 4)

z 0)

o o 400
o E 0
zsoj a6 tLIUOI3A
454
 SECTION 6
CONCLUSIONS/RECOMMENDATIONS

The objectives of this program were successfully accomplished, i.e., to determine if

F- 11 ADBIRT windshield transparencies subjected to simulated service life testing in the
WPAFB Building 68 Durability Facility are experiencing structural degradation similar to
in-service aged windshields, and to gain additional insight into the cause of edge-attachment
cracking and subsequent structural degradation of in-service aged F-111 ADBIRT
windshields.
Test results indicate that simulated service life in the WPAFB Building 68 Durability
Facility is producing structural degradation, but this degradation is less severe in terms of
birdstrike resistance than the structural degradation caused by actual in-service aging.
As concluded in the Reference 2 program, UV light is not a significant factor in the
F- 111 windshield structural degradation problem. The durability facility does not include UV
light, and none of the resultant characteristics of UV degradation are present (e.g., molecular
weight reduction, and embrittlement of the polycarbonate). There is evidence of thermal
history/annealing; however, tensile testing shows that the bulk polycarbonate still retains an
acceptable level of toughness. No evidence of chemical attack of the bulk polycarbonate was
detected by gel permeation chromatography (GPC) or by dynamic mechanical analysis
(DMA). Tensile edge attachment testing did not show any significant reduction in edge
strength.
Craze research and testing from Reference 2 of sealants, cleaners, and other chemicals
used for maintenance and installation of the F- 111 transparency system indicated that
chemical crazing is a likely initiator of fatigue cracks at the transparency edges. The
pressure/thermal fatigue loadings on the F- 111 windshield cause and/or extend cracks caused
by other mechanisms. The fatigue crack tabulation and analysis brought to light the
significance and magnitude of the crack problem. Although the cracking resulting from
simulated service life was random, virtually none of the windshields studied were crack-free.
Even though the entire in-service environment for the F-Ill is not currently being simulated
at the Durability Facility, cracks are still occurring. This cracking does not appear to be
following the same trend as cracking produced by in-service aging. However, the cracking

46
which occurs as a result of simulated service life in the Durability Facility does degrade
birdstrike resistance.
Two additional windshields manufactured by PPG Industries that were not available to
be tested in this program prior to the completion of the technical work have been tested in the
Durability Facility. These were designed by PPG to overcome durability problems (associated
with PPG windshields which were tested in this effort) and were installed with a non-
aggressive dry seal. We recommend that these windshields be examined for edge cracking.
This would allow determination of how much of an effect the wet seal and associated primer
have on cracking. We suspect that cracking is present in those windshields as the
pressure/thermal environment associated with the F-111 and the current edge design are
sufficient alone to cause edge cracking. Other factors such as sealants, chemicals, moisture,
etc., may decrease crack initiation time and accelerate crack growth speed.
Conclusions and recommendations of the Reference 2 program which are considered
applicable to this effort are repeated as follows.
Craze testing of optical and machined surfaces has indicated that machined surfaces are
more susceptible to craze than optical (polished, smooth) surfaces. A possible short-term
solution, which would reduce the influence of moisture and chemical attack, and fatigue,
would be to machine and polish the edges and the bolt holes to a very smooth finish
(removing all sharp edges), and to use moisture and chemical resistant coatings on the interior
and exterior windshield surfaces as well as interior surfaces of the bolt holes. Coatings could
be used to effectively seal the entire transparency, thus greatly reducing or eliminating the
effects of hydrolysis and/or chemical attack from cleaning solutions, etc. Also, the possibility
of field contamination of those areas susceptible (which are the transparency edges and edge
attachments) could be reduced by chemical craze testing all approved cleaning solvents,
sealants, and other substances used, eliminating those substances which may directly attack
the acrylic or attack the polycarbonate in the bolt holes or at the transparency edge. In
addition, better education of field personnel on cleaning techniques may reduce chemical
attack. The effects of flight loads and pressure/thermal loads could be reduced with new
designs by using channel type edge attachments (fastenerless edge attachments similar to
those used by other industries such as automotive/bus transparencies and architectural glass)
or floating bushings. The aft edge attachment on the F-Ill ADBIRT windshield is less than

4"'
optimum because of the constraints imposed by the length of the forward flange on the
existing titanium aft arch. In addition, the bird impact resistance of the existing windshield
configuration could be improved with better edge attachment designs.
A long term solution for windshields such as the F-111, which have demanding
mission profiles, would be to eliminate both chemical craze agents and holes for fasteners in
the transparency. Elimination of the holes in the transparency would in turn eliminate the
stress concentration points and the fatigue cracks. It is possible to develop a channel design
edge attachment windshield which can be changed out by four maintenance personnel in less
than 4 hours with a service life of 4 years or more. Such a design would utilize a two-part
channel, or a channel with an additional leg which itself bolts to the aircraft frame. Safe fast
cure or tacky tape type sealants could be used.

48
 REFERENCES

1. Bowman, Daniel R., Gregory J. Stenger, and Blaine S. West, Full-Scale Birdstrike
Testing of In-Service Aged F-Ill ADBIRT Windshield Transparencies,
WRDC-TR-89-3075, August 1989.

2. Bowman, Daniel R., and Blaine S. West, An Investigation Into the Structural
Degradation of In-Service Aged F-111 ADBIRT Windshield Transparencies,
UDR-TR-90-34, to be published by WRDC, June 1990.

49 *U.S. GOVERNMENT PRINTING omCE: 1,2 64N-127/62-192