Ii!!!
“Wu-13"....“ . .
. . M... k.
. , .1..." 3!. 44.45
. 9.1%., ,. . V . :12." ##3##?
.. i. V . A a .,..u( _
:9. , 53.2%, r. rt V... J .31.. .
. . .
32.2.:

I'D:
{flammuhma
S... 3%..

wife:

.. .V
In .2 .2... . . .33 .
z.- 2. é. , 09. k... V . . Venn I...
kw... 51W... ens.“ , .. .. ., . f #4. iidia!‘
‘ f 3 f I. 7 . . *‘I .1. (‘1.I 2"!
2... ; . . . Li 1 2.5...»
i .fl- . . V Vt. TQQW.
f. .3. 1:1 ‘ , . V V . , T. ‘5
3 .55.“... . V .1
. r. V...» V .
ctwqox~"ls .
mm}; . _. .
mm? .51... r .. ..
V .. .
gnaw-.3... .x,.u!....,s‘...ma...“
...

x
u

.. fl.
..
ma...” tinny”-
. 3.32.115}
. .35.. .. .
z ‘ V . . A ‘ ‘56,...”4
' Y)b¥-.I~. ‘5...)Y‘V K ‘IL.
3-... 1.1. v . 1:... II... .u
9 1...... L. V (3: 91...:
an; .1 , V Ln..." $115.;
‘ q. n .3. if.“ b... p
a. .2: I. . V V - :4.
......u....~.§n .r .2 . V . .
.2. an...

it;

it

2....
is. :9
”3:5- 1.3.x.
3.53.33 any
X... . p
is... .
. 3 .3194?
.L a.
‘

A
I

V 3::
$33....

.55....
V o
6...“..va .

V.
a £21.: J.
3:32-)" I-

f L V
d: Ag 44‘.“

; - 4

s- . 9am. .ar.

$312.23
3:

1......
(2.51:.

3.0!..(i on]:

(to v u u

3 3.52.0).

.0.

.7 1-1.. 3 .

51.9...“ 1‘
J

u.
, .tI-KV do»!
I: . ti. th. l
. 5.:
”3.7024... .A
1.112.: 1:
3.15934.- . .1?! a
~.{:0§.1:|:>=:
:‘ . . : x
it:

‘ .2. 9.3!:
x. 1.3.: x1333:
1. 155.155}: .
.11....gtivtv .. 2:
2... 43:13:.
V r. ’3‘!
V i l. {4“ .XIo
ow . . .41.. . . V .l
V. 5:.
:21}:

in.

x... ‘ I’ll:

V . 2.3%.

.755.

.v. .r.‘ 3.1!. urn .

 

 

wensmr LIBRARIES

WWWWWWWWW WWW WWW WWW 1

31'2 930021

 

LSBHARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled
AN EXPERIMENTAL INVESTIGATION OF THE OPTIMUM JOINING

AND REPAIR OF COMPOSITE STRUCTURES

presented by

Kristin Beth Zimmerman

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Engineering Mechanics

Mntéaofl

Major professor /
Date @"Rfl/Q /?9~5

MSU i: an Affirmative Action/Equal Opportunity Institution 0-12771

 

2W

WW 1‘
WWWW'1WW

W

 

 

 

 

 

WW

 

WW

 

 

 

WW

PLACE N RE'NRN BOXtonmavothl-chockouttmn you'rocord.
1'0 AVOID F INES return on or before date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

;

MSUI AnNflm-tlvo O Inltttwon
0 Action/Emu ppm-unity ‘

 

 

AN EXPERIMENTAL INVESTIGATION OF THE OPTIMUM JOINING

AND REPAIR OF STRUCTURAL COMPOSITES

By

Kristin Beth Zimmerman

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Materials Science and Mechanics

1993

ABSTRACT

AN EXPERIMENTAL INVESTIGATION OF THE OPTIMUM JOINING
AND REPAIR OF DAMAGED STRUCTURAL COMPOSITES
By

Kristin Beth Zimmerman

Because of their high stiffness-to-weight and high strength-to-weight ratios,
composite materials are rapidly gaining acceptance in today’s aerospace industry. As the
technology associated with the utilization of composite materials matures, it is transferred
to many other industries such as the automotive industry. .The primary problems that
need to be addressed in composite structure designs include a thorough understanding of
the characteristics of composites, as well as the repairability of such materials.

The cOmposite materials under investigation are being developed for the
automotive industry, therefore repair techniques must be established to facilitate the
expanding use of automotive structural composite components. Current repair techniques
involve: ( 1) determining the damage mode, (2) selecting the joining technique and
geometry, (3) choosing an adhesive that is compatible with the matrix material, (4)
applying reinforcement patches to both sides of the damaged zone, and (5) assessing the
strength of the repaired composites.

Three different types of composite panels are analyzed. They consist of a chopped
glass fiber in a polyester matrix, a continuous glass mat in a vinylester matrix, and a

cross-woven fabric chemically bonded to randomly oriented glass fibers in a vinylester

matrix. These three samples exemplify most of the composite materials utilized by the
automotive industry.

Artificial damage modes created in the laboratory are used to simulate real composite
damage. Various repair parameters are explored to repair these artificial damage modes.
Results show that by careful selection of geometrical bond line configuration, scarf angle,
filler, and reinforcing patch material, 100% of the composite material’s strength can be
restored. Three-point bending is utilized to assess the restored strengths of the
composites. Results show that the bond line geometry between two separated specimen
parts plays a critical role in the joining efficiency and repair strength of composite
materials.

Based on this research, the fundamental techniques for composite joining and repair
are achieved and the development of an optimal design for superior joining efficiency and

repair strength becomes possible.

This dissertation is dedicated to my parents John C. and Jacquelyn L. Zimmerman
for their relentless, caring, generosity and support throughout my educational career.

There are a few particular phrases that I remember and live by. They state,

The secret of success in life
is to be ready when your opportunity comes,
and do not be afraid to venture, since nothing

ventured, nothing gained.

Another adage states,

A mediocre person knows something happened
A brilliant person knows what happened, but

A genius knows why it happened.

Thank-you mother and dad for instilling in me that the world is out there just
waiting for me. And that I can do whatever I set my mind to.
I’d also like to state a special dedication to my grandfather Lawrence E. Bredahl.

A true educational role model, and a true Michigan State University Spartan!

ACKNOWLEDGEMENTS

The author would like to thank Michigan Materials and Processing Institute for their
financial support.

The author would also like to thank her Ph.D. committee members, Dr. Gary L. Cloud,
Dr. Kalinath Mukherjee, and Dr. Paul M. Parker, for their support and cooperation
during this project.

A special acknowledgement goes out to my advisor Dr. Dahsin Liu for his unending
time, support, and true friendship ........... And for setting off the spark in me to pursue

the very challenging research regarding the repairability of composite materials.

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION Page

1.1 Composite Materials ............................ l

1.2 Composite Repair ............................. 12

1.3 Composite Joining .............................. 14

A. adhesive (bonded) repair using scarf-joining .......... 18

B. mechanical and adhesive joints combined ............ 22

1.4 Previous Research .............................. 23

1.5 Present Study ....................... ......... 23
CHAPTER 2: REPAIRABILITY OF DAMAGED COMPOSITES

2.1 Impact Damage ............................... 26

2.2 Artificial Damage .............................. 27

2.3 Repair or Replace .............................. 27

2.4 Repair Guidelines .............................. 29

2.5 Reinforcing Patch Material and Adhesives ............... 30

2.6 Standard Repair Technique ........................ 30

2.7 Composite Joining Techniques ...................... 32

A. geometric configurations ...................... 33

(a) in-plane bond line geometries ............... 33

(b) through-the-thiclmess scarf joining ............ 38

(c) out-of-plane curvatures ................... 38

iii

CHAPTER 3:
3.1
3.2
3.3
3.4
3.5

3.6

CHAPTER 4:
4.1
4.2

4.3

4.4

4.5

13329

B. associated parameters ........................ 38
(a) reinforcing patch material .................. 41
(b) fillers .............................. 41
(c) adhesives ............................ 44

TESTING AND REPAIR FACILITIES

Composite Cutting Saw .......................... 45
Pneumatic Die-Grinder ........................... 46
Drill Press .................................. 48
Three-Point Bending Fixture ....................... 48
Servo-Hydraulic Tension-Torsion Instron Machine .......... 49
Dynatup Impactor .............................. 49

GEOMETRIC PARAMETERS IN CONIPOSITE REPAIR

Fiber Geometry ............................... 50
Fiber Orientation Analysis (Excel 2415) ................ 51
In-Plane Bond Line Geometries ........... .......... 5 3
A. straight line-crack and circular cut-out .............. 55
B. angled line-crack configurations .................. 55
C. bond line configurations ....................... 55
D. bond line patterns .......................... 56
E. bond line depths ........................... 56
Through-The—Thickness Configurations ................. 57
Out-Of—Plane-Curvatures .......................... 58

iv

22:9

CHAPTER 5: TESTING AND RESULTS

5.1 Impact-Induced Damage .......................... 67
A. Equivalent Damage Hole Size .................. 67

(a) Owens Corning ........................ 67

(b) Excel 8610 .......................... 71

(c) Excel 2415-Warp ....................... 75

B. Effect of Fillers in Circular Cut-Outs .............. 79

5.2 Through-The-Thickness Scarf Joining .................. 79
A. Single-Scarf Joining (Owens Corning SMC) .......... 79

B. Scarf-Hole Study (Owens Corning SMC) ............ 82

5 .3 In—Plane Configurations ..... - .................... 83
A. Bond Line Angle .......................... 83

(a) Owens Corning SMC .................... 85

(b) Excel 8610 .......................... 85

(c) Excel 2415-Warp ....................... 88

B. Bond Line Pattern .......................... 90

(a) Owens Corning SMC .................... 90

(b) Excel 8610 .......................... 92

(c) Excel 2415-Warp ...................... - . 92

C. Bond Line Depth (Excel 2415-Warp) ............. . . 92

5.4 Out-Of-Plane Curvatures .......................... 98

CHAPTER 6: ASSOCIATED PARAMETERS

6.1 Adhesive Study (Owens Corning SMC) ................ 101
A. Evaluation of Adhesives ..................... 101
B. High-Viscosity Adhesives .................... 104

Page
6.2 Reinforcing Patch Study ......................... 107
A. Number of Reinforcing Layers for Butt-Joint ........ 107

B. Number of Reinforcing Layers for Bending Fracture . . . . 113

C. Woven-Patch Orientation Study (Excel 8610) ........ 116
6.3 Aesthetic Repair .............................. 118
6.4 Types of Loading ............................. 118
A. tension and compression ..................... 118
B. critical buckling ........................... 119
C. cyclic fatigue ............................ 126
D torsion ................................ 131

CHAPTER 7: OPTIMIZATION OF COMPOSITE JOINING

7.1 The Strap-Joint ............................... 132
7.2 The 'S-Joint" Design ........................... 133
A. three-point bending ........................ 135
B. buckling ............................... 137
C. fatigue ................................ 137
D. torsion ................................ 139
E. tension ................................ 139
7 .3 Stitching Repair of Torsional Damage ................. 145
7.4 Optimization of Design - A Computer Model ............ 161

vi

CHAPTER 8: SUMMARY ................................ 172
APPENDIX (TABLES) ................................... 174
LIST OF REFERENCES .................................. 238

vii

LIST OF TABLES

CHAPTER 1 Page
Table 1.1 Comparison of Typical Reinforcing Fibers and Their Moduli ..... 3
Table 1.2 Comparison of Fiber and Matrix Properties ............... 4
Table 1.3 The Advantages of Both Mechanical and Adhesive

Joining ...................................... 16
CHAPTER 6
Table 6.1 Listing of the Composite Material’s Characteristic

Material Properties ............................. 120
Table 6.2 Theoretical vs Experimental Load Values for Critical

Buckling .................................... 124
Table 6.3 Experimental Critical Buckling Load Data for the S- and Butt

Joints ...................................... 125
APPENDIX
TABLES CORRESPONDING TO CHAPTER 5:
Table 5.1 Owens Corning Panels: Drilled Hole Survey with Fillers ...... 174
(a,b,C)
Table 5.2 Owens Corning Panels: Impact Study-Specimen

Repaired with Two Glass Patches Per Side ............... 177
Table 5.3 . Excel 8610: Drilled Hole Survey ..................... 178
Table 5.4 Excel 8610: Impact Study For Equivalent Damage Hole Size . . . . 179
Table 5.5 Excel 2415: Drilled Hole Survey with Fillers ............. 180
Table 5.6 Excel 2415: Impact Study-Warp Direction ............... 181
Table 5.7 Owens Corning Panels: Butt-Joint Study ................ 182

(a.b)

viii

Page

Table 5.8 Excel 8610: Butt-Joint Layer Study ................... 184
Table 5.9 Excel 2415: Butt—Joint Layer Study ................... 185
Table 5.10 Owens Corning Panels: Scarf-Joint Survey, Two Reinforcing
(a,b) Patches Per Side ............................... 186
Table 5.11 Owens Corning Panels: Scarf-Hole Survey ............... 188
(a,b)
Table 5.12 Owens Corning Panels: Joint Geometry Study ............. 190
(a,b)
Table 5.13 Excel 8610: Joint Geometry Study .................... 192
(a,b)
Table 5.14 Excel 2415 : Joint Geometry Study-Warp Direction .......... 194
(a,b)
Table 5.15 Excel 2415: Depth-Of-Joint Survey ................... 196
(a,b)
Table 5.16 Excel 2415: Fiber Orientation Study ................... 198
(a,b)
Table 5 .17 Excel 2415: Bending Fracture Layer Study ............... 200
Table 5.18 Excel 8610: Bending Fracture Layer Study ............... 201
Table 5.19 No Damage/Two Patch Layers Per Side Evaluation to

Determine the Influence of the Glass Patches on

Non-Separated Specimen .......................... 202
TABLES CORRESPONDING TO CHAPTER 6:
Table 6.1 Three-Point Bending of Adhesive Coupons ............... 203
Table 6.2 Owens Corning Panels: Adhesive Study ................. 204
Table 6.3 Owens Corning Panels: Impact Adhesive Study ............ 205

ix

12m

Table 6.4 Cumulative Tension Testing Data .................... 206
(a,b)
Table 6.5 Out-Of-Plane Curvature Study on Rockwell SMC ........... 208
(a,b)

TABLES CORRESPONDING TO CHAPTER 7:

Table 7.1 Owens Corning:Torsional Testing Results, Standard Repair ..... 210
(a,b,c,d,e,f,g,h)

Table 7.2 Owens Corning:Butt and S-JointzTorsional Testing, Standard . . . 218
(a,b,c,d, Repair.

e,f,g,h)

Table 7.3 Owens Corning:Torsional Repair Study, Stitching Repair ...... 226
(a,b,c,d,e,f,g)

Table 7 .4 Cumulative S-Joint Results for Three-Point Bending Repair ..... 233
Table 7.5 Tension Layer Study for S-, Butt-, and WW-Joints .......... 234
(a,b) '

Table 7.6 Comparison of All Materials for Each Testing Procedure.
All of the Values in the Table are Average Values .......... 236

CHAPI'ERI

Figure 1.1
Figure 1.2

Figure 1.3

Figure 1.4

Figure 1.5 -

CHAPTER2

Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8

Figure 2.9

CHAPTER3

Figure 3.1

LIST OF FIGURES

Page
Schematic illustrating shear strength phenomena ............. 6
A Bolt Hole in A Bi-Directional Composite Laminate ......... 8
Tailoring the Composite Laminate Around a Bolted Joint ....... 9
Schematic of Bolted vs Bonded Joints ................... 19
Various Bonded Scarf Joints for Monolithic Skin Repair ....... 21
Artificial Damage ............................... 28
Angles of Straight Line-Crack ........................ 34
Bond Line Configurations .......................... 35
Bond Line Patterns .............................. 36
Bond Line Depths ............................... 37
Through-The-Thickness Configurations .................. 39
Repair of Curvatures ............................. 40
Overall Repair Configuration ........................ 42
Types of Fillers ................................ 43
Schematic of Composite Cutting Saw ................... 47

xi

CHAPTER4

Figure 4.1

Figure 4.2

Figure 4.3

CHAPTERS

Figure 5.1

Figure 5 .2

Figure 5.3
Figure 5.4

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Figure 5.10

Excel 2415 : Fiber Orientation Analysis to
Illustrate the Dependence of the Specimen’s
Fiber Geometry and Orientation on the Strength

of the Joint ................................... 52
Patch and Specimen Fiber Orientation ................... 54
Illustration of Different Out-Of-Plane

Curvatures ................................... 59
Excel 8610 Material Undergoing a Scaling Survey ........... 61
Excel 8610 Material. Results for Scaling Study Under

Three—Point Bending ............................. 62
Schematic Illustrating Anti-Clast Phenomena ............... 63
Strain Gaging of Excel 8610 to Study the Poisson

Effect During Three-Point Bending .................... 65
Excel 8610 Material Illustrating the

Poisson Effect ................................. 66

Owens Corning SMC: Drilled Hole Study to Illustrate the
Loss of Strength After Creating Various Void Sizes .......... 68

Owens Corning SMC: Impact-Energy Study. All

Specimen were Impacted then Loaded to Failure

Under Three-Point Bending. The Repaired

Specimen Utilized the Standard Repair Technique ........... 69

Owens Corning SMC: Equivalent Damage Hole
Size Evaluation to Correlate the Damage Zone Size
with the Repair Zone Size .......................... 70

Excel 8610: Drilled Hole Study to Illustrate the Loss
of Strength After Creating Various Void Sizes .............. 72

Excel 8610: Impact-Energy Study. All Specimen
Were Impacted then Loaded to Failure Under Three

xii

Figure 5.11

Figure 5.12

Figure 5.13

Figure 5.14

Figure 5.15

Figure 5.16

Figure 5.17

Figure 5.18

Figure 5.19

Figure 5.20

Page
-P0int Bending. The Repaired Specimen Utilized
the Standard Repair Technique ....................... 73

Excel 8610: Equivalent Damage Hole Size Evaluation to
Correlate the Damage Zone Size with the Repair Zone Size ..... 74

Excel 2415: Drilled Hole Study to Illustrate the
Loss of Strength After Creating Various Void Sizes .......... 76

Excel 2415: Impact-Energy Study. All Specimen

were Impacted then Loaded to Failure Under Three

-Point Bending. The Repaired Specimen Utilized

the Standard Repair Technique ....................... 77

Excel 2415: Equivalent Damage Hole Size
Evaluation to Correlate the Damage Zone Size

with the Repair Zone Size .......................... 78
Owens Corning SMC: Drilled Hole Study Using
Fillers in the Void Regions During Standard Repair .......... 80

Owens Corning SMC: Illustrating the Influence of
Single Scarf Joining, With and Without the use
of Fillers During Standard Repair ..................... 81

Owens Corning SMC: Scarf-Hole Study to
Illustrate both Single and Double Scarfing a Thin
Panel Joint, With Standard Repair ..................... 84

Owens Corning SMC: Off-Axis Joint Geometries

to Illustrate the Effects of Modifying the Bond

Line to the Direction of Load. Standard Repair

Was Utilized .................................. 86

Excel 8610: Off-Axis Joint Geometries to

Illustrate the Effects of Modifying the Bond

Line to the Direction of Load. Standard Repair

Was Utilized .................................. 87

Excel 2415: Off-Axis Joint Geometries to

Illustrate the Effects of Modifying the Bond

Line to the Direction of Load. Standard Repair

Was Utilized .................................. 89

xiii

Figure 5.21

Figure 5.22

Figure 5.23

Figure 5.24

Figure 5.25

Figure 5.26

CHAPTER6

Figure 6.1

Figure 6.2

Figure 6.3

Figure 6.4

Figure 6.5

Page
Owens Corning SMC: V- and W-Joint Analysis Which
Incorporates the Results Found in Figure 5.18 .............. 91

Excel 8610: V- and W-Joint Analysis Which
Incorporates the Results Found in Figure 5.19 .............. 93

Excel 2415: V- and W-Joint Analysis Which
Incorporates the Results Found in Figure 5.20 ........ . ...... 94

Excel 2415: Depth-Of-Joint Analysis to Illustrate

the Strength Contribution From Utilizing Different

Bond Line Depths and Angles. Standard Repair

Was Utilized .................................. 96

Excel 2415: Depth-Of-Joint Analysis to Illustrate

the Strength Contribution From Utilizing Different

Bond Line Depths and Angles. Standard Repair

Was Utilized .................................. 97

Results of Curved Panel Repair After Three-Point Bending
Fracture. Rockwell SMC Material .................... 99

Owens Corning SMC: Adhesive Study to Evaluate
the Influence of the Adhesive System on the
Standard Repair of Damaged Composites ................ 102

Three-Point Bending Results for Adhesive Coupons ......... 105

Owens Corning SMC: Impact-Adhesive Study to

Illustrate the Effects of Utilizing only an

Adhesive System During Repair of Perforated

Impact Damage ............................... 106

Owens Corning SMC: Evaluation of the Strength
of the Glass Patch Reinforcement .................... 108

Owens Corning SMC: Butt-Joint Patch Strength

Evaluation Utilizing Original Thickness and
Repair Thickness Dimensions ....................... 109

xiv

Figure 6.6

Figure 6.7

Figure 6.8

Figure 6.9

Figure 6.10

Figure 6.11

Figure 6.12

Figure 6.13
Figure 6.14
Figure 6.15

Figure 6.16

CHAPTER7

Figure 7.1

Figure 7.2

Figure 7.3

Ease
Excel 8610: Evaluation of the Strength of the
Glass Patch Reinforcement ......................... 111

Excel 2415: Evaluation of the Strength of the
Glass Patch Reinforcement ......................... 112

Excel 2415: Bending Fracture Layer Study to
Evaluate the Restoration of Strength to a Non
-Separated Joint After Standard Repair ................. 114
Excel 8610: Bending Fracture Layer Study to
Evaluate the Restoration of Strength to a Non
-Separated Joint After Standard Repair ................. 115

Excel 8610: Plane-Woven Patch Orientation Study
Standard Repair Was Utilized ....................... 117

Buckling Phenomenon for Owens Coming and Excel
8610 Reference Panels ........................... 122

Buckling Phenomenon for Excel 2415 , both Warp and Fill
Direction Reference Panels ........................ 122

Fatigue Results for Owens Corning Reference and Repair Panels . 127
Fatigue Results for Excel 8610 Reference and Repair Panels . . . . 128
Fatigue Results for Excel 2415(Fi11) Reference and Repair Panels . 129

Fatigue Results for Excel 2415(Warp) Reference and Repair Panels 130

The "S-Joint" Bond Line Configuration ................. 134

Comparative Results for the S-Joint, After
Standard Repair and Three-Point Bending Analysis .......... 136

Results for Different Orientations of the S-Joint,
After Standard Repair and Three-Point Bending Analysis ...... 138

XV

Figure 7.4

Figure 7.5

Figure 7.6

Figure 7.7

Figure 7.8

Figure 7.9

Figure 7.10

Figure 7.11

Figure 7.12

Figure 7.13

Figure 7.14

Figure 7.15

Figure 7.16

Figure 7.17

an:
Owens Corning SMC: Torsional Rigidity for the S—Joint

vs the Butt-Joint. Standard Repair Was Utilized ........... 140
Excel 8610. Torsional Rigidity for the S-Joint vs

the Butt-Joint. Standard Repair Was Utilized ............. 141
Excel 2415-Fill. Torsional Rigidity for the S-Joint vs the Butt- '
Joint. Standard Repair Was Utilized .................. 142
Excel 2415-Warp. Torsional Rigidity for the S-Joint

vs the Butt-Joint. Standard Repair Was Utilized ........... 143
Comparison of S-, W-, and Butt-Joints Undergoing Tensile

Failure. Standard Repair Was Utilized ................. 144
Schematic of the S- and ww-Joint under both Tensile and

Three-Point Bending Loading ....................... 146
Torsional Failure Mode for Owens Corning SMC .......... 148

Torsional Rigidity 0f Owens Corning Material
Before Damage and After Standard Repair ............... 149

Torsional Rigidity of Excel 8610 Material Before
Damage and After Standard Repair ................... 150

Torsional Rigidity of Excel 2415 (Fill) Material
Before Damage and After Standard Repair ............... 151

Torsional Rigidity of Excel 2415 (Warp) Material
Before Damage and After Standard Repair ............... 152

Sketch of Drilled Hole Patterns for Stitching During
Torsional Repair ............................... 153

Torsional Rigidity of Owens Corning Material

Before Damage and after Stitching Repair. The

Thread is Nylon, and No Glass Reinforcing

Layers are Utilized. Mesh Density = (2 x 4) ............. 155

Torsional Rigidity of Owens Corning Material Before Damage
and after Stitching Repair. The Thread is Nylon, and NO

xvi

Figure 7.18

Figure 7.19

Figure 7.20

Figure 7.21

Figure 7.22

Figure 7.23

Figure 7.24
Figure 7.25

Figure 7.26

Figure 7.27

Figure 7.28

Figure 7.29

M

Glass Reinforcing Layers are Utilized. Mesh Density = (3 x 6) . 156
Torsional Rigidity of Owens Corning Material Before Damage
and after Stitching Repair. The Thread is Nylon, and NO

Glass Reinforcing Layers Were Utilized. Mesh Density = (4 x 7) 157

Torsional Rigidity of Owens Corning Material Before

Damage and after Stitching Repair. The Thread is Steel,

and No Glass Reinforcing Layers are Utilized.

Mesh Density = (2 x 4) .......................... 158

Torsional Rigidity of Owens Corning Material Before Damage

and after Stitching Repair. The Thread is Steel, and NO

Glass Reinforcing layers are Utilized. Mesh Density =

(3 x 6) ..................................... 159

Torsional Rigidity of Owens Corning Material Before Damage
and after Stitching Repair. The Thread is Steel, and N0
Glass Reinforcing Layers are Utilized. Mesh Density =

(4 x 7) ..................................... 160
Schematic Showing the '4-Wire-Strand" Configuration ....... 162
Results For Owens Corning SMC With a Four-Strand Wire

Grid For the Torsion Repair Study .................... 163
Schematic of the Finite Element Mesh .................. 165
Computer Modeling of Patch Thicknesses. Normalized Stress . . . 166
Computer Modeling of Butt-Joint Gap Lengths. Normalized

Stress ..................................... 168
Computer Modeling of Butt-Joint Gap Lengths. Normalized

Shear Stress .................................. 169
Computer Modeling of Butt-Joint Gap Fillers. Normalized

Stress ..................................... 170
Computer Modeling of Butt-Joint Gap Fillers. Normalized

Shear Stress .................................. 171

xvii

CHAPTER 1

INTRODUCTION

Composite material components are rapidly gaining acceptance in today’s
aerospace and automotive industries. This is primarily a result of their high stiffness-to-
weight and strength—to-weight ratios. The technology associated with the utilization of
composite materials is slowly maturing, therefore the need for fundamental knowledge
regarding composite materials is pertinent. With the knowledge in hand, many questions

regarding the fundamental characteristics of composite materials can be addressed.

1.1 Composite Materials

This introduction is intended to give a very general understanding of composite
materials, i.e., what they are, how they work, and how they are used. Subsequent
sections will describe the fundamental mechanisms of composite repair in much more
detail.

A thorough definition of composite materials is found in volume 1 of the
Engineered Materials Handbook on Composite Materials [1], it states, ”Composite
material is a combination of two or more materials (reinforcing elements, fillers, and
composite matrix binder), differing in form or composition on' a macro-scale. The
constituents retain their identities; that is, they do not dissolve or merge completely into
one another although they act in concert. Normally the components can be physically

identified and exhibit an interface between one another."

2

Composites such as cellulose, asbestos, and glass fibers with organic resins have
been used since before World War II. These early composites were the precursors to the
advanced composites developed today. Advanced composites are similar to their early
version, except the reinforcing fiber is much stiffer, i.e. , has a higher elastic modulus.
A comparative example of typical moduli for unidirectional composites is outlined in
Table 1.1 [2].

A unique property of composite materials is their anisotropy. Conventional
metals on the other hand are isotropic, meaning their strength and stiffness are
independent of the direction of loading. If a metal structure is deformed too far, it stays
deformed. A composite’s anisotropy means that it can have a strength and stiffness 15
and 30 times greater, respectively, in one direction over another. If it is deformed too
far, it breaks. Table 1.2 illustrates the diversity of properties between graphite fiber and
epoxy resin which are frequently used for composites [2].

Composites follow what is referred to as the rule of mixtures. For example, if
half of the volume is fiber and half resin, the composite properties will be the average
of the two individual properties. Therefore, knowing the volume fraction is very
important in both the design and repair of composite structures. For example, according
to the numbers in Table 1.2, if loading is applied along a sample exclusively made up
of fiber, then the tensile strength equals 300,000 psi. However, if half the volume is

resin, the tensile strength is approximately:

W = 160,000 (W)
2

Table 1.1 Comparison of Typical Reinforcing Fibers and Their Moduli [2] .

 

 

 

 

 

 

 

 

‘ REINFORCEMENT MODULUs, (10‘) PSI l
Canvas (cellulose) 0.5
Asbestos 2
Continuous Glass 6
Aramid 10
Graphite 15
Boron 25

 

 

 

 

Table 1.2 Comparison of Fiber and Matrix Properties [2].

 

 

 

 

 

 

 

 

 

 

—

 

GRAPHITE FIBER EPOXY RESIN I
Tensile Strength (PSI) 300,000 20,000 I
Tensile Modulus (PSI) 30 x 10‘ 0.5 x 10‘
Specific Gravity 2.1 1.1
Thermal Expansion -3 x 10‘5 +400 x 10"
(in/in/°F)
Temperature Limit 2,000 300
('F)

5

In addition, pulling at 90° to the fibers is the same as pulling on only the resin.
Therefore, the tensile strength of the sample is 10,000 psi since half of the volume is
fibers.

The purpose of the resin is to hold the fibers in place and to transfer loads from
one fiber to another. Since composites exhibit low shear strength properties, they are
consequently, designed to minimize shear loading of internal load transfer via shear.
Figure 1.1 illustrates the shear strength phenomenon.

The vast difference between the properties of fiber and resin lends to a design
flexibility not possible with isotropic metals. For example, the design of a helicopter
rotor blade demands very high loads in the blade’s length direction because of its need
to withstand centripetal forces. Loads across the width are not equally significant. A
solid metal blade would have the same strength in all directions, therefore its width
direction is overdesigned. To design the same blade out of composite materials, most
of the fibers would run the length of the blade and very few across the width.
Therefore, the composite blade is tailored to the direction of loading while providing
adequate strength and a large reduction of weight [2].

Since polymer composites follow the rule of mixtures, they can be designed to
optimize directional stiffnesses. If the same amount of fibers run at both 0° and
90° then the strength will be the same in both directions. Strength at any angle away
from the fiber axis will decrease and will be approximately equal to the strength along

the axis times a function of the angle [2]. The largest deviation from the fiber axis, for

 

Figure 1.1 Schematic Illustrating Shear Strength Phenomena.

7
a plane-woven fabric, is 45°. Therefore the strength is reduced from that of the uni-

directional fiber direction. Consequently, when designing composites structures, it is
most advantageous to identify the direction of loading and strictly design strength and
stiffness to that direction.

Continuity of fibers results in anisotropy while discontinuity or random fibers
result in stress concentrations. For example, mechanical joints, such as bolt holes, are
diffith to design because of the unique free edge effects which do not exist in
conventional metals. In addition, any hole cut through fibers will produce a stress
concentration or a weakness [2]. Figure 1.2 illustrates the effects of a bolt hole in a bi-
directional composite. The result is similar to designing a uni-directional laminate with
loads at 90° to the fibers. A common design technique for alleviating the stress
concentrations at and around bolted joints is to design and specify fibers at +/- 45° to
the load as illustrated in Figure 1.3. Another common technique for designing bolted
joints is to move the hole further away from the edge so more fibers can absorb the load
before the bolt ean pull out. Bolted joints are used frequently, but they often introduce
complicated fabrication problems as well as added weight [2]. Also, it should be noted
that with bolted joints and riveted joints, the compressive loads developed during torque
down and "bucking" often create delamination in the composite material.

The added weight and fabrication problems of bolted joints leads to the
introduction of bonded joints. Bonded joints are frequently used and the design and
manufacturing procedures are the same as those for bonding metals, except surface

preparation is easier. With the introduction of bonded joints comes the introduction of

 

$828
BDLT

LOAD
L

 

 

 

 

 

 

 

 

 

 

 

in IAAA

 

 

 

Figure 1.2 A Bolt-Hole in a Bi-Directional Composite laminate.

W—FIBERS

BULT

LOAD
‘— __L

 

 

 

Figure 1.3 Tailoring the Composite Laminate Around a Bolt Hole.

10

resins and adhesives used during repair. An important parameter to consider when
selecting an optimum adhesive system is its temperature usability and capability.

Temperature is a true design constraint in that the resin will degrade if the
operating or design temperature is too high. Some temperature regimes for resins are
listed below:

1. epoxy - up to 300°F

2. bismaleimide - up to 450 °F

3. polyimide - up to 600 °F
All three regimes are utilized frequently because of the ever increasing need for more
durable adhesives. However, it should be noted that processing sophistication and
cost increase with temperature.

The resins listed above are thermosetting resins. They set into a solid and form
an infusible mass when properly catalyzed and heated. However, they can encounter
difficulties from moisture and other solvents [2]. About 1% by weight of '
water will diffuse into thermosetting resins. Moisture acts as a plasticizer, therefore
reducing the mechanical properties of the resin. It also causes micro-cracking, much like
fatigue loading, if the composite is repeatedly heated and cooled. Paint and edge sealants
minimize moisture transport into and away from the composite, but nothing completely
eliminates the effect [2]. Thermoplastic resins, i.e. , PEEK, have been evaluated as
substitutes or replacements for epoxy and other thermosetting resins. These resins are
currently used for organic matrix composites, and they can be repeatedly heated and

reformed. They absorb very little moisture, however they are attacked by hydrocarbon

11

solvents.

Fabrication of organic matrix materials is very important to the overall aspect of
composite design and repair. Distinctive features in the fabrication of thermoset
composites are the critical processing parameters of time, pressure, and temperature
during the cure cycle. All of these parameters determine the mechanical properties of
the molded laminate. The same chemical processing takes place during bonded or
laminated repair.

Monomers, i.e. , small molecular units, are the building blocks of polymers.
Resins are created by chemically combining between 50 to 100 monomers into one big
chain. These chains are then bonded to dry fiber and presented as the product called
'prcpreg" [2]. Prepregging is simply the application of resin to dry fiber. It is often
available in sheet form and most advanced composite laminate structures are formed
through the use of preimpregnated fibers. The fabricators receive prepreg, form it into
a shape, and cure it to keep that shape. The polymer changes from a liquid to a solid
by forming chemical bonds between polymers. These bonds form in two different ways:
by extending the length of the starting polymer chains and by forming bonds between
chains. The bonds between the chains, i.e. , cross links, determine the temperature
resistance and brittleness of the resin.

The cure cycle is also critical since it must allow any entrained air, residual
solvent, absorbed water or water produced as part of the chemical processing, to escape
or to remain inside as a condensed liquid. Voids, and reduced laminate strength, are the

result of not controlling the volatiles. The cure cycle must allow for:

12
1. trapped air to escape,

2. moisture or other solvents to escape or be kept in the liquid state,
3. excess resin to be squeezed out,
4. the mass of fibers and resin to be compacted, and,
5. the resin polymers, hardeners, catalysts, additives, etc. , to chemically
react to form a specified solid, dense mass.
Cure cycles are typically specified by time, temperature, and pressure.

The increased use of composites has introduced questions concerning cost
effectiveness. Many studies and case histories show a 20% cost ' reduction in the
manufacturing of composite parts over metals. This cost reduction is primarily due to
the reduction of assembly time. Assembly time is typically four times greater than

fabrication time. This is what makes composites more cost effective compared to metals.

1.2 Composite Repair

One of the primary questions regarding the utilization and implementation of
composites is their repairability. Many techniques have already been developed to
investigate composite repair [3-12]. However, new composite materials are being
developed and employed at a steady pace thus creating the necessity for a systematic
repair study to restore the strength of damaged composites. A general sequence of
questions to address regarding the repairability of composite structures are:

1. to find the damage,

2. to define the extent of the damage,

13

3. to calculate damage effects and,

a) to use as is,

b) to design a repair scheme,

c) to scrap the part,
4. to specify the repair method if repair is the option,
5. to accomplish the repair,

6. to evaluate the repaired structure.

Evaluating composite damage follows two modes, that of visual inspection, and
that achieved by an overhaul. Visual inspection occurs frequently and with good success,
but the visually hidden defects, such as those hidden under paint, grow slowly and can
eventually grow to a catastrophic size. However, if the rate of growth is slow, visual
inspection will detect the damage before catastrophic damage occurs.

Overhaul maintenance requires the removal of a particular part which is then
inspected often by using ultrasonics or may radiography. Repairs made on location,
i.e. , depot-level repairs, often include both visual inspection and overhaul maintenance
[13-17].

The extent of the damage is frequently evaluated using ultrasonic and x-ray
radiography inspection [18-22], and the effect of the damage must be critically reviewed
by the repair designer. The repair designer must have a comprehensive understanding
of not only the fabricated structure but also the processes available for efficient testing
0f the repair.

Repair techniques developed to date fall into one of the three categories listed: (1)

bolt~on patch, (2) bond-on patch, or (3) cut away damaged material and laminate new

l4

materials. Most of these repair schemes are depot-level repairs and are individually
designed after the extent of damage is determined [23-33]. The major problems in
composite repair for which acceptable solutions have not been found are:

1. moisture penetration into the laminate,

2. acceptable, easily stored materials,

3. repairs for high operating temperatures (over 300°F),

4. repair of joints.

1.3 Composite J olning
The repairability of structural composites must be addressed, not only in the realm

of repair, but also that of joining. An effective method of joining is the key to a
successful and efficient structural repair. Kedward [5] supports this statement, but
discusses further that the weight advantage associated with the use of composites
frequently becomes eroded at the joint. Therefore, the methodology for joining
composite structures, must address the following aspects:
1. the micromechanics level - the fiber-matrix interface phenomenon.
2. the macromechanics level - at the interfaces between layers as
characterized by the so-called free edge problem.
3. the structural level - the interfaces between two or more separate
components as in the conventional joint.
A unique set of challenges is created by the inability of most composite systems to

deform plastically and ultimately reduce the stress concentrations combined with the

15

anisotropy in stiffness and strength properties.

Another parameter to consider during joining is whether or not to incorporate
mechanical methods, adhesive methods, or both. There are definite advantages for each.
Some are stated in Table 1.3. As a general rule of thumb, bonded joints are more
suitable for lightly loaded joints; this is where their high efficiencies can be realized.
Mechanical joints, on the other hand, are predominantly used for high-risk, highly loaded
joints that are subjected to severe fatigue and environmental conditions. The major
disadvantages of the mechanical joints are their increased weight, part count, associated
costs, and surface modifications [33].

Adhesive bonding has been an accepted and widely used technique for composite
repair in aircraft structures. This is mainly due to the high bonding efficiencies and
improved fatigue life. Extensive use of adhesive bonding is utilized in many secondary
structures on aircraft as well. Many articles can be cited regarding the repair of aircraft
[34-75]. For example, 62% of the Boeing 747 wetted area and 35,000 ft2 of Lockheed’s
C5A are adhesively bonded. Some aircraft have even incorporated adhesive bonding of
primary structures, i.e. wing stiffeners, fuselage longerons, and fuselage skin panels.
The most noteworthy company to employ this technology is Fokker Aircraft Co. in their
Fokker F-27.

The use of adhesive bonding has expanded greatly in the past few years to
complement the latest technology in advanced composite structures. It is widely agreed
that the most efficient adhesively bonded joint is the composite-to-composite or

composite-to-metal splice in the form of a scarf or stepped lap joint [76-92]. With the

16

Table 1.3 The Advantages of both Mechanical and Adhesive Joints [3].

Advantages of

Mechanical Joints

Advantages of

Adhesive Joints

 

Tolerance to the Effects of

High Joint Efficiency Index

 

 

 

 

 

Fatigue Loading (relative strength/ weight of the joint)
Ease of Inspection Low Part Count
Capability for N 0 Strength Degradation of Basic
I Repeated Assembly Laminate by use of Cutouts
I High Reliability Low Cost Potential
No Special Surface Potential Corrosion Problems

Preparation Required

 

 

 

Minimized

17
increased usage of bonded assemblies, the requirement of a well defined repair method

is essential to avoid processing anomalies, mistakes during fabrication, and costly
scrapping of large assemblies, etc.

Adhesive joining is an integral factor and procedure in a majority of thin panel
repair. Therefore, it is critical to understand the bonding mechanism by which thin to
moderately thick structures exhibit. Also, it is important to note the remarkable tolerance
for large bond imperfections. However, thicker composite structures exhibit great
sensitivity to both large voids and porosity. In thin composites, i.e. , those that are
2.5mm or less, the flawed bonds are often strong enough, since in a real structure, in
which random bond flaws are surrounded by nominal perfect bonds, any flawed bonds
divert some of their share of the load to the adjacent sound bonds. This effect is
documented extensively by LJ. Hart-Smith [93]. Hart-Smith also comments that the
primary effect of bond flaws is generally a reduction in the thiclmess, or section
modulus, of the members. However, it is important to note that if an adhesive bond, in
a thin adherend, is created without bond line pressure during the cure cycle of the resin,
then in fact an increase in the section modulus can occur. This increase is primarily due
to the air pockets or voids created during resin cure. These voids ultimately reduce the
overall strength of the joint. Flaws in thick bonded structures can propagate
catastrophically, therefore mechanical fasteners are advocated as a ”fail-safe” load path
[93]. As a general rule, Hart-Smith states that it is best to restrict the use of adhesive
bonding to those applications where there is no possibility for local flaw growth. Also,

it is unwise to design or build a purely bonded joint which is weaker than the original

1 8
adherend [93].

In summary, repair documents have been developed by each manufacturer using
their own preferred processing methods and procedures. Depending on the application
and environment surrounding the material, very little variation is required to sustain a
good repair technique.

A. Adhesive (bonded) Repair using Scarf-Joining

As stated previously, the two general approaches to consider during composite
repair are namely bolted or bonded procedures. The factors which determine a particular
approach are: the specific component, laminate thickness, damage size, accessibility, load
requirements, and repair capability.

When considering bonded repair, it is important to categorize the severity of the
damage. For simplicity, three categories of bonded repair are non-structural, secondary
structural, and primary structural repair. Non-structural damage includes scratches,
dents, and other defects confined to the surface of the laminate, so cosmetic repair is
sufficient. Secondary structures usually refer to those components that consist of thin
laminate construction (less than 2.5 mm thick). Upon repair, strength is not a critical
factor though the restoration of stiffness and stability is. Primary structures are those
components that are critical to the operation of, for instance, an aircraft or an
automobile. Bonded repair of such structures is much more complicated since the repair
must be capable of transferring more load [94-98]. Refer to Figure 1.4.

Two types of bonded repair, for both secondary and primary structures, are an

external patch and a flush scarf joint. The bonded flush scarf joint provides maximum

19

.5 name». coecem m> eczem— ue cause—Em v; 2:!"—

zupeazuflau
mun. wwwu... w .H. .D><

Om... I“
me... u
2.1.9

zuccmnotzmo Duh-Em
“mm .ijm mammxnd

 

 

 

 

 

 

 

 

qaxnn
810
DOB.
23...
8am
on»: is
pxouuxzzon
I:
- _
I _ _ r .
xzt _ _ *V" z
8.3
2. 2

 

SHEA

 

 

2:9.—
mag

Good
89¢
Good
89$
83:
23.3

 

141/me 111911132113 INIUr

20

joint efficiency and is applicable where load concentrations and eccentricities must be
avoided. The bonded flush-scarf repair has been found to achieve 60% of its unflawed
or undamaged strength [10]. These repairs require careful preparation and are very time-
consuming regarding machining and application. However, the advantage of scarf
repairs is their load carrying capability, which is higher than most other repair schemes.

Searf bonded repair concepts have been developed for repairing monolithic
laminates up to 10 mm thick. These repairs are assumed to be made in place, where no
access is available from the back side of the damage. The repair is unique to the
particular damage zone and requires the full ultimate design allowable smugth of the
skin. Figure 1.5 illustrates some variations in bonded scarf joints for monolithic skin
repair.

The other major type of bonded repair is the utilization of external patches. In
this case, the load is earried over the damaged zone. Patches are applied over the repair
zone in a stacking sequence which allows for “fairing-in“ of the patch to the adherend
[50]. This alleviates large stress concentrations at the patch ends.

Ultimately, searf-joining with the application of external patches gives the
optimum restoration of the original unflawed adherend strength.

A great deal of documentation regarding scarf joining focuses on the repair of
thick (more than 2.5 mm) skin laminates and honeycomb structures. This is because,
in most cases, thin panel damage is considered non-structural or secondary structural and
is not required to maintain large structural loads.

Recently there has been an increased interest in the application of structural

21

 

RELATIVE

SCARF' - JOINT CONFIGURATIONS JOINT
STRENGTHS

 

 

 

 

1007.

 

A. BASELINE

 

 

 

B. PRECUREIJ SUPPORT

 

 

 

 

 

C. STEP LAP

 

 

 

f 822

D. PRECURED PATCH

 

 

 

 

f 787.

E. ND INSIDE DDUBLER

 

 

 

 

717.

 

F. STEEPER SLOPE

 

M

_/‘ 587.

E. CUNTINUQQS PLIES

 

 

 

 

 

 

Figure 1.5 Various Bonded Scarf-Joints for Monolithic Skin Repair [5].

22

composite materials by the automotive industry. Therefore, the topic of thin panel repair
must be addressed. The literature is in agreement that scarf joining is sufficient for thick
panel repair, but is it the optimum repair design for thin panel repair? This is one of the
questions addressed in this research, and will be formally discussed in Chapter 2.
B. Mechanical and Adhesive Joints Combined

The fatigue life of cracked holes is of major interest, especially to the aerospace
industry. In the past few years a great deal of effort has been focused on devising an
appropriate repair scheme. Full scale fatigue testing of the Royal Australian Air Forces
Mirage III was undertaken to address the problem of repairing cracked rivet holes [47].
Steel interference-fit bushings were installed at the bolt holes which virtually inhibited
crack initiation from the area of the bolted fastener. Bushings are viable for bolted
joints, but not acceptable for riveted joints. A study was conducted to evaluate different
means of utilizing riveted fasteners without the enhancement of fatigue crack initiation.
Conclusions from the study revealed that the use of adhesively bonded rivets significantly
increased the joints life to failure as well as reduced the initiation of fatigue crack
growth.
Some further conclusions are listed below [31]:

(a) the adhesive acts as a barrier to inhibit crack initiation which might

otherwise have been accelerated by environmental interaction;
(b) the adhesive acts as a non-metallic interlayer, thus separating the
rivets and hole surface and reducing the potentially deleterious effects

of fretting;

23

(c) the adhesive provides improved load transfer characteristics at the

section, both before and after crack initiation.

1.4 Previous Research

Prior to this dissertation, the repairability of impact-induced damage in Owens
Corning SMC was explored. The material was made up of randomly oriented chopped
glass (1" fibers) in a polyester matrix. The impact resistance and the notch sensitivity
of the composite were characterized by tensile and flexural tests. An equivalent damage
hole size evaluation was identified through the comparison of impact damage and notch
sensitivity, and it was concluded that none of the damaged portion should be removed
from the impacted composite during the repair procedure. The damaged material was
then repaired with a technique combining resin injection and the applieation of
reinforcing patch material. Resin injection was performed to seal the matrix cracks
which would otherwise result in high stress concentrations from geometrical
discontinuities. Reinforcing patches were employed to compensate for the strength
reduction caused by fiber breakage. It was revealed, from various experiments, that the
repair technique was very effective in restoring the tensile and flexural strengths of the

composite after impact-induced damage [99-101].

1.5 Present Study
Based on the findings from previous research, the fundamental repair techniques

for SMC composites are verified. In view of the variety of parameters associated with

24

the repair of automotive components, such as, the material, its structural configuration,
and loading type, the necessity of a broader study to gain overall knowledge concerning
composite repair is essential. In this study, in addition to the Owens Corning SMC
composite, two other composite materials manufactured by Excel Pattern Works are also
investigated. The first material, Excel 8610, contains a Continuous glass swirl mat in a
vinylester matrix. The second material, Excel 2415, is made up of a cross-woven mat
chemically bonded to randomly oriented chopped glass fibers in a vinylester matrix.
Both of the Excel materials were produced by RTM (Resin Transfer Molding) processes.
These three composite materials incorporate the three major types of fiber geometries and
microstructures utilized in today’s automotive industry.

In composite repair, both the damage mode and repair parameters are of primary
concern since they have significant effects on the repair technique and the bonding
efficiency. In order to cover as many types of composite damage as possible,
perforation, bending fracture, tension, compression, torsion, and fatigue are investigated.
Various repair parameters, such as the use of reinforcing patches, scarfing angles, and
different bond line configurations are also examined. Based on these studies, it is also
possible to present an optimal repair design for damaged composite structures.
Furthermore, different adhesives are examined, and some special repair considerations
are also discussed.

This dissertation outlines the following topics in chapters 2-7: the repairability of
damaged composites, testing and repair facilities, materials and geometric parameters in

composite repair, testing and results, associated parameters, and optimization of

25

composite repair. Chapter 8 summarizes the aforementioned topics and lists some

recommendations for further studies.

CHAPTER 2

COMPOSITE REPAIR

Composite materials are made of at least two constituents: the fiber and the
matrix. Both of these constituents play an important role in the behavior of composite
structures. Depending on the structural configuration, loading type, and boundary
conditions, the damage modes of a composite structure can be very different. Therefore,
the study of composite repair becomes overwhelming. However, in terms of damaged
microstructures, the damage modes can be divided into the following categories: fiber
breakage, matrix cracking, fiber-matrix debonding, and delamination. This
categorization makes it possible to handle composite repair in a more systematic way.
For example, fiber breakage can be restored with compensating fibers, such as the
reinforcing patch material, while the matrix cracking can be repaired by introducing resin

into the cracked, debonded, and delaminated areas.

2.1 Impact Damage
Various types of damage can take place in automotive structural composites.

Among them, impact-induced damage is of major concern because of its high probability.
Therefore, the repair for impact-induced damage in the form of indentation, perforation,
and bending, must be addressed. However, since these damage types are dependent on
the impact parameters such as impact velocity, loading direction, and contact area, the

actual damage size can vary from one specimen to another. Hence, artificial damage is

26

27

more suitable for a systematic and controllable study. In this study, an efficient standard
repair technique is developed to address impact damage. This particular technique is

discussed in detail in section 2.6 and is evaluated in Chapter 5.

2.2 Artificial Damage

As mentioned above, the composite materials used in the automotive industry are
likely to experience impact damage. In view of the damage modes associated with
impact, the investigation of artificial damage in the form of a line crack and circular
perforation was performed. A straight cutting line was used to simulate bending fracture
initiated from an impactor with a linear nose. However, if the impactor has a Small
pointed head, then perforation, in the form of a hole, may occur. A circular cut-out can

be used to simulate this type of damage. See Figure 2.1.

2.3 Repair or Replace
Once the damage mode in a composite structure is identified, a repair strategy
must be established. If a composite structure is severely damaged, then it is desirable
to actually cut and replace the particular damaged section of the structure. However, this
technique can become quite expensive in terms of capital investment, since it requires
stocking a large variety of body components in each respective repair shop. However,
if structural replacement is not pertinent, then a form of artistic repair should be
employed. This technique is currently used to repair automotive bodies made of

Conventional metals. It is based on the individual technician’s judgement and skill

28

 

 

 

 

 

 

 

 

 

(a) Straight Line-crack (b) Circular Cut-out

Figure 2.1 Artificial Damage

29
though some fundamental guidelines should be followed.

2.4 Repair Guidelines

When damage takes place in a composite structure, both the material property of
the composite and the geometry of the structure can undergo change. The change in
material property usually requires replacement or reinforcement, while repair ean be
applied to structures which have undergone only geometrical changes. If the structural
strength is to be restored, then these two factors must be addressed. These are the
fundamental requirements for composite repair.

In a damaged composite structure, both the geometric discontinuity and
irregularity ean create high stress concentrations; therefore, they must be eliminated from
the damaged structure. However, removing the damaged portion of the structure may
further degrade its strength. Consequently, it is important to take advantage of the
damaged section of the composite by utilizing it as a filler during repair. For example,
the damaged section can be mixed with resin to restore structural integrity. If this
procedure is not sufficient, then reinforcing materials and adhesives can be added for
further strength restoration. Hence, the stress concentrations due to geometrical
discontinuity are reduced. However, stress concentrations due to material mismatch can
also occur when introducing the reinforcing patch materials and adhesives to the repair
zone. To reduce the stress concentrations to a minimum, it is advantageous to utilize

l'€=pair materials which are compatible with the composite material being repaired.

30
2.5 Reinforcing Patch Material and Adhesives

Since the composite specimen panels investigated in this study are all made of
glass fiber-reinforced thermosets, glass fabric or reinforcement should be used in the
repair procedure. Therefore, the compatibility of the glass fabric with the fibers of the
composite is satisfied. In addition, it is important to recognize the geometry of the fibers
in the composite. Then, if possible, the reinforcing patch material should be aligned with
the designated geometry and direction to achieve the highest restoration of strength. The
reason that woven fabrics I are selected as the reinforcing patch material in this study is
because of their high strength (as opposed to a chopped fiber or swirled mat) and their
ease in handling during repair.

For most automotive applications, a room-temperature curing adhesive can be
used to produce effective mechanical and chemical bonding at the reinforcing
patch/specimen interface. It is very important to maintain chemical compatibility with
the matrix of the composite and the repair adhesive, since this will help to reduce the risk
of debonding. In this study, an epoxy adhesive system manUfactured by Marblot (trade
name Maraset 658-resin, 558-curing agent) was selected as the adhesive for Owens
Corning SMC repair, because of its low viscOsity, short room temperature curing cycle,
and good strength to failure. A comprehensive study on some other widely used

adhesives was performed and the results are outlined in Chapter 6.

2.6 Standard Repair Technique

The standard procedure for repairing glass/epoxy specimens begins by following

31
the proceeding steps described below [101]. This repair technique is performed on test

specimen measuring 2" x 6" x 0.10" (50mm x 150mm x 2.5mm), primarily in light of
the testing facilities and the materials available, as well as the types of composite damage
to investigate.

(1) WWW, surrounding the artificially damaged (cut
or drilled) area on the top and bottom surfaces of the specimens to improve the
mechanical bonding.

(2) CW from any debris created by the abrading process. This will
insure a better chemical bond between repair materials and the composite.

(3) W113, Low viscosity of the epoxy is critical for
proper wetting of the bonding region around the damage zone. In addition, low viscosity
resin demonstrates superior permeability through the woven glass reinforcing patch
material.

(4) MIX—Ml! to each specimen. When the epoxy is ready, it is poured and spread
evenly to fully wet the repair zone. For reinforcement purposes, glass fillers are
sometimes added to the epoxy.

(5) MMMWW over the repair zone and
massage additional epoxy into the patches until they are evenly saturated. The patch
material is a plane-woven mat with a thickness approximately 0.2mm. With one side
repaired, each specimen must be flipped over and again epoxy and glass reinforcing

patches are put into place.

(0me to both sides of the wet, repaired specimens. Then

32
the repaired specimens are placed between two aluminum plates. Finally, the stacked

aluminum plates containing the repaired specimens are placed in a press and are loaded,
compressively, to approximately 100 psi (0.7 MPa). This load squeezes much of the
excess epoxy and air voids away from the repaired area.
(7) Way at room temperature for minimum of one hour.
(8) WW Once the repaired specimens are fully cured, they are
removed from the press, trimmed, measured, and prepared for testing. The repaired
specimens acquire a glossy smooth finish. However, their thickness is slightly greater
(about 1.0mm greater) than that of the original specimens, because of the addition of the
glass reinforcing patches and resin. The small change in thickness is acceptable in the
repair prowdure since the primary concern is the restoration of strength.

The aforementioned standard repair technique has been verified to be an efficient
and effective procedure for structural composite material repair [101]. Consequently, it

is the standard repair technique utilized throughout the current study.

2.7 Composite Joining Techniques

Composite repair is actually composite joining. Therefore, repair designs often
focus on the bondable surface area surrounding the damaged zone. If this area is
increased, then more contact surface area is created for the reinforcing patch to
mechanically, and chemically bond to the specimen. The larger bonding area therefore
reduces the risk of debonding between the reinforcement patch and the damaged

composite. The technique being used to increase this bonding surface is called scarfing.

33
In addition to scarfing, the bond line geometry can also affect the joining efficiency. The

following sections illustrate various geometries involved in composite joining.
Furthermore, the type of filler used in the joining area is also discussed since this too

plays a very important role in repair strength.

A. Geometric Configurations

The geometric configuration of the bond line is critical to the repair of (a) in-
plane, (b) through-thethicloress, and (c) out-of-plane damage. The following sections
give a brief overview of these three different forms of composite structure repair.

(a) In-Plane Bond Line Geometries

The bond line geometries were prepared by cutting each specimen to designated
angles of 0°, 30°, 45°, 60° and 90° with respect to the in-plane direction of load. Each
of the bond line geometry specimens were evaluated with respect to their depth of joint
(in the plane of the specimens and their bond line angle to the direction of the loading
force. This study was introduced to determine a correlation between the direction of
applied force and the direction of the bond line. The different bond lines are illustrated
in Figures 2.2-2.5.

Another bond line geometry consisted of cutting known shapes onto the 2" x 6"
(50mm x 150mm) panels. Again, all of the bond line geometry specimens were
separated joints and were repaired using the typical procedure listed in Section 2.6,
utilizing two plies of reinforcing material per side. The shapes that were cut into the

specimens were in the form of: V-, U-, W—, UU-, and WW-joints. All of these contours

34

 

 

 

 

 

/

 

 

 

 

 

 

 

 

(a) O-degree . (b) 30-degree (c) 45-degree

 

 

 

 

 

 

 

 

 

(d) 60-degrcc (e) 90-degree

Figure 2.2 Angles of Straight Line-Crack.

 

\/

 

 

 

(a) V-Joint

35

 

U

 

 

 

 

w

 

 

 

(d) UU-Joint

(b) U-Joint

 

\/\/

 

 

 

 

m

 

 

 

(e) 5.10m:

(c) W-Joint

Figure 2.3 Bond Line Configurations.

 

 

 

 

.(a)

36

 

 

 

 

(b)

Figure 2.4 Bond Line Patterns.

 

 

 

 

(C)

 

 

 

 

(8)

37

 

 

 

 

(b)

Figure 2.5 Bond Line Depths.

 

 

 

(C)

 

38

were performed in a 2” x 2" (50mm x 50mm) area in the center of the specimen panel.
(b) Thmugh-The-Thickness Scarf Joining

In searf joining, a bonding area is created that bevels away from the damaged
zone and encompasses the area outward from the zone. Beveling or scarfing the
surrounding area of the damaged zone also reduces some of the stress concentrations
resulting from geometrical discontinuity in this area. The particular scarfing angles of
interest in this study are 5°, 10°, 15°, and 90° with respect to the plane of the specimen
surface. The 90° scarf or so—called butt-joint does not require scarfing. Illustrations of
through-the-thickness scarfing are shown in Figure 2.6.
(c) Out-Of-Plane Curvatures

The aforementioned sections outlined in-the-plane and through-tire-thickness
repair. This information became the baseline reference for the testing of different repair
techniques. Subsequently, the repair of composite structures with out-of-plane curvatures
incorporated the same technique for repair as the two previously described geometries.
See Figure 2.7. The only modification of the repair scheme was in the applieation of
pressure at the bond line during the cure cycle of the epoxy. Further discussion

regarding out-of-plane curvature repair is outlined in Chapter 5 .

B. Associated Parameters
Once the geometrical bond line configuration has been selected, the next few
repair parameters to evaluate are: (a) the reinforcing patch material, (b) the fillers, and

(c) the adhesive selection. The following three sections review these parameters.

39

 

 

 

 

 

 

 

 

 

l l J
(a) Butt Joint

L \/<r J
(b) Single Scarf Joint

l X T
(c) Double Scarf Joint

Figure 2.6 Through-The-Thickness Configuration.

4O

 

Figure 2.7 Repair of Curvatures.

41
(a) Reinforcing Patch Material

The reinforcing patch material selected for all repairs in this study was a plane-
woven glass approximately 0.2mm in thickness. The primary reason for selecting this
material was its ease of handling and its orthotropic properties. The orthotropic
properties allowed for tailoring during repair for optimum efficiency of restored strength.

As addressed previously, two layers of reinforcing patch material per side give
sufficient reinforcement to most of the repaired structure. In fact, the reinforcing patches
simulate a double-lap joint repair on the composite structure. Figure 2.8 illustrates the
details of the double-lap joint design. Some additional parameters associated with the
reinforcing patch material are investigated further, for example, the reinforcing material,
the number of patches, and the size of the patches.

(13) Fillers

For both the scarf and butt joints, different fillers for repair are also presented.
If a scarf-joint is selected as the composite joining technique, then different fillers may
be utilized in the scarfing zone. The effects of the different types of fillers on composite
joining and repair is of major concern. In this study, four types of fillers were
implemented, namely, pure resin, chopped fibers, rolled fabric, and a strap filler. The
fillers are presented in Figure 2.9. The ”strap-joint" filler was an innovative design
implemented into the repair scheme for the repair of large voids. Results are discussed
and illustrated in Chapter 7.

Further comments should also be made regarding the use of chopped glass fibers

(CGF) and a rolled glass patch (RGP) as fillers at the joint. The application of these

42

Reinforcing Patch

 
 

 

 

 

 

 

  

Filler and Resin

Airposite Specimen

Figure 2.8 Overall Repair Configuration.

43

 

 

 

 

 

(a) Pure Resin

 

 

 

 

 

 

(b) Chopped Fibers

 

 

 

?

 

(c) Rolled Woven Fabries

 

 

 

X

 

 

 

(d) Strap Joint

Figure 2.9 Types of Fillers.

44

materials tends to reduce the brittle resin content at the bond line, thereby creating a
more flexible joint. This aspect should be noted and implemented during the repair
design.
(c) Adhesives

The adhesive selected for this study was a two-part room temperature setting
epoxy. The critical factors involved in choosing an optimum adhesive system are to (1)
check its chemical compatibility with the matrix material in the composite, and (2) make
sure its viscosity is low enough to sufficiently wet the woven-glass patch material. If the
chemical and mechanical bonds between the reinforcing patch material and the composite
adherend are sufficient, then delamination of the patch during testing will be alleviated.

A survey to evaluate different adhesive systems is discussed in Chapter 6.

CHAPTER3

TESTING AND REPAIR F ACHJTIES

Designing a repair facility to accommodate various aspects of composite repair
became a very important and beneficial factor in this particular study. Previously, there
were very limited machining facilities to prepare all of the desired repair configurations.
To date, all of the specimen designs are more effectively and efficiently prepared in the
laboratory. Both the quality and quantity of design specimen can then be achieved

throughout the testing procedure.

3.1 Composite Cutting Saw

The critical component of the composite repair facility is the modified lapidary
saw. The lapidary saw, which is contained in a thin-walled steel tub, was completely
dismantled, and modified to become an exquisite wet/dry composite cutting saw. An
advantageous feature of the saw is its blade orientation with respect to the traversing
cutting table surface. The space and orientation allows for the implementation of
different size saw blades. For this particular study, an 8" diameter diamond-edge blade
is utilized. The speed of the diamond-edge saw blade can be modified by simply
tensioning the drive pulley from the motor.

A precision saw table was designed to mount directly atop a traversing guide
apparatus. The guide apparatus travels parallel back and forth past the diamond blade,

and also has 5" of travel normal to the blade. Since the table was designed with two

45

46

precision fence slots, a 180° mitre was implemented for cutting various angles into the
test specimen. Figure 3.1 illustrates a schematic of the composite saw.

' The composite saw can be used in both the wet and dry state. Since there is
neither burning nor significant edge damage in the glass composite material during the
cutting procedure, the dry mode of cutting is used with the application of a shop-vac
nozzle at the location of the blade to draw the dust and debris. For thicker specimen
panels and carbon fiber panels, because of excessive burning heat along the cutting line,
a portable mister may be hooked into the system and mounted directly at the cutting edge

of the blade.

3.2 Pneumatic Die Grinder

Another feature of the repair facility is the application of a pneumatically
controlled die grinder. The mini-die grinder was designed to accommodate either a 2" ,
3", or 4" sanding disc. Three different grades of sand paper were purchased, but the
coarsest grit (36) was used almost exclusively, due to its efficiency in providing a
sufficiently rough or abraded surface.

The die grinder is very light weight and easy to maneuver. It is simply hooked,
via a 1/2" diameter hose, into an air-fitting and is ready to use. The primary use of the
apparatus is in the scarfing procedure of composite repair, as well as, during the surface
abrading portion of the repair procedure. To alleviate excessive dust in the preparation

area, a vacuum system (a shop-vac) is mounted, via a nozzle, on the work bench and all

47

VACUUM HOSE

  
    
 
  

NOZZLE

DIAMOND SAW
BLADE

SAW TABLE

MITRE
TOP-VIEW

W

i

[90' ADJUSTMENT
FOR SAV TABLE

 

 

 

 

 

 

 

 

 

 

 

 

CUT AWAY SIDE-VIEW

Figure 3.1 Schematic of Composite Cutting Saw.

48

of the grinding is demonstrated at the intake nozzle end.

3.3 Drill Press

A table-top drill press (1/2” chuck) was also utilized to further develop the repair
facility. This apparatus was used almost exclusively for the circular cut-out portion of
this study. In part of the circular cut-out study, a plug was press fit back into the drilled
hole of the specimens. These plugs were created by using a core drill or hole saw in the
drill press apparatus.

For the study on in-plane bond line contours, the drill press was used like a scroll
saw by utilizing a tin coated end mill in the chuck. This technique worked remarkably
well in the cutting of special in-plane curvatures on the specimens. However, a severe
drawback incurred with hole and contour preparation in composite materials is the
durability of the high spwd steel and carbide tools. They aren’t tough enough to
withstand the hardness of the matrix supported glass fibers. A possible solution would

be to utilize a diamond impregnated drill bit and end mill.

3.4 Three-Point Bending Fixture

The testing procedure used to evaluate the particular repair technique utilizes the
three-point bending fixture on the Instron testing machine. Each 2"x 6" (50mm x
150mm) specimen is placed in the three-point bending apparatus and loaded until failure
occurs. The span between supports on the apparatus is 4" (100mm) with the supporting

pins located symmetrically at 2" (50mm) from the center and the loading pin located at

49
the midspan. During testing, the loading pin is always symmetrically positioned atop the
repair zone. The maximum strength is calculated using the flexural formula and is
recorded, along with maximum load, flexure, and damage mode at failure for each
particular specimen. Strength results are then compared to the various repair techniques,

configurations, and the reference undamaged strengths of each material.

3.5 Servo-Hydraulic Tension-Torsion Instron Machine

Tests such as tension, compression, buckling, torsion, and cyclic fatigue utilized
a servo-hydraulic tension-torsion Instron machine. Specimen preparation did not require
the use of end-tabs, however, each specimen was abraded at the gripping area to alleviate
slippage in the grips. Further discussions regarding the actual tension, compression,

buckling, torsion, and fatigue analysis and results can be found in Chapter 6.

3.6 Dynatup Impactor

To evaluate the similarity between artificial damage and impact-induced damage, a
series of tests were conducted by first impacting, to the point of perforation, glass/epoxy
panels measuring 4" x 6" (100mm x 150mm). The Dynatup impactor utilized a tup
which measured 1/2" (13mm) in diameter. After impact, the perforated specimens were
trimmed to the dimension of 2" x 6" (50mm x 150mm) and tested under three-point
bending to evaluate the residual flexural strength integrity. Results of this testing are

outlined in Chapter 5.

CHAPTER 4

GEONIETRIC PARAMETERS IN COMPOSITE REPAIR

In the study of composite repair, both fiber geometry and matrix type are critical
parameters. The following sections discuss the various fiber and joining geometries that
were evaluated, as well as the effectiveness and efficiency of these geometries in the

repair of damaged structural composites.

4.1 Fiber Geometry

Three different fiber geometries were investigated. They were: (1) a random
chopped glass (1" fibers) in a polyester matrix, (2) a continuous glass fiber mat in a
vinylester matrix, and (3) a mixture of cross-woven glass and random glass in a
vinylester matrix. The random glass fiber was formed using SMC (sheet molding
compound) while the other two configurations were formed using RTM (resin transfer
molding). The fiber volume fractions for all of the materials were approximately 50% .
These three fiber types represent the majority of composite materials used today in the
automotive industry. The materials mentioned above are subsequently designated as, (l)
Owens Corning SMC, (2) Excel 8610, and (3) Excel 2415, respectively. All of the
specimen panels have an approximate thickness of 1/8" (3mm).

Because of their random fiber orientation, both Owens Corning SMC and Excel
8610 have isotropic properties. However, since Excel 2415 consists of an unbalanced

cross-weave, it exhibits two dominant directions, namely the orthotropic principle axes.

50

51

The higher strength direction is called the warp direction. Normal to the warp direction
is the fill direction. Hence, the warp and fill directions should be evaluated separately.
In all of the subsequent investigations regarding Excel 2415, both the warp and fill
directions are identified. Based on preliminary testing, the average flexural strengths of
the undamaged specimens are as follows: 290MPa for Owens Corning SMC, 300MPa
for Excel 8610, 478MPa for Excel 2415 in the warp direction, and 375MPa for Excel

2415 in the fill direction. These undamaged values are used as reference values for

comparing all repair strengths.

4.2 Fiber Orientation Analysis (Excel 2415)

This study was conceptualized because of the extreme strength associated with the
warp direction of the 2415 material. A survey was designed to prepare several 2" x 6"
(50mm x 150mm) specimens from an Excel 2415 panel. Each set of three specimens
were cut and referenced to the direction of warp. For example, specimens were cut with
an angle of (0°, 30°, 45°, 60° and 90°) from the warp direction. All of the specimens
were loaded to failure utilizing the three-point bending apparatus. After failure, each
specimen was repaired utilizing two layers of reinforcing patch material per side and
Resin Services resin. After repair, each specimen was again loaded to the point of
fracture. Results are illustrated in Figure 4.1. The two lines at the top of the plot
illustrate the average non-damaged strength of both the fill and warp directions after the
application of two layers of glass reinforcing material per side. These two lines reveal

the influence of the glass patch strength on the repair of non-separated specimens. It

52

6855 83 DEA—om 2855 452 05 ..e 59.86 05 :e 0.93 «5:3 23
e5 5:588 be". 05 he coco—e530 05 3832: 0. mean—«.2 8:555 LEE ”Ben .83 _.v 2:»E

accrues zopomma “a; :8: u..oz<

 

 

 

om mu om we on m. o
W W b r W _ t . . r . F . . _ . . L 0
egg oz 0
axiom a
mom «we be ~\uo§<o 02143.... I oo.
mew mun. be «huge 021%.; 1..
com m
s m 3
con W
n. m I.
m H
n 09. \I
W
d
.68 m\
ncom

 

 

 

53
should be noted that the 45° orientation angle did not fail under the initial three-point

bending loads. In fact, it seemed to twist in the fixture and demonstrated flexure values
well above 1" (25mm) within the 4" ( 100mm) testing span. In addition, results showed
that the 45° orientation maintained the highest repair efficiency. This is primarily due
to the fact that the 45° fiber orientation creates the largest off-axis angle between the
specimen’s fiber orientation and the bending direction. Hence, its flexural property is
more dominated by the matrix prOperty instead of the fiber property and the repair
efficiency for matrix is higher than that for fiber. Refer to Figure 4.2 for an illustration
describing the patch and specimen fiber orientation.

The most important contribution of this survey was the information regarding the
dependence that the fiber orientation has on the flexural strength of the specimen.
Consequently, if any off-axis (off the specimen’s warp direction) orientation
is introduced into the repair design, then the strength is changed. This aspect was

documented throughout subsequent testing of the Excel 2415 material.

4.3 In-Plane Bond Line Geometries

The geometrical parameters play very important roles in composite repair since
they have significant effects on bonding efficiency and repair strength. Various
geometrical parameters such as shape, size, and configuration were investigated in this
study.

Depending on the loading situation, a damaged composite structure may exhibit

a small portion of truncation or an unseparated damaged zone. Therefore, the

54

  
  

“WOVEN

 

 

 

 

 

 

 

 

 

 

 

 

 

PATCH MATERIAL

 

 

 

 

 

 

 

 

 

 

 

; VAR? WARP
DIRECTION /
,3? ~ —
DIRECTIEN ~3,I'»'2[LL
ALPHA=0 IEGREES ALPHA-30 IEGREES
WARP
yRP
'—’ —7 '——-—>VARP
\nLL FILL
ALPHfi-lls DEGREES ALPHA=60 DEGREES
tFILL

 

ALPHA-=90 DEGREES

Figure 4.2 Patch and Specimen Fiber Orientation.

55
repairability of both the truncated and unseparated cases must be investigated. The

specimens prepared with circular cut-outs fall into the truncated category while the line
crack, due to failure, falls into the category of unseparation. These two scenarios
illustrate that the joining technique must be tailored to meet the requirements of the

individual damage modes.

A. Straight Line-Crack and Circular Cut-Out V

The straight line-crack and circular cut-out were investigated because of their
similarities to the damage modes involved in a composite material during an impact
event, i.e. , bending fracture and impact-induced perforation, respectively. Refer

back to Figure 2.1, which illustrates these two types of artificial damage modes.

B. Angled Line-Crack Configurations

In Figure 2.2a, the crack line is oriented parallel to the direction of loading or the
bending direction. However, this is a very idealized case, therefore the evaluation of line
cracks at varying angles to the direction of loading must be evaluated. The evaluation
revealed the influence of the bond line angle on the repair strength. Figures 2.2
b,c,d,and e, illustrates the angles of interest, i.e., 0°, 30°, 45°, 60°, and 90°, respectively.

The results of this survey are discussed in Chapter 5.

C. Bond Line Configurations

In addition to a straight bond line, the boundary along the line of damage (the

56

damaged boundary contour), may illustrate a more complex configuration. In order to
simulate the different possibilities, various configurations of bond lines were investigated.
For example, the V-joints and W-joints shown in Figures 2.3a and 2.3e, respectively,
were deve10ped utilizing the concepts from the previous straight line-crack study. The
U-joint and UU-joint, depicted in Figures 2.3b and 2.3d respectively, were two
modifications aimed at reducing the stress concentrations along the bond line edges of
the V— and W-joints. In addition, an S-joint, as shown in Figure 2.3e, was also
investigated. The S-joint was derived from a circular configuration 'which, theoretically,
should result in uniform bonding efficiency and repair strength when subjected to loading

from any direction. The S-joint is discussed further in Chapter 7.

D. Bond Line Patterns

With the same joining configuration as the V-joint, the joining efficiency and
repair strength can be affected if the joining pattern is repeated within the same repair
zone. The configurations shown in Figures 2.4a, 2.4b, and 2.4c, are all based on the
V-joint and are designed with identical depths. However, by altering or repeating the
number of V—configurations in the V-joints, the repair strength and efficiency is affected,
which introduces another area of concern in composite repair. The results of this survey

are discussed in Chapter 5.

E. Bond Line Depths

This study investigates the depth (in the plane of the specimen) of a particular

57

joint. The V-joint was the reference configuration for this study, and the depth of the
V-joint was modified. Figures 2.5a, 2.5b, and 2.5c illustrate the varieties of V—joint

depths investigated and the results of the analysis are outlined in Chapter 5.

4.4 Through-TheThickness Configurations

As mentioned before, the bonding configuration in the thickness direction also
plays an important role in composite joining and repair. Figure 2.6a illustrates butt-
joining, i.e. a 90° bond line, which is normal to the specimen surface. If the angle is less
than 90°, then the type of joint is referred to as a scarf-joint. Figure 2.6b illustrates a
typieal scarf-joint with scarfing angle alpha (a). One of the important aspects of
studying scarf-joining is to investigate the effects of the scarfing angle on the joining
efficiency and repair strength. Since Figure 2.6b is designed with searfing on only one
side of the joint, it is considered a single-scarf-joint. The double-sided scarf, or double
scarf-joint, as shown in Figure 2.6c, is another possible technique for composite joining
and repair.

In addition, it is important to note that when utilizing the scarfing technique in the
laboratory, the artificial damage modes of interest are both the straight line-crack and the
circular cut—out. Both of these damage modes represent possible models to simulate
impact damage. The straight line-crack, with scarfing, is prepared by cutting the 2" x
6" (50mm x 150mm) specimens in half and machining the designated 5°, 10°, and 15°
scarf angles. Again, the scarfing procedure is demonstrated to determine if the increased

surface area around the damaged zone will indeed increase the flexural strength of the

58
repaired joint. Furthermore, a circular-scarf joint study was investigated. Each

specimen was prepared by drilling a designated diameter hole through a 2" x 6" (50mm
x 150mm) composite specimen. The hole was drilled to simulate a perforated zone in the
composite material. The scarfing regions were then prepared radially outward from the

hole.

4.4 Out-Of-Plane Curvatures

The material used for the of out-of-plane curvatures was an SMC made up of
28% , by volume, random chopped glass in a polyester matrix. The actual structure was
a liftgate designed by Rockwell Plastics for Ford’s Aerostar Minivan. Six different
curvatures were evaluated throughout the structure. Each curvature was cut to the
dimensions of 2" x 6" (50mm x 150mm) and loaded to failure under three-point bending.
A detailed discussion regarding the repair and testing of curved panels ean be found in

Chapter 5. Figure 4.3 illustrates the variety of curvatures investigated in this study.

59

 

 

//[//l

R1 R6 R2, R5

Figure 4.3 Illustration of the Different Out-Of-Plane Curvatures.

CHAPTER 5

TESTING RESULTS

Three-point bending was selected to asses the repairability of damaged composite
structures because of its similarity to many real loading situations and its ease in
specimen preparation (compared to uniaxial loading which requires the application of end
tabs). The dimensions of the individual specimen were chosen to be 2" x 6" (50mm x
150mm), because of the availability of the testing apparatus and the versatility of possible
repair design. However, the dimensions of the test specimens raises a question of its
identity. Is it a beam-type or plate-type structure? To verify that the 2" x 6" (50mm x
150mm) specimens could be analyzed using beam theory, a scaling experiment was
conducted. In this experiment the material Excel 8610 was utilized. Three samples of
each of the following widths, 0.5” (12.6mm), 1"(25.4mm), 1.5”(38mm), and 2"(50mm),
were cut, measured, and fractured under three-point bending. Refer to Figure 5.1. The
results, illustrated in Figure 5.2, reveal a negligible effect due to the width dimension of
the test specimen.

This discussion leads to the observance of another phenomenon which can occur
in composite plates undergoing three-point bending. The phenomenon is referred to as
"anti-clast". This phenomenon is illustrated in Figure 5.3, and can be described as a
saddle type bending of the structure. If the dimensions of the specimen were more like
a beam then this phenomenon could be neglected. This effect occurs, not only due to

the geometry of the specimen, but it is predominant in woven-fiber composites which

61

EXCEL 8610 MATERIAL: SCALING STUDY

 

 

 

 

 

 

so"

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10511 1.0" L501 2.011

 

 

 

 

Figure 5.1 Excel 8610 Undergoing a Scaling Survey.

62

scam 9:3 8a 233 are“: 28 .25 an 2:2".

$.28; 1.5.; zuzomam
m; o; md

 

 

0N . o
. , b _ 0
.manm oz\me§<o oz a
too.
icon
* I
m m x -8»
a

 

 

 

00¢

(de) HroNz-lars

63

 

R "

Figure 5.3 Schematic Illustrating the Anti-Clast Phenomena.

64

exhibit substantial amounts of anisotropy. In other words, it occurs in composites which
contain two different, yet dominant fiber directions.

To evaluate this phenomenon, bi-directional strain gage rosettes were mounted to
the surface of the Excel 8610 material. Figure 5.4 illustrates the placement and
orientation of each of the rosettes. One rosette was mounted symmetrically at the center
of the specimen and directly underneath the loading pin. The second rosette was
mounted midway between the center of the specimen and the edge, and the third was
located very near the Specimen edge. The specimen was loaded under three-point
bending, therefore all of the strain rosettes were mounted to the bottom (tensile) side of
the specimen so as to not interfere with the loading pin. The results reveal that readings
taken from the center-mounted rosette were purely axial, therefore verifying the negation
of transverse strain. Also, the transverse strains increased in a direction outward toward
the edge of the specimen and the results are illustrated in Figure 5.5. These results
verify the possibility of an anti-elastic phenomenon being produced in the 2" (50mm)
specimen during three-point bending. Therefore, this phenomenon is documented and
during the calculation of stress, is negated by utilizing the flexural formula.

Calculating the stress at failure for repaired composite specimens is a very
complicated, and nearly impossible, task. This is due to the variance in the thickness
dimension caused by the application of reinforcing materials and resin. An accurate
assessment of the fiber content in the damaged zone is impossible. Consequently, it was
proposed that the original thickness of the specimens would be utilized in all subsequent

stress calculations due to flexural loading. By setting this precedence, the comparison

65

//RSG RDSETTES

//
i// ../
i/ A

 

 

 

 

 

 

/\
V
.<

 

 

 

V V

BENDING STESS SHEAR STRESS STRAIN DURING
ANTI-CLAST

Figure 5.4 Strain Gaging of Excel 8610 to Study the Poisson Effect During Three-
Point Bending.

66

000m

SE :85. 2.. "cogs: area: 28 .86 on 2:5

3851223 22um .23

 

 

 

 

ll-
ooo.‘ ooon . ooo~ ooo. o no
_ . . _ a 0 VNV
x s. S
. o
.\ 3
. . room 5
. uo
. W
N
. . toms x
.. _
a .. . . -ooo. )
m.
c 3
Boga am) a fiomup m
Aozoue .225 5.3% ..u x _
Antone .53 Bag ..u o m...
89 m.
U
(

67

of original strength to recovered strength became the differential comparison during load
to failure between the two situations. This differential mode of comparison was

maintained throughout all testing procedures.

5.1 Impact-Induced Damage
A. Equivalent Damage Hole Size

An evaluation of the equivalent damage hole size began by utilizing the Dynatup
impactor to create perforated impact damage in several 4" x 6" ( 100mm x 150mm)
specimens. In this analysis, parameters such as impact velocity, impact energy, and total
energy were documented. After impact perforation, each 4" x 6" (100mm x 150mm)
specimen was trimmed to the dimensions of 2" x 6" (50mm x 150mm) and fractured
under three-point bending. The residual flexural strength was then evaluated. The
flexural strengths of the impacted specimens were compared to the drilled hole specimen
to determine the equivalent damage hole size created by the impactor upon perforation
of the specimens. In addition, the repairability of circular cut-out and impact-induced
damage was also found. The motivation of » this analysis was to determine a correlation
between the void zone, damage zone, and repair zone ,_ to evaluate composite damage.
All three different fiber geometries were evaluated and the results are discussed in the
following sections. Actual tables containing the raw data values can be found in the
appendix of this document.
(a) Owens Corning SMC

The results for the Owens Corning material are illustrated in Figures 5.6—5.8. Figure

68

 

8% 2.5 ”85> usage as? saga ea 23 2.. egg: 9 been 23. 855 62m “558 22.5 en 2:3...
EEE ”Sm So:
no and mud di
b p b b b b P o
f
AHOOF
. -oou .5.
mo
m
con 9
. I.
e H
100* \r/
. W
d
noon m\
50: zmao o
mmm<oom as: l. nooo
mo<z<o oz 0 .

 

 

 

 

DON.

.3359; goon Ego 2: Baa: Samson naive 2:. .2655 .5835. 32:
223m 2 woos.— 55 oQoooE 803 5.50on =< Seam 3.85389:— HUS—m mEEoU 2030 En 83E

$283 655 5&2. SSE
o¢ mm. on new cu m... OFF m o

 

_ _ _ L _ O
o9
o o so 0 o
o o oou 5
no
m
oon m
W! X C H
in X f .d
x xx room m\
a
* U030 02 O IOOQ
32am”. ozv oEoEz o
92%. fits Emanum a .

 

 

 

005

7O

.85 Son has.

05 .23 03m oooN ouoEaQ 05 922.60 8 ecu—3.9m— oflm 03: anon—an 283,33 "02m mom—coo ”530

EES mum Bo:

md Nnd mud N _..o

 

 

 

Ho: zuao a
mum 5oz mo<z<o 535:3 Ix.
mo<z<o oz 0

 

nausea

_ o
.. o2

oom m...
a a
... 3N.—
.. oon O
1 I.
c H
loos. )
. W

d

room 1px
z

tooo

f

 

 

CON.

71

5.6 illustrates the amount of residual flexural strength in each specimen after drilling a
known hole size into each of the specimen. A linear least-squares line is also depicted
in the diagram. Figure 5 .7 illustrates the residual flexural strength after perforated
impact damage. The plot also depicts the flexural strength of particular specimen after
the repair of perforated impact damage. Some of the impacted specimens were also
repaired utilizing the repair scheme outlined in Chapter 2, and then fractured under three-
point bending. In addition, it is advantageous to utilize the debris surrounding the
perforated zone as the reinforcing material. Therefore, the superior flexural strength of
the repair specimens, depicted in Figure 5.6, is primarily due to the strength of the glass
reinforcing patch material.

To correlate the equivalent damage hole size, Figures 5.6 and 5.7 are
superimposed and represented in Figure 5 . 8. Note that the resulting equivalent damage
hole size in the SMC material is greater than the 1/2” (13mm) impactor. Having
examined the damaged specimen further, it is verified that the damage zone is actually
larger than the void zone. Therefore, the repair zone must encapsulate the larger damage
zone area.

(b) Excel 8610

The results for the 8610 material are illustrated in Figures 5.9-5.11. The trend to
notice in this evaluation is that the total energy absorbed into the panel is again quite
high. Therefore, the equivalent damage hole size, illustrated in Figure 5.11, reveals a

hole size slightly larger than 1/2" (13mm), though smaller in size than the SMC material.

.85 2.5 ”85> «€35 .22 593m B 33 o... 25.2: 9 away. as: BEE 5.8 .8xm an 2.9".
EENV mum Bo:

 

 

 

md mhnd mud mu Pd 0
b b n o
a
Tom:
D D
m com
2 m
con
#09.
room
So: zmao a .
mega oz 0

 

 

00m.

(0cm) HloNaals

73

Ca.

3358.. use... 265m 2.. Bass 5::8..« 8.33. 2:. $523
.5835. 325 225 2 333 85 38:5 203 Saga .2 seem Eucméaé 5.8 36 e .m 2am

3283 SEE 5&2. .52

 

 

on on ma cu m. o. m o
_ _ _ _ L _ p _ 0
82mm”. ozv BEE; o
5&2. fits BEEN". x
mo§<o oz o 2:
o o o o o x 0 com 5
23x .mm
N
x x x x can m
an .. H
* *
r09. )
. W
d
1000 MW
a
iooo

 

 

CON.

74

.88
25a his. 2.. .23 8a 25 ego 2.. 2238 2 Sosa—gm 85 2o: $2.3 53:5 ”28 ram :6 2:5

955 mum So:
no ammo and auto o

 

 

 

 

S

i

no

3

N

9

I—

H

j.IoorV )

. W

d

100m Inw\
r

mum So: uo§<o 532:8 * room
So: zmao o .
‘ mo<z<o oz 0

 

 

005

75

The difference in equivalent damage hole size between the SMC material and the
RTM is primarily a result of the fiber geometries of both materials, i.e., the SMC is
random, chopped glass and the RTM is continuous swirl glass. The fiber continuity of
the RTM material tends to be a bit stiffer under the same loading conditions. Note that
the actual void region in the 8610 material is visually larger than the void region formed
in the SMC material. However, the equivalent damage size data reveals a zone that is
smaller than the SMC. Again, this difference in equivalent damage hole size is primarily

due to the difference in fiber geometries.

(c) Excel 2415-Warp

The results for Excel 2415 are illustrated in Figures 5.12-5.14. The results illustrate
equivalent hole sizes between 1/4” (6.35mm) and 1/2" (12.6mm). This would imply that
the SMC and 8610 panels are absorbing more energy than the 2415 panels. Note that
the fiber geometry of 2415 is a rigid cross-woven mat with the addition of random
chopped glass chemically bonded into it. The rigid structure of the 2415 material
displays very little deflection upon impact. The results show that it also absorbs a larger
amount of energy. Furthermore, there is also a direct correlation between the total
amount of energy absorbed in the panel and the size of the damage zone. If the panel
absorbs a high amount of energy upon impact, then the damage zone envelopes a much
greater area than the void zone or perforated hole zone. This conclusion is verified by
comparing Figures 5 .7, 5.10, and 5.13. Note also that the fiber orientation throughout

the drilled hole survey and the equivalent damage hole size study is in the warp direction.

aim o_o> 25.53 ”grew b¢< 595% .o 33 2.. Swazi 3 him 2o: .6th ”flea .8xm a. .n Bani
Asimov mum 5oz

mNod md 0506 mud mm _..o o

. afisumoéo oz 0
Ho: 2on a .
$938 3 .l loo.

 

76

 

(DdW) HlONaals

 

 

 

77

225.8... :83. 26:5 2.. Baa: 858% BE»... 2:. .msoso
€8.85 525 223. 2 883 85 .9an 203 Beacon .2 seem 98m-.§§ ”a: .85. a .n 2:5

$283 555 5E2. 458
0% mm on nm em my 9.. m o

 

 

_ _ _ _ _
5&2. ants mmssnommzbm a o
9.5% ozv amssnoEoEz. o ..
dcluo<z<o oz o 1o?
mfislmogo oz 0 .
room 5
. Ho
3
o woon w
o o I.
a. o o o H
oz 0oz \/
W
an d
x x oom {ox
x
a
ooo

 

 

, con

78

. 6.3m
2.5 £8”. 2.. .23 8a 25a «go 2: aspcoo 3 8:236 85 2o: guano .5335 as.“ .86 z .n Bani

SSE mum Bo:

 

 

 

mmoo no mood mud no; o
. . . magnmoéo oz o o
Bo: zmmo a .
mum Bo: mo§<o 5.42238 * too.
a m room 5
u - .. a
m n _.N:
n. m noon 9
u .. l
I .. H
D . 00¢ \/
”n W
3 .o
m oom (ox
ooo

 

 

- CON.

79

Additional tests on both the warp and fill directions are discussed in the following

sections.

B. Effects of Fillers in Circular Cut-Outs

The effect of the circular cut-out and the influence of using fillers in the damage
zone is illustrated in Figure 5.15 for the Owens Corning SMC material. It can be seen
that by increasing the hole size, the integrity or strength of the specimen is reduced
significantly. Note that a non-dimensionalized hole size of 2R/W is used in the diagram.
During the repair procedure various fillers, such as, resin, chopped glass, a plug and a
strap were implemented into the void zone. The results conclude that regardless of the
damage size, a filler/patch combination can be added to restore the original flexural

strength of each specimen. This trend stayed consistent for the Excel materials as well.

5.2 'I‘hrough-The—Thickness Scarf Joining (Owens Corning SMC)
A. Single-Scarf Joining (Owens Corning SMC)

In this particular study, each 2" x 6" (50mm x 150mm) specimen was cut into
two halves and a single-scarfing angle was created along both edges of the cutting line.
The standard repair technique was performed and the results are illustrated in Figure
5.16.

The. results show that both the 5° and 10° single scarfs are superior to the 15°
single scarf. This would verify that increasing the bondable surface area between the

reinforcing patch and the repair zone increases the flexural strength of the repair.

80

J

.eaoz Ego ”55o 223x 25 2.. 5 ease as»: Sam 28. 3:5 9% £58 ”530 2 .n 2&5

:35 mum Bo:

 

 

 

no and $3 «.6 _
mo<z<w oz m . o
mud: oz 4
+8.2: 23¢ a ooF
8.3.... 3.... o «
oudc mum: x
5.2. “3.5 c 4 « oou 5
a
_.N_.._
oon 9
l
c H
x
n m m w
. 1000 (
a a a
o
X 100@

 

 

 

con

81

diam Eon—Em mega
aha—m he 8: 05 255—3 23 5.3 $552. room 035m .5 8:02.?— 05 won—58:— .UZM 9550 ”530 2 .n 2:»:—

Amoeoos 5oz< am<om
oo m. 2 o o

 

Hoop

4

 

 

 

o .
m m room nib
m . a
w a m
m oon m
4 c H
..oo¢ )
. W
Goémss uo<z<o oz In Moo
zotomlaoo a room (

aoTaoo o .

aobuaom o .

e “.923 oz 0

 

 

 

con

82

However, the 5° scarf is not superior to 10° scarf. A possible reason for this behavior
is that the 5° scarf removes more material around the joining area than the 10° scarf, and
the removal of more material creates a larger damaged zone. In addition, it should be
noted that the variance in the repair strengths among the cases is not very significant.
In view of the effort involved in specimen preparation, it was determined that the butt-
joint proved more effective and efficient for thin composite panel repair.

During the analysis of the single scarf-joint, notations of loading either from the
top or the bottom of the specimens are included in the diagram. The notations refer to
failure loading exerted on the top (scarfed) side of the panel or bottom (non-scarfed) side
of the panel. The top side of the joint experiences compression during loading while the
bottom side experiences tension. The strengths from both types of loading do not reveal
a significant difference, which implies that the scarf repair technique is valid for repair
on panels which may experience both tension and compression. Again, note that
regardless of the scarfing angle, 100% flexural strength can be restored in the specimens

by simply utilizing a butt-joint with two layers of reinforcing patch material per side.

B. Scarf-Hole Study (Owens Corning SMC)

An analysis was also conducted on the Owens Corning SMC material to determine
if scarfing the area around a 1/2" (13mm) drilled hole or circular void would increase
the strength of the repaired joint. The scarfed region was created by grinding a 5° bevel
radially outward from the void. This bevel encompassed a 2" (50mm) radial region

outward from the center of the void. Initially, only one side of the damage was scarfed,

83
and then both sides were scarfed. This simulated a single and double scarf-joint

respectively. All scarfing was compared to both an undamaged 4" x 6" (100mm x
150mm) reference specimen group, and another set of 4" x 6" (100mm x50mm)
reference specimen with a 1/2" (13mm) drilled hole in the center. The results are
illustrated in Figure 5.17. They reveal that regardless of a single or double scarf joint
repair technique, at least on this particular material, the flexural strength integrity can
be restored by using the standard repair technique discussed initially in section 2.6. In
fact, the repair strengths are much greater than the undamaged specimen strengths.
Again this is primarily because of the high rigidity and strength of the glass reinforcing
patch material. Note also the reference lines that are drawn from the left edge to the
right edge of the plot. The lower line represents the average reference undamaged
specimen strength, while the upper line represents the average reference undamaged
specimen strength after two layers of glass reinforcing material have been added to both
sides. This upper line illustrates the patch strength influence during non-separated

specimen repair, and is discussed further in Chapter 6.

5.3 In-Plane Configuration Analysis
A. Bond Line Angle

A series of tests were conducted to evaluate the effects of altering the bond line
of the separated joint repair specimens. By altering the bond line a correlation was
developed to determine the effects of the direction of bending force with respect to the

bond line configuration. In this study, all of the joints were separated and the two

 

one: 265%
5;» .26. .25 5.; a 3:38 22.8 o5 29% 53 2.58:. 3 33m came-8 62m «558 22.6 : .m 2:3".

£5 3 u 0383 n 29.3 Ho... am<om

N F o

_ _
fioégs uo<z<o oz ..I o

Goégs «Zoom :83 «\mogo oz .. -
_ uo<z<o oz 0 ooF
m M Box ..~\. a

Saab”. So: a} x

 

 

 

oom .2...

no

m.

oon 9

l.—

H

oo... )

W

a. w

x ooo (

..................... m -.:...:........-.:-......-..-..

ooo

 

 

can.

85

mating parts of the specimens were joined in a butt-joint type configuration. The initial
study consisted of an analysis of bond lines measuring 0°, 30°, 45°, 60°, and 90° to the
direction of the loading pin.
(a) Owens Corning SMC

The results for Owens Corning are illustrated in Figure 5.18. The plots show that
at an off—axis angle of 30° the strength of the repaired specimens is restored. In fact, an
off-axis angle of about 15° would probably restore 100% of the flexural strength. Note
the high repair strengths of the 60° and 90° bond lines. They are approximately the
same, though ideally the 90° bond line should be superior since this bond line is
complemented by the 6" (150mm) length of the specimen plus the double lap-joint repair.
Again, the straight lines depicted in subsequent figures illustrate both the average
undamaged strength (solid line) and the averaged undamaged strength (dashed line) with
the application of two layers of reinforcing material per side.
(b) Excel 8610

The same testing outlined above was conducted for both of the Excel materials.
The results for 8610 are illustrated in Figure 5.19. For the off-axis analysis the results
followed a trend similar to that exhibited by the Owens Corning SMC. Again, 30°
restored the strength of the specimens, as well it appears that 15° would again satisfy the
strength requirement. A different trend is exhibited though for the 45 ° angle. There
does not seem to be any increase in strength from the 30° angle. This change in trend
may be a result of the different fiber geometry between Owens Corning SMC and Excel

8610. Another trend to highlight is the increase in flexural strength between 60° and 90°.

6855 a? .83. 235m .33 co 8825
05 9 25 zoom 2: 9.5632 .6 33cm 05 233.: 3 850886 .52 8295 62m 9.250 M2355 8 .n “coma

A8883 m..oz< Emzomo 5.2.

 

 

 

 

 

 

oo on oo ow on 9 o
r . _ F p _ . L _ . r r . p _ . p O
Aoo.oo.m¢.on.ov $2.2. x
@353 uo<z<o oz I. r
uo<z<o oz 0 Goo.
Goém>$ manna zofia «\uogo oz .1. .
oom :81
w H
m
oon
a. m
AW H
x 1:00.? )
x W
s d
H 1100m MW
..-.............:.-.:w .............................. 1ooo
x ooh

 

 

zgaaoaazzgeoagsm.gfifiz
8825 as s a: 28 05 25?: 3 seem 05 25...... 9 850580 .52 ”335 ”28 .oam 2 .m 2&5

A8833 u..oz< Emzomo 5.2,
oo 2. oo m... on m, o

 

 

 

h b h p b b b I - Bl — P n b F —
~ Aoododtoné mics x o
Gosmss mafia :05. «$228 oz .. ..
‘ Goémzo mo§<o oz ..I oo.
uo<z<o oz 0
oom B
on no
m.
oon
w, . m
H
x x noow )
om . M
.............. fl.......:......:.........:..-.........i..............oom mx
floom
1

 

 

 

005

88

It is approximately 84%. This trend supports the idea that in repair, care must be
exercised to not build in a repair strength that is too high since this can shift the damage
into a different area.

The shifting of damage to another area is a topic of concern. Throughout all
testing, remarks were documented regarding the failure modes, and the amount of flexure
in each specimen at failure. The failure modes are listed as either catastrophic (brittle)
or non-catastrophic (ductile). There seems to be a direct correlation between the amount
of flexure in a specimen at failure and the failure mode. If the specimen fails in a non-
catastrophic mode then the load being applied to fracture the specimen is being
distributed further into the specimen, and its flexural value increases. However, it
should be noted that this higher strength value could induce excessive loads elsewhere
in the composite panel. Further comments regarding the failure modes are discussed in
the appendix of this document.

(c) Excel 2415-Warp

Results for Excel 2415 reveal a couple of different trends. The results are
illustrated in Figure 5.20. The bond line contour survey illustrates the same trends as
shown by the other two materials, with the exception of the comparison of fiber warp
direction to fiber fill direction. Figure 5.20 reveals that the strength of the fiber warp
direction is restored at the 60° mark, as well, it indicates the incapability in restoring the
strengths by using the reinforcing patch material in the cross-woven 2415 material.
Consequently, the repair strength does not surpass the original undamaged reference

strength of the composite until a bond line angle of approximately 50° is achieved. A

89

5835 a? one: 285m .33 .o
8:025 as s 2.: 2.8 as 2582 s 385 2.. 223...: s 85288 .52 32.8 "a: 32m own 23E

@283 5oz< Emzouo 5.2.

_ _ —

om mu. 0% we on 0.. o

 

mm<3|mo<z<o oz 0
Gosmss amssnuogo oz .l

Gododtoné €22. x

0953 «Emma zoos. «\mméuuogo oz .. ..

30%
X%

 

 

100p

 

j

CON

con

00¢

com

1000

1005

 

 

00w

(de) HloNz-Ials

90

possible solution to this incapability is to repair the 2415 warp direction specimens with
a sample of the Excel cross-woven mat. However, since the reinforcing material is very

rigid, it will probably delaminate during failure rather than add more repair strength.

B. Bond Line Pattern

The secondary testing involved looking into the idea of cutting a known in-plane
contour into each of the specimens. The contours were in the shape of a V-,U-,W—,UU-,
WW- and S-joint. This investigation proved to be a very important.

The following sections present the results of the joint geometry study for all three
different fiber geometries. All specimen were repaired following the standard
procedures. The Owens Corning composite utilized Maraset as an adhesive, while the
Excel materials used Resin Services resin.

(a) Owens Corning SMC

Since the 60° and 90° bond lines demonstrated such superior results (100%
increased strength), it was determined that subsequent bond line contours would
incorporate bond line angles from 60° to 90°. The resulting contours were symmetric and
based on the form of the V-joint. The V-joint was cut into a 2" x 2" .(SOmm x 50mm)
central area on the repair specimen. The slope of each side of the V was measured to
be approximately 60°. The U-joint was designed to investigate if rounding off the
comers of the V-joint, would reduce the stress concentrations at the bond line. The
results show that indeed this was the case. The strength values, illustrated in Figure

5.21, were predominantly higher for both the U- and the UU- joints. Results also show

91

 

 

 

.2.“ 2:2". 5 28m £83. 2.. §e§8s 523 ”iii .53.; o5 -> 65 Esau 2.95 a.“ 83E
zoFéaoEzoo .55.;
25> m 33.3 :.>
b — P C
._.z.o?.m o
$3.35 a
9.5.an D 3.00?
”3.33 02 o
fioégs «Zoom :23 «\uogo oz I. woow 8
II.
can mm.
N
c m
roo¢ H
0
\w/
o x x com moo
.................... dill-......x.....:............:..ooo (
m m ..oo\.
m

 

 

 

 

92

that regardless of a V-, W-, or WW-joint, at least over a 2" x 2" (50mm x 50mm) area,
the strength values are approximately the same.

An additional contour was also evaluated under the title of S-Joint. This set of
testing was designed to try to find an optimum bond line joining contour. Further
discussion regarding the S-Joint is postponed until Chapter 7.

(b) Excel 8610

Results for the V-Joint analysis again utilize the optimum angle of 60° in the V-,
U-, and W-joint design. Again the repair results, shown in Figure 5.22 are superior in
restoring flexural integrity, though the results for the V- joint versus the U-joint are not
consistent with those exhibited by the Owens Corning material. However, the UU-joint
surpasses the W-joint like before.

(c) Excel 2415-Warp

The results of the V-joint contour study, shown in Figure 5.23, revealed that the
V- and W- joints are superior to the U- and UU-joints. However, note that all joint
configurations were successful in restoring the flexural strength of the specimens, even

in the warp direction.

C. Bond Line Depth (Excel 2415-Warp)

The results of the joint geometry analysis provoked inquiry into a correlation
between the in-plane depth (in the specimen’s length direction) of the joint and its
corresponding strength. The material used for this particular survey was Excel 2415 in

the warp direction. This investigation was designed to determine an optimum depth of

93

a; as»: 5 28m 9.33. 2.. 8.20985 523 iii .53.? as -> 5.8 .8xm «an 23E

20:.(«30E200 5.2.03
>>>> 33.3 :.>

 

- _
$3.35 Ezomo 5.2.
25.3 Ezouo 5.2. a
“5523 oz
396% manna :23 «$323 oz .. ..

O

 

 

loop

 

CON

com

I
O
0
d.
(Daw) HloNaals

I

1
O
0
ID

loom

 

 

con

94

.2 .n 2:5 5 28m 5.33 2.. 8.5985 523 ”2.22 .52.? o5 -> "a: .85 an 23:.

ZOEEDOEZOQ .55..
25> sacs qfl>

 

 

 

L . -
33.3.5 Ezouo 5.9.
32.3 Ezouo 5.2. 4
afiznuo<z<o oz o

fioém>$ «25mm Iota «\améuuogo oz .. ..

100p

ICON

icon

 

00+

com

1000

100x.

 

 

com

(0cm) HloNz-ials

95

joint, from correlating the results of the 60°-90° bond line investigation. The two
different contours that were evaluated were the V- and W-joints. The V-joint was
designed over the following specimen areas: (1) 4" x 2" (100mm x 50mm), (2) 3" x 2"
(75mm x 50mm), (3) 2" x 2"(50mm x 50mm), and (4) 1" x 2" (12.6mm x 50mm). The
depths of each joint are listed first, therefore depths of 4"(100mm), 3"(75mm),
2“(50mm), and 1" (25mm) were analyzed. The depth design of 4"( 100mm) incorporated
a 76° bond line, while the 3"(75mm), 2"(50mm) and 1" (25mm) depth designs
incorporated a 72°, 63° and 45° bond line, respectively. Plots illustrating the results of
this analysis are shown in Figures 5.24-5.25. In addition, an analysis was conducted on
a set of V-joint specimen with a 1" (25mm) joint depth. In this study, the edges of the
V-joint were slightly searfed to check if scarfing would reduce the stress concentration
along the bond line, and possibly shift the failure mode, or increase the strength of the
joint. Figure 5.24 reveals the results of both the V-joint and the V(scarl)-joint. The
results illustrate that the flexural strength of the joint was not substantially increased by
utilizing the scarfing technique. Further results show that for the 4" ( 100mm) depth
design the flexural strength was increased by approximately 50% . This strength result
was similar to the 75° bond line result from the preliminary bond line orientation study
on Excel 2415. The steep bond line allowed the reinforcing patch material to utilize a
larger bondable zone as well as more of the parent material. In the shallower joints, e.g.
1" (25mm) depth, the parent material cannot help out as much under bending.
Therefore, it was concluded from the results that, any bond line depth equal to or greater

than 2" (50mm) is capable of restoring 100% flexural strength in the repaired specimen.

96

33:5 33 .3 395m .835 as
”soon 2.: Eco 220:5 23:5 52". 8:33.60 593m 2.. 9935 9 6222 26295.3 "23 scum a.“ as»:

AEEV :Euo pz_o_r>
ooF on on no
_

_ _
8m x as
an x o3
8m x at
8... x 85
Aomxnuvamém
am§iwo<z<o oz

X

 

100..

DO!!§‘

ICON

W

T fl
0
s;

(de) HloNz-ials

 

O
O
V.

 

coo

* X
Xfi
%
W]
O
0
ID

 

con

97

6355 83 is. .523 .892 .5
”sea 2.: 95o .5.qu 255: 52m 8:35.50 593% 2.. 35...: 9 ”i=2 .52..ng ”a: .83 a.“ 23E

AEEV :Euo 56?;

 

on ma
. _ 0
an x as a
an x o3 a .
@933 «39. .62.. «\dcuuogo oz I. too.

cogs manna zoo». «\aménmogo oz !.
afiaumoéé oz o

 

 

 

dcumoséo oz a room m
j 3
woon w
uln—
H
oo... \1
W
x d
m oom wt
3
room

 

 

 

DON.

98
The W—joint design was developed for only two different depths; 2"(50mm) and

1" (25mm). The 2" (50mm) depth utilized a 76° bond line, while the 1" (25mm) depth
utilized 72°. The results are plotted in Figure 5.25, and they reveal that regardless of
a depth of 2"(50mm) or 1"(25mm), the flexural strength can be restored. It should be
noted that the angle for both of the W-joints exceeded 70° . Consequently, it was
determined that the bond line angle is the dominant factor in restoration of separated joint

strength and also in bond line design [101].

5 .4 Out-Of-Plane Curvatures

The material used for this evaluation was an SMC made up of 28 % random
chopped glass in a polyester matrix. Six different Curvatures were evaluated throughout
the structure. Each curvature was cut to the dimensions of 2" x 6" (50mm x 150mm)
and loaded to failure under three-point bending. After damage, the specimen were
repaired, using a technique similar to that outlined in section 2.6, and three-point loaded
again to failure. Again, the parameter under differential comparison was the load to
failure. The results illustrated in Figure 5 .26 reveal that the standard repair technique
utilized for flat panel repair is also superior for curved panel repair. It is worth noting
that during the repair process there was no pressure exerted along the bond line during
the curing cycle of the epoxy. This did not seem to complicate matters. To make sure
there was no problem with increased voids, an additional test was conducted to repair a
flat panel of the same material without the application of bond line pressure during the

resin cure cycle. This value is illustrated in Figure 5.26 as the reference repaired

.3.an 02m =23?”— éaoefiz 9.35m i.e.—.02.: 85. :25. .05; 3:5 z. £32. on.“ 2:2".

 

 

 

mmmofiioo
om . mm 5.. mm mm E om
r — — b — b O
.
m hoop
C m 100w
f
. i
o m woo...
x x x j
x oom
x a
ooo
Suzi uo<z<o oz 0 1oox.
fluzé ommzfix u .

 

 

 

com

(de) HIONaals

100

specimen and demonstrates the same repair strength value as the specimens with applied
bond line pressure. There are actually six marks revealed on the plot, three for

pressurized, and three for non-pressurized.

CHAPTER 6

ASSOCIATED PARAMETERS

The purpose of this particular chapter is to completely address some of the
associated parameters regarding composite repair, for example, selection of adhesives,
the determination of the number of reinforcing patch layers, and the evaluation of

different types of loading.

6.1 Adhesive Study (Owens Corning SMC)

The first parameter to evaluate was the choice of adhesive to be used, and the
effects of the adhesive on the joining strength. The material that was used exclusively
for this investigation was the Owens Corning SMC. The reason for using the SMC
material was its random fiber geometry, which is not influenced as much by loading
direction. Also, the SMC material had displayed superior repair results from utilizing
the Maraset epoxy. However, the Maraset epoxy was abandoned for subsequent testing
on the Excel material due to compatibility problems.

A. Evaluation of Adhesives

The results of this study are illustrated in Figure 6.1. All joints were 90° butt-
joints utilizing two layers of reinforcing glass material per side. In this study, six
different resin systems were investigated. All were two-part, room-temperature curing
systems. The different adhesive systems are outlined as follows.

(1) Fusor - This is manufactured by Lord Industries and is designated as a structural

101

102

6258.50 com—«Eon
he :83— EooSam 05 co 589$ 02852 05 we 8:26:— 05 Baa—gm 9 him 9:83.; .038 9.550 «530 _.e Boar.—

 

 

 

m2Em>m 2.8m
2 NB mom mm in «03...
b _ — p — C
n. .Woop
. S
m m m m
Lu 0 a room m
w a 9
mm m n. M
a i m oon \/
m a M
f D
(
o
-oow
uo<z<o oz 0 T
2% 2.8m a

 

 

 

com

103

adhesive. It is very viscous (thixsotropic) and does not saturate the patch material
uniformly.
(2) 3M - This is another structural adhesive and responded quite similarly to the Fusor
material. The high viscosity system did not permeate the patch material completely,
consequently creating an adhesive pocket at the bond line.
(3) Resin Services - This particular epoxy resin system showed superior bonding
strength and did not demonstrate any delamination between the patch material and the
adherend. The viscosity of this system was very low and seemed to fully saturate the
patch material during repair. Therefore, this particular resin system became the adhesive
of choice. '
(4) TCC205 - This is another low viscosity epoxy resin system that turned in superior
results during the study. There was no delamination of the patch material from the
adherend and the bonding strength was very close to that of the Resin Services adhesive.
(5) TCC076 - This particular material is a low-viscosity polyester. It saturated the patch
material well but displayed some delamination during testing. The results were
consequently very poor.
(6) Maraset - Since Maraset was used exclusively for the Owens Corning material, a
comparison to the Resin Services epoxy became pertinent. The results indicated that
indeed, the Maraset and Resin Services resin are complementary to the repair of the
Owens Corning SMC material.

The adhesive study was conducted to verify the present resin systems being used.

During the preparation of this particular study, ten different adhesive coupons were

104

developed. Each coupon was cut into three specimens and characterized for their
flexural strengths. The results, illustrated in Figure 6.2, verify once again that the Resin
Services system displays superior flexural integrity and rigidity. The modes of failure
were also reviewed and both the vinylester and polyester samples proved to be much
more flexible than the rigid epoxy. The amount of adhesive rigidity is a very critical
parameter in composite repair. In general, Figure 6.2 illustrates that the more rigid the
adhesive, the higher the strength of the joint will be. However, there will be a high
stress concentration along the bond line. Therefore, it is advantageous to utilize an
adhesive system that responds very similarly to the original matrix of the composite in
order to mimic the flexural response and lower the stress concentrations along the bond
line.
B. High-Viscosity Adhesives

The aforementioned results imply that the high viscosity of the Fusor and 3M
material may be superior for perforation damage repair and cosmetic type repair. It is
the objective of this particular study to further evaluate their applications. Several 4" x
6" (100mm x 150mm) panels were impacted to the point of perforation. Then each
specimen was trimmed to the standard testing dimensions of 2" x 6" (50mm x 150mm).
After trimming, much of the damaged zone material was pushed back into the void
region and utilized as filler while the rest was sanded off using the die grinder. The
repair procedure was a bit different for this survey. Instead of repairing with epoxy and
two layers of reinforcing patch material, just resin was placed into the damaged zone of

each sample. Figure 6.3 illustrates the results. Fusor, not the 3M material, displayed

.2880 02.23 8. 35.3. "52.3 .=_£-§fi 3 25E

m2mem>m 25mm

300 2 who now mm Sn moms.”—

b b L b P

 

105
Elm
[I]

DO

 

100—
IONF
TO¢P
romp
iomp

(0cm) HloNaals

 

 

CON

106

duos—on .895 Baotou we :83. 9.55
629$ 02853 co .:5 «55:5 we 38cm 05 3.53:— 3 Exam o>ao€<385 .055 9.530 2.036 no 23E

 

 

 

 

mzmfifi 2.3m
goo 2 «so now mm 2n moms...
_ _ L P r h . 0
oo.
8
I.-
m m m w.
9
Il—
m u H
oon m
d
D
(
oo...
mo§<o oz 0
ouciuiofioéz a

 

 

- com

107

superior results. However, an interesting observation revealed that the Resin Services
system also displayed complementary results. This again is primarily due to the superior
bondability of the Resin Services system. The line of failure cut through the center of
the void, while the more thixsotropic adhesives debonded from the edge of the void. The
major findings of this study were that the strengths of the repaired panels were not
disturbed regardless of voids in the repair region. Similar results took place in Fusor,
that is, the Fusor material enhanced the aesthetic repair of the small holes or indented

regions at least up to 80% of the composite’s original strength.

6.2 Reinforcing Patch Study
A. Number of Reinforcing Layers for Butt-Joint

The main reasons for utilizing a plane-woven glass patch for reinforcement was
because of its ease in handling and effective repair strength for all of the different fiber
geometries. In addition, compatibility of material and consistency of fiber direction are
two factors which must not be forgotten during the development of an optimal patch
design.

The influence of each glass reinforcing patch on the ultimate strength of butt-joint
repair in Owens Corning SMC is illustrated in Figures 6.4-6.5. In each case there are
two layers of glass patch material on the top (compressive) side and one to four layers
of patch on the bottom (tension) side of the butt-j oint specimen. Note that the average
undamaged specimen strength is 290Mpa. Also note that two and three layers of glass

patch restore approximately 100% of the undamaged specimen strength, and four layers

108

58283.52 62 age 2.. co .385 05 Co 83.5 65 £58 .830

A323 mama... Iota

to 25E

 

 

 

v n N F o
r [P F L o
oo.
m
m oou
u
«— 0 con
u 0 Am
1oo+
u 1
6235 uo<z<o oz I. woom
93 z. 2.8m o .
56 2. dog a room
58 oz a
mo<z<o oz o .

 

 

can.

(can) HroNaals

109

228055 ”85.3.: fie... 05 as Bozo 2.. ”€55 5&3. fir. 2335 02m ”568 2.25 no 2:5

323 mamas: Iota

 

 

 

 

e n a P o
F — _ o
4 a
« « oo.
m
u m
m oow co
m m.
m 9
n u oon m
my D _U u \l
c M
.8... o
(

H
u
U

Qoémss uo<z<o oz

mmuzxo..¢ madam oz_N_.Eo whammy.
mmmzxog ._<z_o_mo oz.~_.__S gamma
uo<z<o oz

1
O
0
ID

I

004'

 

 

 

coo

110

far excwds the 100% restored strength and could therefore cause failure in an
undamaged portion of the specimen. These superior repair strengths verify that the
plane-woven geometry of the glass reinforcing patch material maintains a higher flexural
strength than the random chopped glass composite itself. An additional comment should
be made regarding the results in Figure 6.5. Note the repair strength of the butt-joints
that utilized the repair thickness dimension rather than the original thickness dimension
during the calculation of stress. It was determined that the repair thickness dimension
did not demonstrate the significance of the relationship between increased load to failure
due to increased patch layers. Therefore, the repair thickness dimension was not used
for all subsequent calculations of repair strength.

Additional testing was conducted on the Excel materials. The results for Excel
8610, shown in Figure 6.6, reveal again that only two layers of reinforcing patch
material are necessary for complete restoration of flexural strength. Again, the Excel
8610 material is a continuous swirl mat, which is not as flexurally stiff as the plane-
woven glass repair patches.

The results for Excel 2415, illustrated in Figure 6.7, reveal the need for three (fill
direction) to four (warp direction) patch layers for complete repair. The increase in
patch layers for the 2415 material is primarily a result of the fiber geometry of the
specimen. The 2415 consists of a cross-woven fiber geometry, with a weave thickness
of approximately 1mm, while the reinforcing patch material is a plane weave with a
weave thickness on the order of 0.2mm. This difference in thickness implies that the

strength of the 2415 fibers is much greater than the patch material. Consequently, more

111

45588:? as». 35 2.. 3 593m 2.. ac Soggm ”28 32m 3 as»:
3.5 mama: :25

 

 

 

m. ¢ n a F o
_ b F . _ o
co.
oom
x
x can
x .
m -oow
x x .
x
uuuuuuulw uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu icon
6395 mo§<o oz I. .
6353 «Iowa :05. «\uogo oz .. room
8523 oz 0
5.2. Ea u 1

 

 

COB

(de) HiONaals

112

.Eesécsm =28 ago as .6 £55m 2.. a 833m «so $6
323 392.. 5:3

n v n N P

he Bani

o

 

- p _ _
Amoégsaméumzaum 52a «\uogo oz
Gags amssuuogo oz

afizuuogo oz
x dcuuo<z<o oz
.3 denommzfim
améuoumzfim

<4 *¥%

 

taco]:

 

 

c

1

floor

ICON

'

 

con

00¢

 

 

(de) HlONaals

113
patch layers are needed to restore the desired strength. Figure 6.7 verifies this

conclusion.
B. Number of Reinforcing Layers for Bending Fracture

This particular study was introduced to evaluate the strength contribution of the
glass patch layers in non-separated specimen repair. Specimens were loaded to the point
of three-point bending fracture and then repaired with two, three, four, and five
reinforcing patch layers on the bottom (tension) side of the joint. Two layers of patch
were used on the top side of the joint, since two patches seemed to be able to sustain the
top (compressive) load induced on the specimens. The results are illustrated in Figures
6. 8 and 6.9. Results revealed that two patch layers restored the flexural strength for the
fill direction, while three layers were necessary to restore the strength in the warp
direction. The results were also compared to the butt-joint layer survey in Figures 6.4-
6.7. The conclusions were virtually the same, i.e., in general, a separated joint requires
more layers of reinforcing patch material than a non-separated joint.

Comparative results for this survey are illustrated in Figures 6.4-6.7, and are
highlighted as follows. For non-separated repair, the increase in strength values
compared to the original undamaged strengths are: (l) Owens Corning (64%), (2) Excel
8610 (52%), (3) Excel 2415 - warp (41%), and (4) Excel 2415 - fill (40%). These
results again verify that the random fiber and continuous mat swirl fibers can be repaired
quite efficiently with the plane-woven glass reinforcing patches, while the Excel 2415

material requires a bit more strength in the patch material to complement its own fiber

geometry.

 

114

eEEZE§B?§2.
.52. 3888-52 a S Swap—om no coup—38¢ 05 833m 2 :55 8.53 838.."— wcmvcom ”Bea 63m

323 mama: 52a

 

 

 

 

 

 

e n a P o
p b b — O
cum—(mum a .
83:32: uozumubm a
cogs afizuuogo oz .I woo?
afisuuogo oz o 4
0953 «Zoom 10?. «\afizuuogo oz - .. room
.
¥ 1.000
x ..
w x oow
4 fi 4
. oom.
w” m
.. o
my ....... up .................................. 44 oo
1oon
L7

 

 

com

woeaa

(Paw) HloNaals

115

sis. Basso 55.
.52. 3888.52 a 3 59.95 mo ecu—23'3— ofi 3.3—gm 9 33m Ban..— 2325. «£38 53w «Bum o6 2:me

323 $83 5.5.

 

 

 

 

 

m . e n a P o
— b — p o
Barium a
89332: uozumuaum a .
uo<z<o oz o 18.
93195 «the :05. whose oz .. ..
m 0352: 8.33 oz .l
oou m...
a m
w a o1 luv oon w
. m.
x x 18¢ )
. W
x d
........ x...!......:.!:.l-............-....-..---............j..oom mx
. a
x 18m.

 

 

can.

1 16
C. Woven-Patch Orientation Study (Excel 8610)

A plane—woven glass fabric was utilized during repair to supplement the loss of
fibers during failure of the specimen, as well as to add flexural rigidity. However, it
should be noted that the plane-woven patch material has two dominant strength
orientations namely the 0° and 90° directions. As stated in Chapter 2, it is very
important to identify the fiber direction in the material that is under repair and
supplement the strength of that direction with the strength from the reinforcing patch
material. However, the misalignment of the woven patch with respect to the fiber
orientation of the damaged material is also a concern. The results of this investigation
are in Figure 6.10.

The patch orientation study was conducted utilizing the Excel 8610 material. The
repair procedure was consistent with that which was described previously. The only
repair modification was the patch orientation angle. Two layers of glass patch were
incorporated per side. The first patch orientation angle was (0° ,90°). This was
considered the reference angle. The second orientation was (+45°-45°). Note the
reduction of flexural strength. The third ease exhibited an orientation on the first layer
of (0°,90° , while the second layer was (+45°/-45°). Subsequent orientations were,
(30°,-60°) and (15°,-75°). The results were very interesting since they revealed that only
for the orientation angle of 45° was there any appreciable reduction in flexural strength.
It was therefore determined that the orthotropic behavior of the glass reinforcing material
is not a critical factor if its orientation angles stays within 45° of the damaged material’s

fiber direction. In fact, if an isotropic orientation is required, then a combination

87;.“ 3

seem 838.5 53. 8833.. "28 .85. 2 .o 23E

@8303 u..oz< zopfizymo 5.2a

GT\+V
:5on EMS An¢u\+v 8&8

 

117

 

zo_._.<._.zm_mo Io._.<n_ fl
mega Oz 0

 

o
100'
5
com .\d
3
N
9
oon m
a .\.l
T m
00¢ .u
T (
icon
a

 

 

cow

118
utilizing (O°,90°) and (+45°/—45°) layers can be installed.

6.3 Aesthetic Repair

Aesthetic repair was discussed briefly in Chapter 1. To reiterate, aesthetic repair
is a form of non-structural repair and lends itself to a lot of artistic freedom. The study
outlined in section 6.1B describes the aesthetic repair scheme utilized to restore strength
in impacted panels. Again, Figure 6.3 revealed that a thixsotropic structural epoxy
system should be the adhesive of choice for aesthetic repair. As a note, the thixsotropic
adhesive resembles the material referred to as "bondo" . To date, bondo containing
random chopped glass fibers can be purchased at any auto-body repair shop. The study

outlined in section 6.18 supports the use of a product such as this.

6.4 Types of Loading
The major types of loading that were discussed and evaluated throughout Chapter

5 were three-point bending. This section is focused on discussing the additional loading
types, i.e. , tension, compression, torsion, and cyclic fatigue, that were used to
completely characterize each fiber geometry. Throughout a majority of the testing, the
specimen dimensions of 2" x 6" (50mm x 150mm) were adhered to in order to alleviate
the complication of any other testing parameters. This also helped with keeping in a

differential and comparative mode between testing techniques.

A. Tension and Compression

119

The evaluation of tension and compression utilized the servo-hydraulic Instron
machine and was conducted under stroke control. Readings of stress and load were
monitored simultaneously during compressive displacement. The specimen dimensions
for tensile testing were 2" x 6" (50mm x 150mm) (gage length = 2.75", 70mm) while
the compression specimen were cut to 2" x 3" (50mm x 75mm) (gage length = 1.5",
40mm), this dimension was also utilized to instigate compressive failure while preventing
buckling.

The determination of Young’s modulus and Poisson’s ratio were evaluated via
tensile testing of each material with the application of bi-directional strain rosettes. '
Large strain gage areas were utilized to create more of an averaging effect of axial and
transverse strain. This helped to alleviate the possibility of an extraneous strain readings
due to the placement of the gage on either a resin rich or fiber rich zone. Table 6.1
contains all of the characteristic material properties, such as tensile and compressive
moduli, tensile and compressive strengths, and Poisson’s ratio, for all three composite

fiber geometries.

B. Critical Buckling

The evaluation of buckling was also conducted on the Instron machine and was
monitored under strain control. This required the use of an extensometer. The
extensometer was modified to incorporate a 50mm gage length and was fitted with 50mm
knife edges on both ends. This special modification is of particular interest when

analyzing strains in composite materials. The larger gage length and knife edges produce

Table 6.1

120

Listing of the Composite Material’s Characteristic Properties

E= YOUNG' S MODULUS, E‘= COMPRESSIVE MODULUS,
T=ENSILE STRENGTH, a = COMPRESSIVE STRENGTH
0 = POISSON S RATIO
0- = REFERENCE UNDAMAGED STRENGTH
a = REFERENCE UNDAMAGED STRENGTH W/2 LAYERS OF
GLASS REINFORCING MATERIAL (with resin) PER SIDE
P = PATCH STRENGTH INFLUENCE

OWENS CORNING:
E = l.8Mpsi,E,= 2.0Mpsi
a,= 19 ksi, a¢= 24.2 ksi
t9 = 0.3
a = 42 ksi
ap= 69 ksi, P = 27 ksi

EXCEL 8610:
E = l.56Mpsi,E,= 1.63 Mpsi
,= 23 ksi, a,= 26 ksi
r3 = 0.24
a = 44 ksi
ap= 66 ksi, P = 23 ksi

EXCE 2415-WARP:
E = 2.71Mpsi,Ec= 2.15Mpsi
,= 26 ksi, a,= 29 ksi
0 = 0.25
a = 69 ksi
a,= 98 ksi, P = 29 ksi

EXCEL 2415-FILL:
E = 2.26Mpsi,E,= 1.64Mpsi
,= 35 ksi, ac= 34.2 ksi
0= 0.15.

:P=76ksi,P=22ksi

121

an averaging effect of the strain over the surface of the specimen. This is very critical
for composite materials for the simple reason that it allows for more of a ”whole-field"
approach to strain mapping rather than ”point-by-point" . The whole field approach for
composite material evaluation helps to prevent the positioning of the extensometer
directly atop a fiber, or resin rich zone. The averaging of strains gives a much more
accurate strain mapping of the surface of the composite undergoing deformation.
Although this testing was conducted under strain control, stress versus strain was
monitored and recorded simultaneously.

Critical buckling analysis was conducted for all three of the different fiber
geometries. The theoretical analysis for buckling utilized Euler’s buckling formula in

the form:

Pm =itrJiEI
L2

where: E = Young’s modulus
I = second moment of area

16 = specimen’s effective length

Even though Euler’s formula is designed for isotropic, homogeneous material analysis,
the theoretical calculations were within 20% of the experimental values. Figures 6.11-
6.12 illustrate the buckling phenomena during loading while Tables 6.2 and 6.3 list the

theoretical versus experimental values for P“.

122

CRITICAL BUCKLING LUAD HR WENS CURNIMS

 

m: m]
w j

CRITICN. BLCKLIM L041} FUR EXCEL 8610
137: "Per

.. /"2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 6.11 Buckling Phenomenon for Owens Coming and Excel 8610.

123

CRITICAL BLCKLING LEN] EUR EXCEL 8415 FILL

 

 

m: 1P0!

 

 

 

 

/

3'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CRITICN. BLCKLILG LMD FDR EXCEL E415 UMP

A

ni' ml

.. ,

Figure 6.12 Buckling Phenomenon for Excel 2415.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

124

 

 

 

 

 

 

Table 6.2 Theoretieal vs Experimental Load Values for Critical Buckling.
MATERIAL LENGTH WIDTH THICK MODULUS CRITICAL CRITICAL
L. (m) (In) (In) (Mpsi) LOAD LOAD
PM P...
(lbs) (lbs)
OWENS 1.30 1.931 0.104 1.80 2120 2114
CORNING
EXCEL 1.40 1.965 0.123 1.56 2040 2501
8610
EXCEL 1.38 1.923 0.113 2.71 2280 2597
2415
WARP
EXCEL 1.3 1.918 0.114 2.26 2024 2264
2415 .
FILL
——_ j _
1W
OWENS CORNING: 0.3%
EXCEL 8610: 18%
EXCEL 241m: 12%
EXCEL 2415F: 11%

 

 

 

 

 

 

 

 

 

Table 6.3

125

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Experimental Critical Buckling Load Data for the S-, and Butt-Joints.
SPECIMEN AREA MODULUS CRITICAL CRITICAL COMMENTS
ID (sq.in) (Mpsi) STRESS LOAD
(kpsi) (lbs)

OCND-15 .228 1.80 7.79 1760 NO REPAIR
OCBJ-6 .237 1.80 12.52 2980 BUTT-JOINT
OCSJ-6 .242 1.80 19.28 4600 S-JOIN'I‘
86ND-15 .244 1.56 +7.80 1752 NO REPAIR
8631-9 .250 1.56 17.39 4200 BUTT-10M
8681-14 .229 1.56 15.53 3480 S-IOINT

24FND-15 .223 2.26 7.44 1664 NO REPAIR

24FBJ-12 .213 2.26 14.08 2908 BUTT-JOINT
24FSI-14 .216 2.26 15.46 3368 $401th
24WND-15 .220 2.71 9.90 2110 NO REPAIR
24WBJ-9 .209 2.71 14.17 2780 BUTT-JOINT
24WSJ-11 .212 . 2.71 18.65 3990 S-JOINT

 

 

 

 

 

 

 

126
C. Cyclic Fatigue

The evaluation of fatigue began by utilizing the results from the buckling analysis.
Specimen containing the same 2" x 6" (50mm x 150mm) dimensions were utilized. It
was decided that the cyclic fatigue testing would be run in a fully reversed (R=-l)
tension-compression mode, since this mode simulates a worst case scenario for the
loading of structures. The critical buckling load data was used to establish the fatigue
load limits. The tensile and compressive loads were established to be 50% of the critical
buckling load. This particular load value corresponded to a strain value of approximately
0.0002 or .02 % strain.

The series of specimens were analyzed under strain control, therefore an
extensometer was implemented. Again, this extensometer was modified to evaluate
average strains over the surface of the composite. Two strips of double stick tape were
utilized under both of the knife edges to prevent any slipping of the extensometer during
the cycling process.

The idea behind the fatigue testing was to cycle both the undamaged or reference
panels, and the repaired panels for 100, 1000, and 10,000 cyclic intervals. It was
originally planned to run specimens to 100,000 cycles but the facilities and time did not
permit this series of tests to be run. After cycling, the panels were fractured under
three-point bending to see if there was any degradation of flexural strength due to cyclic .
fatigue. The results are illustrated in Figures 6.13-6.16. Note that all four fiber
geometries were evaluated and the parameters being compared were the reference

strength, butt-joint and the S-joint. The butt-joint and S-joint were the chosen repair

127

£05m hon—om v5 8:89.3— mEEoO 2830 3.. 3.80m Sauce 2 .c Bzmfi

mw._o>o mo mumzaz

 

 

ooo.o . ooo. . oo.
w , L h b o
oo.
com mo.
.2 . ._ s.
n <
N
4 a oon 9
. 1
o c H
AV 100.? .\.l
m 0 w W
1. mo
Toom (
55?...“ o T
$57.56 a room
~85... ones a
uo<z<o oz 0 .

 

 

 

COB

128

.288 once as 8883 28 .95 8.. £39— 2.»an : .o 2:2".

mmqo>o LO mumEDz

 

 

 

000.0 — 000. P 00 P
. t 0
Too.
oou B
m m
N
m w con m
e H
0 Too.‘ \W/
o o a .0
Av T000 Inw\
5.91m o a
.PZ.O$I._.—.Dm 4 700m
Ska o3o>oiuo<z<o oz a
85.3 oz 0 .

 

 

 

00m.

129

285.. :83. ea 8:2»? £55: 8am 8. £32. 2&3. a .o 28E

mw1_0>0 LO mwmzaz

 

 

000.0 P 000. P 00 P
p p o
. .8.
room
AV ..
1000
d m m
o . 00¢
w a
1000
5.0..lm o a
E-Ofilgm 4 1000
”30:3. 01.05.00.330 02 n .
Ann—Imago 02 o

 

 

 

 

005

(Daw) HIONaars

130

.225. .33. 8“ 8:283 35628 .85. 8.. £89. 8qu o. .o 28E

mun.0>o “.0 mmmzaz

 

 

ooo.? ooo; o8
b P o
100.
4 room
a. q , -oon
, q .
Av 0
roo¢
B
mm m n. 8....
m ...z.o..1m o
s o 553.? a ooo
moons. ozo>ouuo<z<o oz a
afiznuogo oz 0

 

 

- 005

(de) HLONBELLS

131

configurations from the standpoint that they demonstrate two opposing forms of in-plane

repair. The S-joint is described in complete detail in Chapter 7.

D. Torsion

The torsion analysis was conducted by also utilizing the servo—hydraulic Instron
machine. This testing was conducted in stroke or rotation control. Each specimen was
machined from the dimensions of 2" x 3.5" (50mm x 90mm) into a dogbone
configuration of 1.5” x 2" (40mm x 50mm). This evolved through preliminary testing.
Preliminary observations revealed that the 4" (100mm) gage length of the typieal 2" x
6" (50mm x 150mm) repair specimen could not be used since it did not exhibit failure
in the gage length after a torsional rotation of 50°, which was the rotational limit on the
Instron. During testing,rotation angles in 5° increments were utilized to better reveal the
maximum torsional load to failure. Results for the torsional testing are postponed until
Chapter 7 as well as a further discussion regarding some intriguing developments for

torsional damage repair.

CHAPTER 7

OPTIMIZATION OF CONIPOSITE REPAIR

This research and development of composite repair evolved into a comprehensive
study to evaluate optimum repair designs for various types of composite panel failure.
The following sections outline three special approaches for thin panel composite repair.
The final section discusses the, "optimization of design, " computer model that was

utilized in conjunction with the experimentation.

7 .1 The Strap-Joint

The idea of implementing a strap-joint to repair small circular voids was a novel
approach focused primarily at reducing the degree of catastrophic failure during bending
fracture. The strap-joint procedure consisted of cutting two glass patches approximately
2" x 8" (50mm x 200mm) and knotting one to the other separately which eventually
created two knots at both of their centers. This double knot was then placed into the
void region of the damaged specimen and the tails of the knot were drawn outward along
the top and bottom surfaces of the specimen. The knot stabilized the position of the filler
as well as reduced the stress concentrations in the repair zone. The design demonstrated
a continuous patch from the top of one surface of the joint, through the center of the joint
to the opposing surface, as shown in Figure 2.9e. The results illustrated in Figure 5 .14
verify not only the restoration of flexural strength, but also the reduction of the degree

of catastrophic failure in the repair zone.

132

133 '
7.2 The S-Joint Design

In-the-plane repair of thin panel composites was primarily the focus of this study.
In evaluating many creative joining designs, it became apparent that there should be an
optimal "isotropic" design for joining two separated composite panels. The S-Joint
design attempts to optimize all of the advantageous repair results determined in the
results section (Chapter 5) of this study.

Looking back to the study of in-plane bond line geometries, or specifically the
"In-Plane Bond Line Geometries, " the results concluded that a bond line angle of 90° to
direction of load was indeed the optimum design. When the bond line angle study was
taken one step further it revealed that a V-joint configuration, which utilized a bond line
angle of 76°-86°, depending on the particular design, exhibited superior results. It was
proven in the study titled, "Depth of Joint" , that the superior strength results encountered
were primarily due to the bond line angle rather than the depth of the joint.

All of the preliminary bond line analysis supported the fact that if the bond line
could be maintained at an angle of 90° to the direction of force, then it would be the ideal
bond line configuration. This is where the idea of the S-Joint came from. Figure 7.1
illustrates the S-Joint bond-line configuration. Note that the design is based on the
adjacency of two circles, offset on a 45° axis. The repair zone was established to be 2"
x 2" (50mm x 50mm) to maintain symmetry. Also, it was proven in the "Depth-Of-Joint
Study" that a 2" (50mm) bond line depth was sufficient for the restoration of flexural
strength. Note also that the centers of both circles are offset along the 45° line, to allow

for the radii to be greater than 0.5” (13mm). The idea of

134

'S-JDINT'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L——e 4

NOTE: ALL DIMENSIONS IN INCHES.

Figure 7 .1 The 'S-Joint" Bond Line Configuration.

135
utilizing a modified circle as a bond line supports the isotropic bond line model which

attempts to keep the direction of force always at 90° to the bond line.
A calculation for the diameter of the S-j oint circle must be determined before
repair takes place, and the circle diameter is dependent on the size of the area nwding

repair. A sample calculation for a 2"x 2" (50mm x 50mm) repair zone is as follows.

Using the pythagorean theorem,

32 + a2 = (2r)2

a = (2)"2 r

Since the repair zone is 2"x 2" (50mm x 50mm), (r) results in,

2" =2r+a=2r+(2)"2r
= 3.414r

r = 2"/3.414 = 0.5858”

A. Three-Point Bending

The first series of tests designed to evaluate the S-Joint was the three-point
bending analysis. It was decided that the analysis of the S-Joint would primarily be
compared to the butt-joint, which is the least desirable bond line configuration. Results

for this series of tests are illustrated in Figure 7 .2. They reveal that indwd the S-Joint

136

29:92 26.3 .5882fi e5 .83. 285% 8.2 .263 2.. 8.. £33. 25858 S. as»:

323 zocéooczoo 5.2.

._.Z_Ool._.5m ._.z_m00lm 0.250102

 

doéui afizfleuumogo oz 1.
02620 mzmsouwo<z<o oz .l
96 3.0 .590 a

mzuso a

 

0

'

I00F

I00N

 

 

I
O
O
‘-
(DdW) HroNaars

l
O
0
ID

I

I000

 

 

00x.

137

is superior in flexural strength to the butt-joint by a margin of 45%, for the 0° joint
orientation. The testing was conducted utilizing the standard repair procedure outlined
in section 2.6, which implemented two layers of glass reinforcing patch material per side
(a double-lap joint).

An additional set of evaluations were setup to also vary the orientation of the 8—
Joint within the 2" x 2" (50mm x 50mm) repair area. The different orientation angles
were 45° and 90°. The results are illustrated in Figure 7.3. The results are slightly
lower than the 0° orientation. Consequently, it appears that the repair area is actually

too small to achieve valid symmetric testing of flexural strength for the S-Joint design.

B. Critical Buckling

The next series of tests were conducted to establish, not only the S-Joint’s critical
buckling load, but also the compressive limit load to be utilized during fatigue testing.
As a note, the S-Ioint was not evaluated for compressive strength since its minimal gage
length of 2" (50mm) allowed buckling to occur. The S-Joint specimen were abraded at
both ends to alleviate slippage in the grips on the Instron machine. Again, the theoretical
calculations were evaluated using Euler’s theory, and the results of this survey are listed
in Table 6.3. The results revealed that the S-Joint, at an orientation of 0°, was again

superior to the butt-joint with regards to its critical buckling load.

C. Fatigue

With the critical buckling load information in hand, the next survey was to

138

22.92 ”528 seguefi e5 .25 Esesm 5.2 .283 2.. co «8.385 8an .8 338m 2 2:5

$2383 557m

 

 

 

 

oo mm o_o ow om on o
«Zoom 125. «>333 E_o....m a o
$948 “.323 oz 0 .
Amzuzov “.223 oz 6 flooF
«Zoom :82... Q6538 5.2...” o .
fioéuss Essa SNImo§<o oz 1. S
fioéeé mzusonmo<z<o oz I. Ioou “lo.
M
o oon m
m H
. oov
m \w/
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII d
oom mx
ooo

 

 

005

139
cyclically fatigue both the S-Joints and the butt-joints at 100, 1,000, and 10,000 and

cyclic intervals. This was investigated to evaluate if there was any loss in flexural
strength after cycling. The results are shown in Figures 6. 12-6. 15. The most prevalent
observation was that during cycling, if there was any gap along the bond line, then the
glass patch material tended to locally delaminate atop of the gap. In some cases, this
initiated bending failure at a lower load, however, the patch material still distributed the
load very well, and failure of both joints was primarily due to patch failure rather than

patch delamination.

D. Torsion

The servo-hydraulic Instron was again instituted to conduct torsion studies on the
S-Joint and butt-joint. Again, both ends of the specimen were abraded to alleviate
slippage in the Instron grips. Rotations of 5° were applied to each specimen, while a
simultaneous reading of torque was recorded. This procedure was carried out until the
maximum Instron rotation of 50° was achieved. The results for torsional rigidity are
illustrated in Figures 7.4-7 .7 . The 0° S-Joint again revealed superior strength to the butt-

joint.

E. Tension
A comparison of tensile strengths was undertaken to verify if the 0° S-Joint was
superior to the W-Joint and the Straight-Line joint. Results for the three joints

undergoing tensile failure are illustrated in Figure 7. 8. However, the results reveal that

140

.8283 Boz< zocfiom
00 m... 0.4 Mn 00 mm 0m 0.. 0. m

_ _ _ r _

85...: a; .68.. .523 eaves. 2.. a, 2.3m 2.. a. 5.2.... .228... .02... 2.58 .830

 

 

52.001m .Il
525010.50 II

 

100

I00«
1000
fi00N

IomN

 

 

con

«522$

(SCH-U!) Nousaor

141

85...: a? :83. 28:3 3.2.3.... 2.. a, 2.3m 2.. 8. 0.2»... 38.38. .29. .8... m... as»...

$8.83 mnoz< 29251
cm

0N

0

av
.

09 m

 

 

 

._.z_0_.lm I

525015.50 I

0
Ion
0.
loop Mu
. O
n
3
700w \/
“lo.
.._I
ICON m.
(
TOMN

 

 

con

142

68...... a? :83. 285.... 3.2.3.. 2.. 9 .583 o... .o. 0...»... 2.5.8... .52...“ .85... o.» 2......

.8883 502... 2925”.
00 me 0.. mm. 0+0 0N 0m m». 0. m 0

_ _

 

 

— — 0
100
m
700—. HO
O
n
3
Ion. )
w.
P
IOON m
(
IOmN
5.001m .ll
E_001Em It

 

 

 

000

143

.§....: 83 has. ..§§« .2622... 2.. 3 263 o... a. .96.»... 38.2.: .33-n.§ .85 E 8:3".

$8.503 302.. 20.25”.
on By o¢ mm mm mm cu m— op m o

 

 

 

 

 

h _ _ _ P n O
100
m
100.. ud
O
n
3
Ion. )
w.
-_I
ICON W.
l..\
10mm
._..Z_O_JIm .Il
._.Z_O....l._.._.3m ll
COM

144'

 

A!

 

 

 

 

con

...8....: 3? sag”. 3.2% 22...". 2.25 26935 3.2.5.. as. :3 ..m 3 8.5.58 m... 2:2".
3.5. min. mmfio ...o mmmzaz
n N P o
b b L 0
10m
m 5
w 100.. H.
< 3
m n N
4 4 . w
d 709. H
\w/
Q
108 We
(
«Emma 02 II
#26733 < TomN
._.z_o..lm n
._.z_o_..l.r5m o

145

the W—Joint is superior to the S-Joint. The reason for this phenomenon may best be
interpreted by observing Figure 7.9 and the particular bondline loading configurations.
The three-point bending bondline for the W-Joint follows a consistant 83°, while the 8-
Joint bond line traverses though both 0° and 90°. It has been shown that the optimum
bond line angle for bending should be close to 90° while remaining invariant the direction
of loading. The S-Joint is the optimum configuration to satisfy this criteria. Therefore,
the W-Joint exhibited superior results in this particular loading configuration because the
bond line was not allowed to shift. However, if the bond line was shifted to a parallel
position to the W-Joint’s 83° then an insufficient butt-joint type bond line would be
created. Additional comments should also be made regarding the tension bond line
configurations. Again the W—Joint retains a consistant 7° while the S-Joint again
traverses through 0°, 45 °, and 90°. Therefore, the results show that the optimum bond
line configuration for tensile loading should be close to 0° to the direction of loading.
Since the W-Joint maintains this set of criteria, it remains the optimum joint for pure
tensile loading.

Ultimately, the results shown in Figures 5.20 and 7.2 reveal that the W—Joint is
superior in three-point bending and in tension. However, note again the bond line
differences between the S-Joint and the W-Joint. If the W-Joint were loaded along the
edge of its ”W-Shape" , then it would simulate a butt-joint configuration which is

much less superior, in strength, to the more “isotropic” configuration of the S-Joint.

7 .3 Stitching Repair for Torsional Damage

146

'VW-JUINT' ’S-JUINT’

 

 

TENSION
BDND LINE (7,7,7,7,7, 7,7, 7,7) (0,90,0,45,0,90,0)

THREE-POINT
BENDING (83.838383838383823) (90,0,90,45,90,0,9U>

 

 

 

 

 

 

 

 

 

 

 

BUND LINE

. . . , . TENSION
7° 7° 7°7° 7° 7°7 7° 7° 0 9.01330906 0 LOADING

A s 7 a a l ft”. A A

I I l

0° 3-PDINT
. / BENDING
90° ._ 9 LOADING

83° 0

90° " 9o

’ /

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v vvvv
7° 7°7° 7°7° 7°

000 6 60 00

¢ $1 TENSION
V V V V . LOADING
7 7 7, 0. 90.0.4 .

Figure 7 .9 Schematic of the S- and WW-Joints under both Tensile and Three-point
Bending Loading.

147

The effects of torsional damage are quite different from the damage modes created
through impact, bending, tension, compression, and fatigue failure. None of these
damage modes experience any appreciable delamination of the matrix material in the
composite during failure. This aspect is what separates torsional damage repair from the
rest. Therefore, a different technique must be implemented to restore the torsional
rigidity of the specimen after failure.

Preliminary tests revealed that to create failure, the torsional coupons would have
to be trimmed into dogbone specimens with a gage area of 1.5" x 2" (37mm x 50mm).
This again was necessary to create failure at a rotation angle of 50° . The failure mode
that occurred is illustrated in Figure 7.10. The first mode of failure started on the edge
of the specimen in the form of matrix cracking. The crack soon propagated across the
top and bottom of the gage length and met, through delamination, at the middle.

The first screening of repair followed the conventional form used for the bending
specimen. The results are listed in Figures 7.11-7.14, and reveal that the original
torsional rigidity could not be restored utilizing this repair technique. Therefore, it was
decided to try a combination of mechanical and adhesive bonding in the form of stitching
[102]. Consequently, this repair mode shifted into a through-the thickness mode rather
than in-the-plane repair.

Two different thread types, nylon and steel, were used during this repair
procedure. The parameter under differential comparison was the thread density in the
repair zone. Some typical patterns of stitch design are illustrated in Figure 7.15. The

holes were drilled with a 3/32" drill bit. The nylon thread was approximately 1/32"

148

 

 

 

 

 

 

 

 

 

 

 

T 5,, CRACK
3.5 ‘ GROWTH
2.75 T (DELAMINATION)
1.75 .L /l *—
.75
T 2.0 #125

 

 

 

NOTE: ALL DIMENSIONS
IN INCHES.

Figure 7.10 Torsional Failure Mode for Owens Corning SMC.

149

5%.. 285.» .82 as. 0355 2&3. 35¢: 9.58 .53 .o 5...»... €295

. $0833 ”:02... 20.25..

.2. as»...

 

 

 

on me 0... on on mm ON 9 0.. m o
b h — L P b - r b 0

Tom
loop
109.
ICON
10mm

cum—(mmmlomogo 1..

min—mm 02lwo<2<o 02 ll

 

 

con

(scan—U!) anoam

150

.98.. 2.2....» 5.2 .5 09.5.5 280.. area: a...» .8...» .o 3.2»... 3.2%... N: “.5»...

.8883 Boz< 20.25..

mm

on

mm
L

0N mm or m

_ _ _

 

 

 

   

 

  

' .
""||“

cum—(mumlomoga II
madam 02|uo<2<o 02 ll

700

700—.

109.

102..

IomN

 

 

oon

(sqI—U!) anoam

151

e.g.”... 285.» .32 .2... age 2o...» 3.5.»: 5.... n...» .8... .c 5...»... access as 2:»...

$00523 502.. 20:50..
on me 0... W». cm mu cu mm or m c

 

 

F . . _ . . . 0
100
Ian:
Ion P
ICON
iomN
ochmumlamogo It
«3m... ozuuo§<o oz .l.

 

 

 

con

(sqI—U!) anoam

152

e.g.... .565.» .3... .2... 09.5.5 2&3. 38...: :53. m...» .8... .o 5...»... .2228. 3 ... as»...

.8233 302.... 2925..
,8 9. 0+ Mn om. mm om MP 3 m o

 

 

— — o
I on
l 00 F
I 00 F
I CON
1. CON
Gum—(mmmlommuéo t I
man—mm OZIM0<I<O 02 ll.

 

 

 

con

(SQI-U!) anoam

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

k 1.... 7 kg, .
1:.w :i
B. (224) K] d {313) Kl (M)
l I» A
4— \l mwm
__|

 

Figure 7.15 Sketch of Drilled Hole Patterns for Stitching During Torsional Repair.

154
while the steel thread was approximately 2/ 32 ". Motivation behind this particular hole

and thread size was to (1) not create holes too large since this would create large voids,
and (2) choose a thread diameter that fits well through the hole without too much of a
gap. Any gap that was created between the thread and the drilled hole was filled by the
application of epoxy on the specimen surface. Since the steel thread was less flexible
than the nylon, it could only be drawn through the hole once. In some cases the nylon
thread was drawn through the holes twice, therefore doubling the density at that spot.

In this repair scheme, no glass patch material was utilized, only thread
(mechanical) and epoxy (adhesive). The results of this new form of repair are illustrated
in Figures 7.16-7.18. The conclusions from the nylon thread revealed that the nylon
thread was not rigid enough to enhance the through-the—thickness rigidity. Also, it was
an interesting observation that the failure propagated from the same zones as it had
before in the reference undamaged specimen, which revealed that the edge density of the
stitching pattern is very important, while repair of the central region adds less structural
rigidity.

The results for the steel thread are illustrated in Figures 7.19-7.21. Note the
increase in strength over the nylon thread as well as the increase due to thread density.
By achieving a stitching pattern of 3 x 6, the torsional rigidity was restored. At the
thread density of 4 x 7, the torsional rigidity was 33% higher than the original
undamaged specimen rigidity.

These conclusions were very encouraging, but there was still a question as to what

parameter was contributing the most to the repair rigidity. Was it the portion of thread

155

.3. n base .82 35...: 2.. so»... 2.2853. a»... oz
2... .532 m. 28.5. och. .283. 322.6 .0...» E... omnEaQ 283— 3.552 MEEcU 2030 .3 5.2»... 3:239... o. .h 0.5»...

A888... .302... 2925..
on m... 9... mm on mu cu m. or m o

 

 

 

_ h — — _ P _ C
low
M
IOOP no
0
fl
3
ion. \1
m.
-_u
T.oou m.
. (
nomu
9...». 1mm: zo..>z .:.
man—um 02lu0<2<o 02 I...

 

 

 

con

156

.65 u 9.8.. .82 68...... 2a 29.3 9.26.6.3. .8... oz
2.» .532 m. 28...... 2F £80.. 323cm 3...... E... own—=5 283 3.262 “EEO“. 2030 .o 5.2»... REES. 2 .5 2:»...

.8868. 502.. 29.25..

 

 

 

 

on me 9. on on mm cu m. o. m o
P p p r L b L b F 0
non
m
IOOF no
0
n
3
ion. )
w.
I_.
loom w.
/I\
103
6an 1mm: 20.»: E.
«2%... ozumo§<o oz I.

 

 

con

157

.63. u €89 =8: ...8....: a. 53. «c.2853. .86 Oz
v5. .532 £ .585. 2:. £83— mczozfi .2... 23 03.59 288 3.032 9.500 3030 ..o >3.sz 3:295... w. H 23E

.8283 302... 2925..
on me 0* mm on mm om mm or m o

 

 

 

_ _ . . p _ 0
non
m.
loop no
0
n
3
100.. \./
w.
.._-
loom m.
(
-omu
fix... :8: zo..>z ..
«2...... oznuo§<o 02 ..l

 

 

 

con

158

din. u 5.82. ..8: 3...... as 22.3 9.28%.. .36 c2
Ea ..8.m a. 23...... 2.... c.83— w.....u...m .3... Ba own—55 283. 3.822 w=.EoU 2030 ..o 5.29% .2233. a. .h 23.".

.8833 5oz... 292.6”.
on m... o¢ mm on ma cu m. o_. m o

 

  

 

. _ _ _ _ b b C
..om
{If}--- ........ m.
.IOOF Mu
O
n
3
nom— )
w.
.._..
..oou m
(
uomu
. .2318: .HFm ..
«2...... 02:83.3 oz I.

 

 

 

con

159

.63. n 5.2.5 .82 63...... 2.. e93. ”528...... .36 cz .5
..oo.m m. 23...... 2.... died 9.28.5 .2... ..ca 0»..an Baum 3.532 95.80 «530 ..o .9632 .2568... cm... 2%.".

.8833 302.. 2925”.
on we 0.. on on ma om m. o. m 0‘

 

 

_ b _ P h _ L L 0
10m.
m... .
100.. ud
O
n
3
won. )
w.
._.
woou m
/I\
10mm
8x3 :82 .5um ..
«2...... ozumo§<o oz I.

 

 

 

00m

160

.63.. u 2.2.8 ..82 68...... an 293. ”2.26.2.3. 26.0 oz 6.... 6.3
2 68:... 2: 2.2.3. 2.22.6 .2... 6.... 82.5 26.8. 3.362 esteem .8... .6 2.6.»... .2223 .2 2......

.8883 302.. 29.25..
on 06 0... on o... mm mm mm o. m o

 

 

 

r b _ _ P _ 0
100.
m.
(III. . \\\\\\ IOOF a
I: uuuuuuuu ......\ .0
n
3
1cm. x)
m.
._.
loom m.
(
loom
. .3018... ..mEm I.
22“.“... 0229...... oz I.

 

 

 

oon

161
that was through-the-thiclmess of the specimen or the portion that was in-the-plane? One

more test was conducted to evaluate this question. A torsionally damaged specimen was
repaired by laying four steel wire strips along the gage length on both sides of the
specimen. See Figure 7.22. The wire strands were set in place and coated with epoxy.
The wire ends were left long enough so that during the testing procedure the ends could
also be anchored in the Instron grips. This simulated in-the-plane repair with added
supports in the gage length by the addition of two anchor supports at both ends of the
specimen. The results are illustrated in Figure 7.23, and they reveal that about 50% of
the restored torsional rigidity is contributed by the in-plane stiffeners while the remaining
50% comes from through-the-thickness stiffeners.

The application of such a repair design focuses primarily on those endurable
goods that simply must be repaired rather than replaced. With the absence of glass
reinforcing material, the thickness direction of the repaired specimen is negligibly
enhanced, only about 0.02". This aspect is very desirable when and if the torsional
repair specimen is a spinning shaft, since it would reduce the effects of any complicating

factors due to rotational eccentricities and surface roughness.

7 .4 Optimization of Design - A Computer Model

After the designs for the S-joint were complete, a computer model [103] was
implemented to evaluate an in-plane repair geometry resembling a butt-joint. The finite
element model is based on interlaminar stress continuity theory which utilizes a multiple

layer approach. Therefore, the analysis of through-the—thickness repair can take place.

162

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/
0 i—’ _J ’_’ CLAMPING ZDNE
[\r f
3.5
2.75 I
1.75 l
L/ _, H CLAMPING ZUNE
’75 f‘i ’ , ’— - J

 

 

 

 

 

NOTE: ALL DIMENSIONS
1N INCHES.

Figure 7.22 Schematic Showing the '4-Wire-Strand" Configuration.

163

Scam :83— :oEo... o... .8 3.0 2.3 caisson— « 5.? Dim ”5:80 3030 8.. 2.80: mm... 05w."—

Ammeooa 302.. zoEfiom

om

 

 

 

 

9. 9. mm on ma cu m. 2 m o
_ L L _ — F b _ h C
100
floor
tom.
floom
10mm
.8255 e. was. ..HFn. ..
magma oznmo§<o oz ..I

 

 

con

(Stu—U!) anoam

164

The model incorporates a Hermite cubic shape function to analyze the composite layer
assembly in the thickness direction. By using the high-order shape function, the
transverse shear deformation can be considered in the analysis. The high-order shape
function enables the evaluation of interlaminar shear stress and interlaminar normal stress
at the laminate interfaces directly from the constitutive equations, rather than recovering
them from the equilibrium equations. Consequently, since the theory and technique are
two—dimensional, it is valid for both thin and thick composite laminates. Numerical
results show that there is excellent agreement with the elasticity solution.

The finite element mesh that was developed using the aforementioned model
consisted of six layers and ten elements. The model was subjected to three-point bending
throughout the analysis. Figure 7 .24 is a schematic of the finite element mesh. The six
evaluation layers consisted of: (1) patch material plus resin on the top two layers, (2) the
composite material on the middle four layers, and (3) patch material plus resin on the
bottom two layers. The frwdom of the six different layers allowed modification of not
only the composite material, but also the patch thickness, and moduli. Because of this
freedom, three different surveys were undertaken. The first analysis evaluated variances
in patch thickness. The different thicknesses were based on two layers of patch on the
top side of the composite measuring 0.4mm, and the addition of one, two, three, four,
and five patch layers with thicknesses measuring 0.2mm, 0.4mm, 0.6mm, 0.8mm, 1mm
respectively to the bottom side. Results of this survey are illustrated in Figure 7 .25,
along with a correlation of the experimental data. Note that the values of stress have

been normalized in order to correctly correlate the experimental value with the computer

165

FINTTE ELENENT [RID [f 6 LAYERS BY 10 ELEMENTS
NUDEL IS SET UP EUR THREE-POINT BENDING STN

 

1 PARA Pujs RESTN

PATCH PLUS RESIN
CUPPUSITE; MATERIAL
CENPUSITE NATERIAL

PATCIH PLUS min
PATCN PLUS min

 

 

 

 

 

 

 

 

 

 

 

O‘U'l-n-w’m

 

 

 

 

 

TL

SINPLE SlPPflRT ANALYSIS AREAS

 

 

Figure 7.24 Schematic of the Finite Element Mesh.

166

o.—

.agm 82.2.82 $85.25 6.9. .o ”5.082 3.2.58 as ea»...

.55. mmuzxgi 18E

m6 6.0
_

.vd
E

Nd

_

 

 

1 m6

0...

 

 

Ea ..fizmzmmaxm .i
.32. «3.228 ..I

  

_OE.._OC flocflLOwOL

10;.

1 ON

1 9N

 

 

0.0

883818 OEZI'TVWHON

167

model. The second survey dealt with varying a gap or void size on the composite layers
of the model. Gap sizes that were evaluated were based on a half gap length due to the
fact that the finite: element model is symmetric with respect to the composites in-plane
centerline. The half-gap lengths that were evaluated were 0mm, 1mm, 2mm, 3mm,
4mm, and 5mm. These gaps were created assuming pure epoxy resin was to take their
place. The normalized stress results are listed in Figure 7.26 and 7 .27. Figure 7.27

correlates the shear stress occurring at the bottom of the composite layer two and inward
from the centerline of 5mm. The third survey, kept the half gap length at 5mm and
modified the fillers used in the gap or void zone. The results for both normalized stress
and normalized shear stress are illustrated in Figures 7 .28 and 7 .29. The experimental
results correlate very well to the computer model with respect to displaying the same

trends during loading.

168

.395 332.82 659.3 95 2.2.3.5 he maze—co: .2..an0 ea... 9...»...—

EE. :52”... Q. n20
n N P o

_ — b

10
w

 

 

 

 

Echo: 3:889.

In...

(.55 4<P2w§mmmxw ll
Juno: mkamzoo II

 

o.N

 

 

$83818 GETZI'IVWHON

169

32% 325 39:2ch 559.3 9.0 2.2.2:: «o “$.03: 83925 has 05w...“

AEEV :52”... m) .20

 

 

 

 

... .._. m N. . o
o
I t I 11.1.. 0..
.250: 3:882
a. m.—
JwOOZ mwSQZOO ll
O.N

 

88381.8 ENC-1H8 OBZI'IVWETON

170

sea ..8.....Eoz e.g.... 9.0 5.2.3.... .o 9...».5: 5.2.58 ”a... 9...»...

on

Anon—xv mn_<_mm._.<§ mung... n_<o “.0 Q3300:

 

 

 

 

 

ON 0.. o
b P b P P b by h LP P n
o

1 0.0
0;

Echo: 00:230..
1 n;
1 O.N

<._.<o ..(hZMEEMdXU ll

doc... $5.38 I. 3

 

88381.8 GBZI‘IVWHON

171

on

\

.22.... 525 58.5552 .50.... ..8 5.2-2... 2. ”5.0.82 8.2.58 an... 0...»...

can... 936:... Ed: 93 .0 8.582

 

 

 

ON 0 _. 0
b I LP b L—L L b P b ~ - P P b b
.1111 0
. III
. 111
~ II
~
.. 1 0.0
N
\~
\. 0..
\\ Beta: 3:889.
\\ u m...
\\ .. o.~
Jun—02 mgmioo 1.... ON

 

 

883818 8V3H8 G3ZI'IVW8ON

CHAPTERS

SUMNIARY

Based on the experimental results, some important conclusions and
recommendations can be summarized as follows.

1. A standard repair technique based on reinforcing patch and bonding adhesive
has been verified to be a feasible technique for repairing composites with line cracks and
circular cutouts. Since these two types of artificial damage modes and their combinations
can closely simulate various types of impact-induced damage, the standard technique can
be used for repairing real damage.

2. In order to have good bonding efficiency, the mechanical properties of the
reinforcing patch and adhesive should be compatible with those of the fiber and matrix,
respectively, of the composite under repair. In addition, a low viscosity adhesive and
a woven type of reinforcing patch material can help to improve the repair efficiency.

3. The fiber geometry plays a very important role in composite repair.
Composites consisting of chopped fibers and continuous swirl mats have randomly
oriented fibers. Therefore their prOperties are more matrix dominated. The woven
fabric composite is more fiber-dominated. Hence, the matrix-dominated composite can
be more efficiently repaired than can the fiber-dominated composite since the latter
requires a higher strength reinforcing patch to achieve the required repair strength.

4. In addition to the type of adhesive and reinforcing patch material, the

geometrical parameters are the most important elements for efficient composite repair and

172

173

joining. The geometrical parameters can be divided into three categories: in-plane
configurations, through-the—thickness scarfing, and out-of-plane curvatures.

5. Many in-plane configurations have been examined. It has been found that an
angle between 60° and 90° with respect to the loading direction results in the greatest
repair efficiency. This result can be used in designing an optimal bond line
configuration. A complex bond line geometry gives both higher repair strength and
sufficient bonding area.

6. Bonding configurations associated with the thickness direction are scarf joints.
A smaller scarf angle has been verified to be more effective than a larger angle, though
the difference is not very significant. Therefore, a butt-joint is recommended for thin-
layer composite repair, because of their efficiency in preparation.

7. The effects of out—of-plane curvatures on the joining and repair efficiencies are
also important factors, and should be incorporated into the repair design.

8. Through-the—thiclmess stitching has been verified to be very effective and
efficient for repairing failure due to delamination in torsion.

With the complete information from the geometrical analysis, the adhesive study,
the filler examination, and the reinforcing patch investigation, it is believed that the

optimization of repair design for damaged composite structures can be achieved.

APPENDIX

TABLES

174

Table 5. la Owens Corning Panels: Drilled Hole Survey with Fillers.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX FILLERS I
ID (mm) (mm) (MPa) (mm)

1/4H-1 50.00 2.7 219.9 - 1/4' HOLE |

1/4H—2 50.00 240.9 - - I

l/4R—l 50.69 485.89 13.88 RESIN IN
HOLE

1/4R-2 49.89 540.29 16.18 -

1/4P-1 51.21 477.74 10.38 PLUG IN HOLE

l/4P-2 50.35 523.50 10.94 -

1/4P-3 51.81 471.41 11.14 -

l/4P-4 50.43 558.57 11.88 -

1/4P-5 51.40 529.62 12.62 - |

l/4F-1 51.02 458.95 14.14 CHOPPED 1
GLASS

l/4F-2 50.29 522.90 14.68 m HOLE

1/4F-3 50.54 515.83 10.56 -

1/2H-1 50.00 204.40 - 1/2" HOLE

l/2H-2 50.00 200.0 - - |

l/2R—l 51.63 470.67 11.83 RESIN IN I
HOLE

l/2R-2 51.13 522.76 10.39 - |

l/2P-1 51.19 568.77 11.19 PLUG IN HOLE |

l/2P-2 51.30 473.69 9.97 -

1/2P-3 51.99 445.24 11.98 - I

l/2F-l 51.19 400.35 10.76 CHOPPED I
GLASS

l/2F-2 51.75 458.04 11.22 IN HOLE I

l/2F-3 51.81 532.18 11.08 - I

 

175

Table 5.1b Owens Corning Panels: Drilled Hole Survey with Fillers.

 

176

 

 

  

 

  

 

 

 

 

 

 

 

 

  

 

 

 

  

 

 

 

  

 

  

 

 

 

  

 

 

 

 

 

 

 

 

 

 
 

Table 5.1c Owens Corning Panels: Drilled Hole Survey with Fillers.
SPECIMEN WIDTH THICKJ STRENGTH FLEX FILLERS
113 (mm) (mm) (MP3) (mm)
OCHl-l 49.74 3.11 116.40 13.5 1" HOLE
OCH1-2 50.17 2.99 114.90 13.0 NO FILLER
OCHl-3 49.53 3.06 119.00 13.1
OCCP-l 51.39 472.28 16 STRAP JOINT
OCCP-2 50.05 364.21 16 IN 1" HOLE
IL OCCP-3 49.88 417.46 16 1-3= 1 PATCH
OCCP—4 51.52 462.18 17 2 PATCHES
OCCP-5 50.58 500.37 18 2 PATCHES
OCS-l 50.03 3.08 554.00 15 2 PATCHES
OCS-2 49.92 3.09 509.60 10 2 PATCHES
OCS-3 51.00 3.09 449.80 10 2 PATCHES
OCF-1-3 51,50 3.09 340,296 10.8 CHOPPED
GLASS
OCCF-l 51.15 429.80 - CHOPPED
GLASS

177

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.2 Owens Corning Panels: Impact Study-Specimen Repaired with
Two Glass Patches Per Side.
SPECIMEN . WIDTH THICK STRENGTH ENERGY FLEX HEIGHT R
ID (mm) (mm) (MPa) (Joules) (mm)
(inches)
OCI-l 49.56 2.70 512.9 26.59 10.8 38.5 Y
OCI-2 51.43 414.6 26.65 10.8 38.5 Y
0C1-3 48.27 404.0 28.91 11.2 35 Y
OCI-4 48.39 456.8 28.58 11.6 35 Y
OCI-5 49.80 599.3 28.35 10.9 35 Y
OCI-6 51.68 517.3 29.59 10.7 37 Y
OCI-7 50.55 453.6 28.16 11.5 37 Y
OCI-8 52.28 418.5 27.12 10.6 37 Y
OCI-9 49.60 432.4 23.50 11.6 40 Y
OCI-lO 49.17 505.6 25.46 10.7 40 Y
OCI-ll 49.53 409.4 24.77 11.1 30 Y
OCI-12 48.22 362.1 25.34 11.3 30 Y
OCI-13 51.29 359.8 25.86 10.8 25 Y
OCI-14 48.98 437.8 23.32 11.3 20 Y
OCI-15 49.22 383.9 23.32 10.7 20 Y
0C1-16 48.75 440.8 11.90 9.9 12 Y
OCIR 52.18 2.78 208.8 5.49 15.1 8.25 N
OC2R 49.84 2.98 161.1 13.57 13.4 14.5 N
OC3R 49.66 2.95 161.2 21.15 12.2 21.25 N
OC4R 51.29 2.98 142.4 26.17 11.2 25 N
OCSR 51.00 2.95 150.6 25.77 - 12.3 27.5 N
OC6R 49.36 3.08 138.5 28.82 10.8 27.5 N
OC7R 50.40 3.06 137.4 26.53 11.0 27.5 N
0C25R 49.18 2.74 158.2 33.70 12.7 30.0 N

 

 

 

 

 

 

 

 

 

NOTE:R=REPAIREDEI'1'HERYES(Y)ORNO(N).

 

178

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.3 Excel 8610: Drilled Hole Survey.
SPECIMEN WIDTH THICK STRENGTH HOLE SIZE FILLERS
1]) (mm) (mm) (MPa) (inches)
8610N-1 49.28 3.04 175.06 0 NONE
8610N-2 49.40 I 292.85 0 NONE
8610N-3 50.06 181.05 0 NONE
8610ND-l 50.40 320.75 0 NONE
8610ND-2 51.18 307.34 0 NONE
8610ND-3 49.99 309.42 0 NONE
8610ND4 50.03 345.89 0 NONE
8610H1-1 50.64 86.06 1 CHOPPED
GLASS
8610H4-1 49.25 178.72 3/4 FIBERS
8610H4-2 49.30 117.57 3/4 -
8610H4-3 49.73 212.05 314 -
86101-12-1 49.17 230.41 1/2 -
8610H2-2 50.17 242.32 1/2 -
8610H2-3 50.1 1 135.65 1/2 -

 

 

 

 

 

 

 

 

179

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.4 Excel 8610: Impact Study For Equivalent Damage Hole Size.
SPECIMEN WIDTH THICK STRENGTH HEIGHT ENERGY REPAIR FLEX
ID (mm) (mm) (MPa) (inches) (Joules) vs (mm)
.811.
8610-A 51.33 3.03 - 22.5 24.46 NR
8610-B 51.20 3.00 - 20 21.86 NR
8610-C 50.49 2.95 - 19 22.05 NR
8610-D 51.89 3.08 ~ 18 20.88 NR
8610-E 49.41 3.00 - 18 20.68 NR
8610-P 50.73 3.82 347.78 28.5 31.12 R 13
8610-G 51.68 3.16 208.26 28.5 29.32 NR
8610-H 52.47 3.22 209.27 25 19.49 NR
8610-I 50.24 3.09 202.95 27 24.04 NR
8610-J 53.75~ 3.64 330.05 22.5 15.17 R 14.5
8610-K 51.98 3.49 301.76 28.5 28.52 R 13.3
8610-L 52.18 3.64 356.47 34.5 34.20 R 13
8610-M 53.37 3.23 208.20 34.5 32.86 NR
8610-N 53.08 3.84 385.90 28.5 32.38 R 13.5
8610-O 48.69 3.92 376.69 22.5 15.54 R 13.4
8610-P 50.86 3.12 188.48 30 34.39 NR
8610-Q 51.39 2.96 181.54 29 31.20 NR

 

 

Excel 2415: Drilled Hole Survey with Fillers.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH COMMENTS FIBER
ID 5 (mm) (mm) (MPa) DIRECTION

2415-EXCEL 50.0 2.70 384.3 NO DAMAGE FILL
2415-31 48.93 2.88 353.3 NO REPAIR FILL
2415-32 49.11 2.85 391.9 FLEX =24 FILL
2415-33 49.06 2.93 355.8 FILL
2415-34 48.99 2.70 401.5 FILL
2415-35 49.08 2.95 351.6 FILL
2415-36 48.98 2.95 364.7 FILL
2415-7 49.03 2.87 278.3 1/2" HOLE WARP
2415-8 49.02 2.84 307.6 NO REPAIR WARP
2415-9 49.04 2.80 373.0 FLEX= 18mm WARP
2415-10 48.69 2.79 405.0 WARP
2415-11 49.02 2.85 315.5 WARP
2415-12 48.84 2.92 353.1 WARP
2415-13 49.02 2.92 280.4 3/4' HOLE WARP
2415-14 49.06 2.87 268.1 NO REPAIR WARP
2415-15 49.04 2.78 275.3 FLEX= 15mm WARP
2415-16 48.97 2.94 259.8 WARP
2415-17 49.03 2.81 307.9 WARP
2415-18 49.00 2.92 303.7 WARP
2415-19 49.08 2.90 205.9 1" HOLE WARP
2415—20 48.97 2.87 227.7 NO REPAIR WARP
2415-21 49.04 2.92 159.9 FLEX = 13mm FILL
2415-22 48.99 2.70 181.6 FILL
2415-23 48.97 2.87 233.7 WARP
2415-24 49.12 2.88 211.6 WARP
2415-25 48.98 2.81 195.8 1 1/4' HOLE WARP
2415-26 49.02 2.83 168.3 NO REPAIR WARP
2415-27 48.92 2.87 161.9 FLEX= 13 WARP
2415-28 48.99 2.84 157.0 WARP
2415-29 49.07 2.73 158.6 FILL
2415-30 48.93 2.84 169.4 WARP

 

 

181

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.6 Excel 2415: Impact Study-Warp Direction.
SPECIMEN HEIGHT TOTAL WIDTH THICK STRENGTH REPAIR/
11) (inches) ENERGY (mm) (mm) (MPa) NO
Ft-lstoules REPAIR
2415-E 40.5 2549,3456 51.62 364.55 R
2415-F 40.5 2549,3456 51.77 2.79 361.82 NR
2415-G 34.5 22.65.30.71 52.12 504.89 R
2415-11 28.5 23.81.3229 51.91 2.79 311.01 NR
2415-I 22.5 15.55,21.09 53.94 529.44 R
2415-J 34.5 24.1,32.68 51.89 2.62 381.08 NR
2415-K ' 28.5 23.87,32.37 53.87 444.93 R
2415-L 22.5 13.15.17.83 52.73 2.74 356.11 NR
2415-M 25 16.32.2213 53.92 573.14 R
2415-N 26 15.45,20.95 53.06 502.66 R
2415-0 27 22.99.31.17 52.79 2.71 341.81 NR
2415-P 27 2239,3036 54.55 416.01 R
2415-S 22.5 12.21.1656 51.86 2.79 367.18 NR 4

 

 

 

 

 

 

 

 

182

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.7a Owens Corning Panels: Butt-Joint Study.

SPECIMEN WIDTH THICK" STRENGTH

ID (mm) (mm) (MPa)
B-lO-NG 51.84 2.70 332.5 I
B-ll-ROL 50.90 307.2 9.66 ROLL 2,2
B-lZ-RES 51.66 289.8 8.50 RESIN 2,2 1
B-13-NG 50.90 302.9 0 - 2,2
B-l4-ROL 51.14 304.7 9.02 ROLL 2,2
B-lS-RES 51.10 281.1 8.00 RESIN 2,2 1
B-16-NG 52.46 297.0 0 - 2,2
B-17-ROL 52.22 286.7 9.12 ROLL 2,2
B-18-RES 51.46 187.9 6.08 RESIN 2,2
B-19-NG 51.24 360.1 0 - 2,3
B-20-ROL 51.66 274.9 9.00 ROLL 2,2 |
B-21-RES 51.24 252.3 4.94 RESIN 2,2
B-22-NG 51.06 280.0 0 - 2,2
B-23-ROL 51.30 325.2 9.20 ROLL 2,3
B-24-RES 52.86 254.0 3.20 RESIN 2,2
B-25-NG 50.70 298.6 0 - 2,3

B-l 48.95 266.3 0 - 2,2

B-2 50.45 222.1 0 - 2,2

B-3 49.22 235.4 0 - 2,2

B-4 50.44 243.1 0 - 2,2

B-5 50.30 269.6 0 - 2,2

B-6 49.18 256.4 0 - 2,2

 

GAP MATERIAL: ROL

NG

ROLLED GLASS PATCH IN GAP, RES = PURE RESIN IN GAP

NO GAP

183

Table 5.7b Owens Corning Panels: Butt-Joint Study.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH GAP GAP LAYERS
ID (mm) (mm) (MPa) (mm) FILL TOP,
BOTTOM
Bol-B 49.62 121.2 0 - 2,1
B-2-B 48.92 131.0 0 - 2,1
B-3-B 49.26 132.4 0 - 2,1
B-4-B 49.26 208.6 0 - 2,2
B-S-B 49.44 184.4 0 - 2,2
B-6-B 49.26 206.4 0 - 2,2
B-7-B 49.52 220.9 0 - 2,3 j
B-8-B 49.14 262.0 0 - 2,3
B-9-B 49.92 . 267.9 0 - 2,3
B-l-T 48.95 196.4 0 - 2,3 I
B-2-T 50.43 206.5 0 - 2,3
B-3-T 49.25 455.3 0 - 2,4 I
B—4-T 50.40 353.6 0 - 2,4 I
GAP MATERIAL: ROL = ROLLED GLASS PATCH IN GAP, RES = PURE RESIN IN GAP
NG = NO GAP

 

 

 

 

 

184

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.8 Excel 8610: Butt-Joint Layer Study.
SPECIMEN WIDTH THICK YIELD ULTIMATE FLEX PATCH
ID (mm) (mm) STRENGTH STRENGTH (mm) LAYERS
* (KN) (MPa) TOP,BOTTOM
86B-21 46.67 3.13 738.3 242.3 6.3 2,2
86B-22 47.16 2.90 896.6 339.2 7.2 2,2
8613-23 47.15 3.13 920.8 299.1 6.3 2,2
86B-3l 46.68 3.11 1141 379.2 7.8 2,3
86B-32 48.27 2.99 1184 411.7 8.5 2,3
8613-33 47.07 2.98 1219 437.6 8.1 2,3
86B-41 47.72 3.16 1621 510.4 9.6 2,4
8613-42 47.82 3.07 1310 436.1 8.8 2,4
86B-43 50.60 2.98 1417 473.1 11.1 2,4
86B-51 51.29 2.95 1621 544.9 12.3 2,5
86B-52 50.63 2.95 1863 634.4 13.3 2,5
868-53 50.20 3.06 1796 573.3 9.1 a 2,5

 

 

 

 

 

 

 

 

NOTE: ALL FAILURE MODES WERE CATASTROPHIC

 

185

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.9 Excel 2415: Butt-Joint Layer Study.
SPECIMEN WIDTH THICK STRENGTH FLEX COMMENTS PATCHES
ID (mm) (mm) x (MPa) (mm)
2482-1 52.64 2.81 197.65 8.4 WARP 2
24B2-2 51.99 166.76 8.6 DIRECTION 2
2482-3 53.45 196.54 7.7 RS-RESIN 2
2483-1 52.05 197.92 11.8 WARP 3
24B3-2 52.31 264.20 10.3 DIRECTION 3
24B3-3 52. 15 302. 16 9.1 RS-RESIN 3
2484-1 53.58 481.48 9.9 WARP 4
2484-2 53.49 429.02 10.6 DIRECTION 4
2484-3 51.37 443.76 8.5 RS-RESIN 4
2485-1 51.04 463.75 9.1 WARP 5
2485-2 49. 18 522.62 9.9 DIRECTION 5
2485-3 51.81 484.29 8.9 RS-RESIN 5
248F3-1 49.0 369.47 10.0 FILL 3
248F3-2 48.31 369.47 9.2 DIRECTION 3
248F3-3 48.83 403.12 10.0 RS-RESIN 3

 

 

 

 

 

 

 

 

 

186

Table 5.10a Owens Corning Panels: Scarf-Joint Survey, Two Reinforcing Patches Per
Side.

SCARF FILLERS LOADING
(degree) IN GAP TOP,BOT

ROLLED

PA

waw—IHH

..1

B
B
B
B

5
I-I—awcu

wthflH—IH—I

 

187

Table 5.100 Owens Corning Panels: Scarf-Joint Survey, Two Reinforcing Patches Per
Side.

FILLERS IN G FLEX WIDTH
(degree) IN GAP TOP,BOT (MPa) (mm) (mm)

4 .

B
T
B

FIBERS
1 IN GAP

 

FOR SPECIMEN SSlS-l,2,5,6...ONLY l LAYER OF PATCH IS USED ON THE BOTTOM SIDE.
THE SPECIMEN THICKNESS USED FOR THIS 18 2.70 mm.

188

Table 5.11a Owens Corning Panels: Scarf-Hole Survey.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX SCARF REPATRj
ID (mm) (mm) (MPa) (mm) (degree)

OCSJO-l 101.13 2.70 273.44 20 0 NO P
OCSJO-2 98.32 301.05 18 0 NO
OCSJO-3 98.06 300.49 17 0 NO
OCSJR-l 98.94 589.56 13 0 YES
OCSJR-2 98.43 562.31 13 0 YES
OCSJR-3 102.84 532.29 14 0 YES
OCSSJ-l 100.54 144.18 19 5 NO
OCSSJ—2 100.39 133.13 27 5 NO
OCSSJ-3 101.42 125.95 21 5 NO

OCSSJR-l 98.00 525.89 14 5 YES I
OCSSJR-2 101.54 486.83 14 5 YES
OCSSJR-3 99.75 557.07 13 5 YES

 

 

 

 

 

 

 

 

 

189

Table 5.11b Owens Corning Panels: Scarf-Hole Survey.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX SCARF REPAIR
11) (mm) (mm) (MPa) (mm) (degree)
OCDSJ-l 101.42 140.31 20 5 NO
OCDSJ-2 101.89 143.07 17 5 NO
OCDSJ-3 101.23 160.74 18 5 NO I
OCDSJR-l 101.39 441.30 13 5 YES
OCDSJR-2 99.93 527. 88 13 5 YES
OCDSJR-3 99.60 569.56 14 5 YES
OCREF-l 100.99 311.54 20.2 0 NO
OCREF-2 100.83 329.77 21.5 0 NO
OCREF—3 100.71 333.64 19.8 0 NO

190

Table 5.12a Owens Corning Panels: Joint Geometry Study.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH TffiCK STRENGTH
ID (mm) (mm) (MPa)
JGO-1 52.37 2.70 314.39
JGO-2 52.19 217.02
JGO-3 51.77 292.42
JG30—1 51.86 405.89
JG30-2 50.57 330.79
I JG30-3 52.07 399.91
JG45-1 52.62 528.16
JG45-2 50.77 484.28
JG45-3 53.09 530.77
JG60-1 52.74 685.14
JG60-2 51.43 608.06
JG60—3 51.58 439.18
JG90-1 51.66 541.17
JG90—2 51.61 648.75
JG90-3 50.24 564.17
JGV-l 51.37 575.43
JGV-2 50.83 645.93
JGV-3 50.68 547.28
JGU-l 52.81 725.04
JGU-z 50.64 738.64
JGU-3 51.53 669.07
JGW-l 51.28 614.14
" JGW-2 52.20 587.44
JGW-3 52.01 517.35
JGUU-l 49.68 611.69
JGUU-2 52.46 690.91
JGUU-3 51.76 708.75
JGWW-l 51.94 558.47
JGWW-2 52.78 588.31
JGWW-3 49.64 650.01

 

STUDY INCLUDES: V,U JOINT CONFIGURATIONS CUT OVER A 2"x 2" AREA.

 

 

191
Table 5.12b Owens Corning Panels: Joint Geometry Study.

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH COMMENTS
ID (mm) (mm) (MPa)
OCS-l 51.44 3.00 358.40 S-JOINT
OCS-2 51.68 3.02 417.00 0—DEGREES
cos-3 51.14 3.02 423.00
OCS-4 50.62 2.89 427.60 S-JOINT
OCS-s 50.46 3.10 352.10 45-DEGREES
ocs-6 51.25 3.06 313.90
OCS-7 98.62 3.06 285.30 S-JOINT
11 OCS-8 99.18 3.06 268.50 90-DEGREES
fl OCS-9 99.26 3.06 260.10 ..J

 

 

 

 

 

STUDY INCLUDES: V,U JOINT CONFIGURATIONS CUT OVER A 2”x 2" AREA.

192

Table 5.13a Excel 8610: Joint Geometry Study.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX FAILURE
ID (mm) (mm) (MPa) (mm) MODE

86JGO-1 51.52 3.04 307.0 5.9 CATASTROPHIC

86160-2 51.57 283.0 - -

86100-3 50.68 314.7 - -
86JG30-1 51.38 331.7 6.6 -
86JG30-2 51.72 407.7 6.6 -
86JG30-3 52.76 433.5 8.2 -
86JG45-1 52.92 396.9 9 -
86JG45-2 51.33 407.6 8.8 -
86JG45-3 52.79 399.5 8.5 -
86JG60-1 53.20 484.8 11 NON-CATASTROPHIC
86J660-2 52.05 483.9 10 -
86JG60-3 51.40 474.0 14 -
86JG90-1 52.79 541.4 13.3 -
86JG90-2 49.64 558.3 13.5 -
86JG90-3 51.15 562.2 13 -

 

 

 

 

 

 

 

193

Table 5.13b Excel 8610: Joint Geometry Study.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX FAILURE
ID (mm) (mm) (MPa) (mm) MODE
86JGV-1 52.08 513.6 10 CATASTROPHIC
86JGV-2 50.58 505.7 10.7 1 -
86JGV-3 52.38 490.0 10.9 -
86JGU-l 53.91 441.4 10.2 NON-CATASTROPHIC
86JGU-2 51.63 409.3 11.5 —
86JGU-3 52.24 396. 1 1 1 -
86JGW-1 51.95 418.5 10.5 -
86JGW-2 52.23 509.7 1 1 -
861GW-3 52.97 483.7 10.8 -
86JGUU-1 53.93 548.7 11 -
86.JGUU-2 52.59 551.0 11.5 -
8GJGUU-3 52.81 ' 491.8 12 -
86JGWW-1 51.73 497.8 10.5 -
86JGWW—2 50.96 533.5 11 -
86JGWW-3 51.35 501.5 10.3 -

 

 

 

 

 

 

 

194

Table 5.14a Excel 2415: Joint Geometry Study-Warp Direction.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX COMMENTS
ID (mm) (mm) (MPa) (m) I

2410901 50.95 2.81 716.96 13 |
2410902 48.12 707.96 16

2410903 48.39 654.61 13 PLOT I
2410601 51.66 613.00 12 CRACK STARTS |
2410602 51.58 604.12 15 ALONG SEAM

AND

2410603 51.54 592.68 15 THEN To PATCHJ
241045-1 51.63 411.72 8.1 1
200452 52.4 429.20 8.5

241045-3 51.86 455.28 9.2 PLOT
2410301 53.31 363.51 5.6

2410302 51.86 309.75 5.6

2410303 49.02 396.37 6.6

241001 52.64 197.65

241002 49.09 281.53

241003 53.45 196.54

24JGV-1 49.73 609.13 10.5 ]
2410v-2 52.62 585.39 13 J
2410v-3 50.67 564.62 12.3 |
241001 51.98 597.49 12.4 NOT AS 1
241002 51.96 594.84 12.8 CATASTROPHIC |
241003 . 51.57 507.24 13 FAILED ON EDGE ]
2410w-1 51.89 619.07 10.4 CATASTROPHIC
2410w-2 52.65 612.15 16.8 I
2410w-3 52.13 597.63 12 Pm |

 

195

Table 5.14b Excel 2415: Joint Geometry Study-Warp Direction.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX COMMENTS
ID (mm) (mm) (MPa) (mm)
24JGUU-1 51.05 473.52 14.3 FAILED ON EDGE
24JGUU-2 51.30 493.25 12.5 FAILED ON EDGE
24JGUU-3 52.78 605.03 13 No EDGE
FAILURE
24JGWW-1 51.07 574.18 16
2410ww-2 51.23 553.47 13.2
2410ww-3 50.64 558.95 14.6 PLOT
24WS-1 48.92 2.70 587.30 13.3 S-JOINT
24WS-2 48.38 515.20 11.8 0DEGREES
24WS-3 48.23 529.10 12.5
24ws-4 48.36 450.30 11.0 S-JOINT
24ws-5 49.55 418.30 13.0 45-DEGREFS
24WS—6 49.41 T 374.60 15.6

 

 

196

Table 5.15a Excel 2415: Depth-Of-Joint Survey.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH V-AREA V-ANGLE
1D (mm) (mm) (MPa) (mm) (degree)
24JGV4-1 51.80 2.81 577.97 100 x 50 76
24JGV4-2 50. 83 664. 19 - -
24JGV4-3 50.68 589.66 - -
24JGV3-1 51.26 528.28 75 x 50 71.6
24JGV3-2 52.71 469.24 - -
24JGV3-3 51.58 541.80 — -
24JGV2-1 50.22 547.36 50 x 50 .63
24JGV2-2 49.31 477.83 - -
24JGV2-3 49. 14 488.80 - -
24JGV1-1 51.16 456.54 25 x 50 45
24V1-1 51.11 306.34 SCARFED EDGE
24JGV1-2 51.58 363.84 - -
24V1-2 51.82 284.40 SCARFED EDGE
24JGV1-3 51.89 280.10 - -
24J1-3 52. 19 298.04 SCARFED EDGE

 

 

197

Table 5.15b Excel 2415: Depth-Of-Joint Survey.

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH V-AREA V-ANGLE
ID (mm) (mm) (MPa) (mm) (degree)
24JGW2-1 49.48 532.85 50 x 50 76
24JGW2-2 49.08 458.24 - -
24JGW2-3 48.95 519.88 - -
24JGW1-1 52.00 460.93 25 x 50 71.6
24JGW1-2 51.96 514.29 - -
24JGW1-3 51.75 496.67 - -

 

 

 

 

 

 

 

198

Table 5.16a Excel 2415: Fiber Orientation Study.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX REPAIRED
1D ‘ (mm) (mm) (MPa) (mm) STRENGTH
(MPa)

2490-1 49.11 2.76 343.5 19

2490R-1 51 . 16 9.5 325.47
24902 49. 18 2.81 334.0 19 ,

2490R-2 49.25 8.7 305.45
2490-3 49.06 2.95 337.8 17

2490R-3 49.64 7.5 271.25
24601 49.54 2.75 272 26.8

2460R-1 50.72 18 293.96
2460-2 49.47 2.74 288.5 27

2460R-2 50.33 12 288.80
24603 50.97 2.78 271.0 28.7

2460R-3 51.62 12 261.43
2445-1 47.58 2.90 294.9 33

2445R-1 49.78 10.5 288.57
2445-2 51.24 2.85 259.4 33 .

2445R-2 51.16 12 313.97
2445-3 46.68 2.85 278.3 33

2445R—3 48.00 27.6 337.76
2430-1 47.72 2.85 340.8 23.8

2430R-1 48.88 23 395.56
2430-2 50.31 2.90 356.0 24

2430R-2 51.65 25.4 350.43
24303 49.56 2.84 368.8 26

2430R-3 51 . 10 18 321.43

 

 

199

Table 5.16b Excel 2415: Fiber Orientation Study.

 

 

 

 

 

 

 

 

 

SPECIMEN WIDTH THICK STRENGTH FLEX REPAIRED
ID (mm) (mm) (MPa) (mm) STRENGTH
(MP3)
2401 49.08 2.73 442.7 19
240R-1 50.93 8.5 324.64
240-2 49.06 2.87 439.6 19.4
240R-2 50.66 8.6 281.79
240-3 49.06 2.81 444.0 18
240R-3 51.16 8.0 279.12
24REFO-1 50.96 2.67 434.6 17.3
24REFO-2 51.1 1 2.75 420.0 16.5
24RF90-1 49.11 2.75 370.0 20
24RF90-2 49.07 2.80 369.6 21

 

 

 

 

 

 

 

200

 

 

 

 

 

 

 

 

 

 

 

Table 5.17 Excel 2415: Bending Fracture Layer Study.
SPECIMEN WIDTH THICK STRENGTH FLEX COMMENTS
ID (mm) (mm) (MPa), (mm)
24WRP3- 49.12 2.69 471.5 18.5 -
24WRP3-1 49.21 546.5 8.2 CATASTROPHIC
J FAILURE
24WRP3-2 49.06 2.73 473.5 18.9 -
24WP3-2 49.22 ‘ 556.7 9.4 CATASTROPHIC
FAILURE
24WRP3-3 48.99 2.75 455.6 19.0 -
24WP3-3 49.11 446.8 8.0 CATASTROPHIC
FAILURE
24WRP4-1 51.58 2.87 442.8 18.6 -
24WP4-1 51.60 555.3 8.1 CATASTROPHIC
FAILURE
24WRP4-2 51.35 2.80 437.2 18.4 -
24WP4—2 51.49 585.8 9.5 CATASTROPHIC
FAILURE
24WRP4-3 51.20 2.80 443.7 19.7 -
24WP4-3 51.43 619.2 10.3 CATASTROPHIC
FAILURE
24WRP5-1 49.40 2.78 432.7 18.1 -
24,WP5-1 49.61 800.7 12.5 CATASTROPHIC
FAILURE
24WRP5-2 51.40 2.78 406.8 17.4 -
24WP5-2 51. 14 626. 8 8.9 CATASTROPHIC
FAILURE
24WRP5-3 49.39 2.73 439.8 18.3 -
24WP5-3 50.33 693.2 7.7 CATASTROPHIC
FAILURE

 

 

 

 

 

 

 

201

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5 .18 Excel 8610: Bending Fracture Layer Study.
SPECIMEN WIDTH THICK YIELD ULTIMATE FLEX PATCH
ID (mm) (mm) STRENGTH STRENGTH (mm) LAYERS
(KN) (MPa) TOP,BOTI‘OM
8621 50.92 3.16 974.5 287.6 18.9
8621R 50.99 977.2 288.0 - 2,2
8622 51.18 3.06 894.0‘ 279.9 19.1
8622R 51.28 950.3 296.9 - 2,2
8623 51.3551. 3.04 945.0 298.8 19.3
86-23R 67 920.8 289.3 8.2 2,2
8631 49.25 3.18 1012 304.9 18.2
86-31R 49.27 1272 383.0 9.4 2,3
86-32 49.59 3.14 966.4 296.6 17.5
86-32R 49.86 1538 469.4 8.2 2,3
8633 50.06 3.06 845.6 270.7 16.8
8633R 50.01 1855 594.4 10.3 2,3
86-41 48.79 3.03 485.9 162.8 17.5
86-41R 49.83 1557 510.6 9.5 2,4
86-42 49.22 3.08 547.7 176.0 17.6
86-42R 49.16 1592 512.2 9.8 2.4
86-43 49.28 3.21 979.9 289.5 17.3
86-43R 49.55 1399 411.1 8.3 2,4
8651 49. 12 3.00 808.1 274.3 20.0
8651R 49.98 1635 545.4 10.6 2,5
8652 49.22 3.18 1001 301.7 18.9
8652R 49.50 2046 613.3 9.2 2,5
86-53 49.22 3.13 902.0 280.7 17.6
86-53R 50.11 1546 472.5 7.7 2,5

 

 

 

 

 

 

 

 

 

202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.19 No Damage/Two Patch Layers Per Side Evaluation to Determine the Influence of the Glass
Patches on Non-Separated Specimen.
SPECIMEN WIDTH THICK YIELD ULTIMATE FLEX AVERAGE 5
ID (mm) (mm) STRENGTH STRENGTH (mm) STRENGTH FROM
(KN) 002) (MP!) N0
DAMAGE
OCL-l 51.13 2.98 15601 515.5 11.9 476.12 64
OCL-2 50.95 2.99 1393 458.9 11.6 INCREASE
OCL-3 51.34 2.97 1278 423.4 13.1 OWENS
OCL-4 50.53 _ 2.94 1369 470.3 12.6 CORNING
OCL-S 51.96 2.98 1576 512.5 12.5
24W-1 49.04 2.7 1777 745.8 14.5 675.04 41
24W-2 50.39 2.69 1541 634.1 14.1 INCREASE
24w-3 49.22 2.7 1388 580.4 14.6 EXCEL
24w-4 48.94 2.73 1592 654.9 13.9 2415-WARP
24W-5 49.11 2.76 1895 760.0 15.2
24F-1 49.19 2.79 1385 542.7 14.8 523.98 40
24F-2 49.51 2.84 1230 462.1 14.2 INCREASE
24F-3 50.66 2.76 1243 483.3 14.8 EXCEL
24F-4 49.44 2.89 1442 524.0 13.8 2415-FILL
24F-5 49.18 2.89 1664 607.8 14.3
86:1 49.14 3.14 1530 473.8 14.0 457.10 52
862 49.30 3.17 1546 468.2 15.2 INCREASE
863 49.59 3.22 1474 430.1 14.1 EXCEL
86-4 50.07 3.12 1458 448.8 15.6 8610
865 49.88 3.17 1552 464.6 13.8

 

 

 

 

 

 

 

 

 

 

203

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.1 Three-Point Bending of Adhesive Coupons
SPECIMEN WIDTH THICK MAX. MAX. FLEX COMMENTS
ID (mm) (mm) STRESS LOAD (mm)
(MPa) (MPa)
3M-1 14.54 8.48 62.79 .8752 3 3M-EPOXY
3M-2 11.78 8.55 71.33 .8188 -
3M-3 11.91 8.39 65.10 .7275 3.3
RS-l 12.40 6.33 127.7 .8456 2.3 RESIN SERVICES
RS-2 11.95 6.36 121.2 .7812 2 EPOXY
RS-3 13.23 6.51 103.8 .7758 -
205-l 12.96 8.06 137.6 1.544 3.5 TCC-205-EPOXY
205-2 11.95 7.96 137.5 1.388 3.3
205-3 14.66 7.60 142.2 1.605 3.8
F-l 12.78 9.58 48.76 .7624 2.3 FUSOR-EPOXY
F-2 12.96 9.51 62.54 .9772 2.2
F-3 13.97 8.85 67.73 .9879 2.1
MI 12.80 6.84 120.1 .9584 3 MARASET-EPOXY
M-2 12.87 6.92 90.51 .7436 2.3
M-3 13.93 6.78 101.0 .8617 4.2
DOW-1 12.14 6.03 40.60 .2389 1.2 DOW-VNYLESTER
DOW-2 12.19 5.90 35.13 .1987 .95
DOW-3 14.07 6.03 37.79 .2577 .97
072-1 12.77 6.00 38.98 .2389 1.04 TCC-072-EPOXY
072-2 12.09 5.90 38.28 .2148 .75
072-3 14.18 5.95 56.16 .3758 1.07

 

 

204

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.2 Owens Corning Panels: Adhesive Study.
SPECIMEN WIDTH THICK STRENGTH FLEX COMMENTS
ID (mm) (mm) (MPa) (mm)
FUSOR-1 50.57 2.70 163.85 7.0 HIGH
FUSOR-2 50.76 148.00 8.8 VISCOSITY
FUSOR-3 49.36 161.16 6.4 STRUCTURAL
FUSOR-4 50.69 211.40 5 .24 ADHESIVE
FUSOR-5 52.26 216.67 6.0 WITH GAPS
3M-1 52.92 165.99 8.4 HIGH
3M-2 51.13 97.23 7.5 VISCOSITY 1
3M-3 51.54 227.20 7.8 STRUCTURAL I
3M-4 53.89 266.51 7.4 ADHESIVE
3M-5 56.00 156.86 8.0 WITH GAPS
RS-l 49.63 320.56 7.2 LOW
RS-2 52.65 284.32 5.5 VISCOSITY
RS-3 50.34 288.57 6.5 EPOXY
RS-4 49.58 278.51 5.6 I
RS-5 49.87 320.09 7.6
TCC205-1 51.89 272.50 4.9 LOW
TCC205-2 50.80 324.03 5.0 VISCOSITY
TCC205-3 51.75 312.76 6.3 EPOXY
TCC205-4 50.82 279.33 5.9
TCC205-5 51.71 259.60 4.8
TCCO76-l. 50.46 198.14 4.5 LOW
TCC076-2 50.06 171.03 5.0 VISCOSITY
TCC076-3 49.37 177.92 7.2 POLYESTER
TCC076-4 49.17 169.65 6.3
TCC076-5 49.58 166.00 4.8
MARASET-l 49. 16 207. 85 10. 1 MEDIUM
MARASET-2 49.89 289.03 7.4 VISCOSITY J
MARASET-3 49.74 265.41 7.2 EPOXY |
MARASET-4 48.52 250.56 7.2 I
MARASET—S 50.15 245.65 6.2 J

 

 

 

 

 

 

 

NOTE: FUSOR AND 3M ARE THICKSOTROPIC AND CREATE A NON-UNIFORM
BONDLINE DURING THE REPAIR PROCEDURE. VOIDS AT THE BONDLINE
REDUCED THE OVERALL STRENGTH OF THE REPAIR BY A NEGLIGABLE
AMOUNT.

205

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.3 Owens Corning Panels: Impact Adhesive Study.

I SPECIMEN WIDTH THICK STRENGTH FLEX ENERGY
ID (mm) (mm) (MPa) (mm) (Joules)
OC-4 49.89 2.70 169.39 14.5 24.33
005 49.09 160.21 14.6 26.81
OC-6 48.81 171.24 17.0 26.18
OC-lF 51.48 212.44 19.8 27.83

I OC—2F 52.80 244.81 20.5 28.61

I OC-3F 51.76 219.83 16.3 27.26

OC-73M 52.88 200.55 16 24.72
OC-83M 53.26 193.94 17.7 25.51
OC-93M . 53.54 204.26 18.4 23.13
OC-lOM 49.47 213.29 16.2 27.24
OC-llM 51.73 249.87 16 27.54
OC-12M 50.19 242.13 15.2 25.91
OC-13RS 53.79 210.51 15.8 26.24
OC-14RS 51.20 193.10 15.5 27.04
OC-15RS 51.54 177.90 18.9 26.60
OC-16T2 _ 51.25 191.87 14.5 23.93
OC-17T2 51.89 205.44 16.3 26.82
[ OC-18T2 50.41 221.35 14.6 25.34
OC-19T0 53.16 195.35 16.2 27.27

I 002010 50.43 185.12 16.1 24.26

|| 002110 52.04 200.62 17.3 24.53

[I OC-22Dl 50.01 181.16 16.4 24.92

|| OC-23D2 51.18 179.15 21.8 23.16

|[ OC-24D3 51.23 194.07 15.8 23.07

 

 

 

 

 

 

NOTE: F = FUSOR, M = MARASET, RS = RESIN SERVICES,
T2 = TCC205, T0 = TCCO67, D = DOW DERAKANE.

 

206

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.4a Cumulative Tension Testing Data
SPECIMEN WIDTH THICKNESS MAX. MAX. COMMENTS
ID (mm) (mm) STRESS LOAD
(MPa) (KN)
OCTENI 119.2 18.66 NO REPAIR
OCTEN2 143.0 22.04
OCBJ-lO 50.50 3.02 87.82 13.39 BUTT-JOINT
OCBJ-ll 49.98 3.07 77.38 11.87
0031-14 50.30 3.09 81.80 12.71
OCBJ-IS 50.86 3.00 73.49 11.21
OCSJ-9 52.45 3.02 68.39 10.83 S-JOINT
OCSJ-ll 51.53 3.00 82.82 12.80
OC81-14 52.23 3.07 77.79 12.47
OC81-16 51.81 2.98 73.99 11.42
86TEN1 142.5 22.40 NO REPAIR
86T'EN2 176.4 27.84 ]
8631-6 53.00 3.02 72.87 11.66 BUTT-JOINT I
86131-7 52.43 3.04 75.18 11.98 I
86131-17 51.39 2.97 82.84 12.64
8631-18 50.85 3.00 67.27 10.26
8681-6 53.49 3.09 61.61 10.18 S-JOINT
8681-8 53.39 3.08 85.71 14.09 ‘
8681-17 53.37 3.04 69.11 11.21
8681-18 53.40 3.04 73.75 11.97

 

 

 

 

 

 

207

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.4b Cumulative Tension Testing Data
SPECIMEN WIDTH THICKNESS MAX. MAX. COMMENTS
ID (mm) (mm) STRESS LOAD
(MP8) (KN)
24FI'EN1 244.5 35.49 NO REPAIR
24FBJ-6 48.76 2.81 91.76 12.57 BUTT-JOINT
24FBJ-7 49.21 2.84 83.74 11.70 |
24FBI-17 51.04 2.81 99.66 14.29 |
24FBJ-18 52.03 2.89 74.77 11.24 |
24F81-1 50.78 2.89 70.98 10.40 S-JOINT I
24F81-16 53.00 2.81 81.80 12.18
24F81-17 53.25 2.84 80.03 12.10 I
24FSJ-18 53.86 2.84 70.37 10.75 |
24WTEN1 177.6 24.39 No REPAIR |
24WBJ-7 49.55 2.67 77. 19 10.21 BUTT-JOINT
24WBJ-ll 49.18 2.64 107.2 13.92
24WBJ-12 49.55 2.69 90.50 12.06
24WBJ-15 51.39 2.93 79.12 11.91
24W81-6 53.20 2.55 91.21 12.37 S-JOINT
24WSJ-7 51.54 2.61 71.94 9.68
24w31-8 48.23 2.60 89.18 11.18
24ws1-12 50.58 2.76 84.48 11.79

 

 

 

 

 

 

208

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.5a Out-Of-Plane Curvature Study on Rockwell SMC.
SPECIMEN WIDTH TI THICK MAX. MAX. COMMENTS
ID (mm) (mm) LOAD 0
(KN) (MPa)
R-l 51.56 2.54 174.5 78.71 FLAT PANEL-NO REPAIR
RR-l 902.0 406.8 REPAIRED W/BOND LINE
PRESSURE
R-2 50.95 2.51 327.5 153.1 FLEX AT FAILURE OF
RR-2 1077 503.4 SPECIMEN WAS OVER 1'
R-3 51.79 2.64 416.1 173 ' '
RR-3 1275 530 - -
R-4 51.25 2.43 343.6 170.4 FLAT PANEL-No REPAIR
RR-4 1066 528.5 REPAIRED W/O BOND
LINE PRESSURE
L R-5 51.58 2.41 349 174.8 ' '
RR-5 1034 517.9 ' '
R-6 50.66 2.40 319.5 164.3 ' '-
RR-6 1195 614.4 ' '-
R1-1 57.68 2.47 1248 532.1 CURVATURE ll-NO REPAIR
RRl-l 1256 535.5 REPAIRED I
Rl-2 60.73 2.62 1128 406 ' '
RRl-2 1154 415.3 ' '
R2-1 68.05 2.64 942.3 298.1 CURVATURE 112-NO REPAIR
RR2-1 1468 464.4 REPAIRED
R2-2 72.13 2.51 1146 378.4 - -
RR2-2 1565 516.7 - -
R3-1 56.48 2.70 614.8 224 CURVATURE #3-NO REPAIR
RPS-1 1195 435.5 REPAIRED
R3—2 59.46 2.70 601.3 208.1 ' -
RR3-2 923.5 319.1 - -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.5b Out-Of-Plane Curvature Study on Rockwell SMC.
SPECIMEN WIDTH THICK MAX. MAX. COMMENTS
ID (mm) (mm) LOAD a
(KN) (MPa)

R4-l CURVATURE 174-NO REPAIR
RR4-l REPAIRED

R4-2 60.97 2.90 730.2 213.7

RR4-2 706 206.6

R4-4 54.36 2.56 350.3 350.3
RR4-4 1187 499.9

R5-1 54.96 3.03 303.4 90.22 CURVA'I'URE #5-NO REPAIR
RIG-l 1501 446.3 REPAIRED

R5-2 54.16 3.18 263.1 72.08
RR5-2 1071 293.4

R5-3 49.78 2.98 217.5 76.36
RR5-3 848.3 297.8

R6-1 52. 15 2.81 153 55.75 CURVATURE #6—NO REPAIR
RR6—1 542.3 197.6 REPAIR

R6—2 51.63 3.00 131.6 42.49
RR6-2 477.9 154.3

R6-3 51.67 3.06 174.5 54.11
RR6-3 426.9 132.4

 

 

 

 

 

 

 

 

Table 7.1a

210

OWENS CORNING: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=O.122in, Wm= 1.41,1.39,1.46
REPAIRED DIMENSIONS: THICKNESS=0.142, Wm= 1.42,1.39,1.40
SPECIMEN ID: OCND-B, OCNB-BR: REFERENCE,REPAIRED RESPECTIVELY

ROTATION (°)

REFERENCE
TORQUE (In-lbs)
45

REPAIRED
TORQUE (in-1175)

 

COMMENTS

FAILURE BEGAN BY
CRACKING

ALONG THE SIDES OF THE
SPECIMEN. THE CRACKS
THEN PROPAGATED ACROSS
THE SPF£IMEN SIDES. THE

FAILURE MODE WAS
SEVERE

THROUGH-THE-THICKNESS
DELAMINATION. ALSO, THE

REPAIRED SPECIMEN
FAILED

IN THE SAME ZONE AS THE
REFERENCE SPECIMEN.

211

Table 7.115 OWENS CORNING: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=0.118in, wm= 1.39,1.37,1.4o

REPAIRED DIMENSIONS: THICKNESS=0.147, Wm= 1.41,1.41,1.38
SPECIMEN ID: OCND-C, OCNB-CR: REFERENCE,REPAIRED RESPECTIVELY

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ROTATION (o) REFERENCE REPAIRED COMMENTS
TORQUE (in-lbs) TORQUE (in-lbs)

5 37 41 FAILURE BEGAN BY
. CRACKING

715 55 58 ALONG THE SIDES OF THE
10 71 72 SPECIMEN. THE CRACKS

12.5 84 87 THEN PROPAGATED ACROSS
15 94 99 THE SPECIMEN SIDES. THE
16 98 102 FAILURE MODE WAS

SEVERE
17 103 109 T'HROUGH-THE-TIIICKNESS
18 105 113 DELAMINATION. ALSO, THE
19 107 118 REPAIRED SPECIMEN
mm

20 110 121 INTHESAMEZONEASTHE
21 111 112 REFERENCE SPECIMEN.
22 113 101

23 117 96

24 122 102

25 124 107

26 127 112

27 134 115

28 132 116

29 134 105

30 138 102

31 138 93

32 134 96

33 131 89

34 129 83

35 127 83

40 84 72

44 69 59

48 50 49

50 46 44

 

 

 

 

 

 

Table 7.10

 

212

EXCEL 8610: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=0.115in, Wm= 1.37,1.35,1.41
REPAIRED DIMENSIONS: THICKNESS=0.138, Wm= 1.36,1.35,1.35
SPECIMEN ID: 86ND-A, 86ND-AR: REFERENCE,REPAIRED RESPECTIVELY

(°) REFERENCE REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)
41

SEVERE
THROUGH-THE-

213

Table 7.1d EXCEL 8610: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS =0.123in, wm= 1.35,1.31,1.33
REPAIRED DIMENSIONS: THICKNESS=0.153, wm= 1.33,1.31,1.34
SPECIMEN ID: 86ND-B, 86ND-BR: REFERENCE,REPAIRED RESPECTIVELY

ROT A (0) REPAIRED
TORQUE (in-lbs) TORQUE (in-1b!)

 

Table 7.1e

 

214

EXCEL 2415-FILL: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=0.114in, Wm: 1.33,1.28,1.36
REPAIRED DIMENSIONS: THICKNESS =0.146, w1.:,:= 1.31,1.30,1.31
SPECIMEN ID: 24FND-A, 24FND-AR: REFERENCE,REPAIRED RESPECTIVELY

A (°) REPAIRED
TORQUE TORQUE
FAILURE BY
CRACKING

THE
. THE
THEN AGATED

SEVERE
THROUGH-TIE-

215

Table 7.11 EXCEL 2415-FILL: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=0.113in, Wm: 1.34,l.31,1.33
REPAIRED DIMENSIONS: THICKNESS=0.136, Wm= 1.35,1.32,1.33
SPECIMEN ID: 24FND-B, 24FND-BR: REFERENCE,REPAIRED RESPECTIVELY

ROTATION (°) REPAIRED
TORQUE (in-lbs) TORQUE (in-1b5)

SEVERE
THROUGH-THE-

FAILED
INTHE

 

216

Table 7.1g EXCEL 2415-WARP: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=0.108in, Wm= 1.40,l.39,1.38
REPAIRED DIMENSIONS: THICKNESS=0.143, Wm= 1.43,l.42,l.42
SPECIMEN ID: 24WND-A, 24WND-AR: REFERENCE,REPAIRED RESPECTIVELY

ROTATION (°) REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)
AILURE BY
CRACKING

THEN PROP TED

SEVERE
“ROUGH-THE-

 

217

Table 7.1h EXCEL 2415-WARP: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REFERENCE DIMENSIONS: THICKNESS=0.108in, Wm= 1.35,1.31,1.3o
REPAIRED DIMENSIONS: THICKNESS=0.130, Wm= 1.41,1.39,1.38
SPECIMEN ID: 24WND-B, 24WND-BR: REFERENCE,REPAIRED RESPECTIVELY

ROT A (°) AIRED
TORQUE (in-lbs) TORQUE (in-lbs)
1 FAILURE BEGAN BY
CRACKING

44 THE
SPECIMEN. TIE
THEN PROP TED

SEVERE
THROUGH-THE-

 

218

Table 7.23 OWENS CORNING-BU'I'I‘ JOINT: TORSIONAL TESTING RESULTS, STANDARD

REPAIR REPAIRED DIMENSIONS: THICKNESS =0.120, w,=2.oo,TIIICKNESS=0.121,
W,= 2.00 SPECIMEN ID: OCBJ-12, OCEI-l3: REPAIRED

- -1
TORQUE (HI-lbs) TORQUE (in-lbs)

ROTA (°)
FAILURE BEGAN BY

A

THEN PROPAGATED
THE SPECIMEN. TIE

TO
PATCH FAILURE,
UTILIZATION OF A LARGE
PA

 

Table 7.2b

 

219

OWENS CORNING-S-JOINT: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REPAIRED DIMENSIONS: THICKNESS=0.137, W,= 2.15,THIC1(NESS =0.146, W,=
2.09 SPECIMEN ID: OCSJ-lO, OCSJ-13: REPAIRED

ATION (°) -1 OCSJ-
TORQUE (in-lbs) TORQUE (m-Ibs)

FAILURE BEGAN BY

PA

TO
PATCH FAILURE,
UTILIZATION OF A

DELAMINATION

220
Table 7.2c EXCEL 8610 BUTT-JOINT: TORSIONAL TESTING RESULTS, STANDARD REPAIR

REPAIRED DIMENSIONS: THICKNESS=0.152, W,= 2.06,THICKNESS =0.142, WI:
2.“) SPECIMEN ID: 86BJ-13, 86BI-16: REPAIRED

(°)

TORQUE (in-lbs) TORQUE (in-lbs)

FAILURE BEGAN BY RIPPING

PATCH FAILURE,
UTILIZA OF A LARGE
A

 

221
Table 7.2d . EXCEL 8610 S-JOINT: TORSIONAL TESTING RESULTS, STANDARD REPAIR

REPAIRED DIMENSIONS: THICKNESS=0.144, W,= 2.12,THICKNESS =0.142, W,=
1.98 SPECIMEN 1D: 86SJ-9, 8651-16: REPAIRED

(°)

TORQUE (111-le) TORQUE (in-lbs)

FAILURE BEGAN BY RIPPING

PA

TIEN PROP TED
TIE .
W
TO
PATCH FAILURE, THOUGH
UTILIZATION OF A LARGE
OF PA AREA

 

222

Table 7.28 EXCEL 2415-FILL BUTT-JOINT: TORSIONAL TESTING RESULTS, STANDARD
REPAIR REPAIRED DIMENSIONS: THICKNESS=0.134, W,=2.(D,THICIO\IESS=O.137,
W,= 2.(X) SPECIMEN ID: 24FBJ-15, 24FBJ-16: REPAIRED

ROT A (°) 15 FBI-16 COMMENTS
TORQUE (in-lbs) TORQUE (in-lbs)
FAILURE BEGAN BY

 

223

Table 7 .2f EXCEL 2415-FILL S-JOINT: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REPAIRED DIMENSIONS: THICKNESS=0.134, W, = 2.07,THICKNESS=0.129, W,=
2.02 SPECIMEN ID: 24FSJ-11, 24FSJ-15: REPAIRED

ROTA (°) -1 1 1 COMMENTS

TORQUE (in-lbs) TORQUE (in-lbs)

FAILURE BEGAN BY RIPPING

PA

'10
PATCH FAILURE, THOUGH
A
A

 

224

Table 7,2g EXCEL 2415-WARP BUTT-JOINT: TORSIONAL TESTING RESULTS, STANDARD
REPAIR REPAIRED DIMENSIONS: THICKNESS=0.129, W,=1.94,TIIICKNESS=O.141,
W,= 1.94 SPECIMEN ID: 24WBJ-13, 24WBJ-14: REPAIRED

(9) WBJ-l 4WBJ-14
TORQUE (in-lbs) TORQUE (in-lbs)

F

 

Tabi

Table 7.2h

 

225

EXCEL 2415-WARP S-IOINT: TORSIONAL TESTING RESULTS, STANDARD REPAIR
REPAIRED DIMENSIONS: THICKNESS=0.132, W,= 2.00,TIIICKNESS=0.131, W,=
21X) SPECIMEN ID: 24WSJ-13, 24WSI-14: REPAIRED

A'I'ION (°) WSJ-l 24WSJ-14
TORQUE (in-lbs) TORQUE (in-lbs)
40 FAILURE BEGAN BY RIPPING

TIEPA TIE

THEN

226

Table 7.3a OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.120, W,,,,,= l.22,1.20,l.23
SPECIMEN ID: OCT-1, REPAIRED WITH (3X6) DENSITY NYLON THREAD

(°) REFERENCE REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)

 

227

Table 7.3b OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.118, w,.,,,= 1.21,1.20,1.26
SPECIMEN ID: OCT-2, REPAIRED WITH (4X7) DENSITY NYLON THREAD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ROTATION (o) REFERENCE REPAIRED COMMENTS
TORQUE (in-lbs) TORQUE (in-lbs)
5 34 22 FAILURE BEGAN BY
- CRACKING
7.5 T49 32 OF THE MATRIX ALONG THE
10 61 40 EDGE OF THE SPECIMEN.
12.5 68 48 THEN PROPAGATED ACROSS
15 72 56 THE SPECIMEN. THE
16 75 57 FAn.URE MODE WAS DUE
To
17 80 60 SEVERE DELAMINATION OF
18 84 63 THE MATRIX. THE
19 81 66 REPAIRED SPECIMEN
FAILED
20 84 68 IN THE SAME ZONE.
21 89 71
22 92 75 STITCHING Is CRITICAL
23 95 77 ALONG BOTH EDGES
24 97 79
25 98 81
26 102 82
27 106 84
28 106 84
29 108 86
30 108 84
31 109 81
32 105 82
33 105 78
34 90 75
35 88 72
36 ' 86 67
37 80 66
4o 71 56
50 31 38

 

 

 

 

 

 

228

Table 7.30 OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.120, W,.2.,= 1.24,1.20,1.24
SPECIMEN ID: OCT-3, REPAIRED WITH (2X4) DENSITY NYLON THREAD

ROTATION (°) REFERENCE REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)
S FAILURE BEGAN BY
CRACKING
MATRIX
TIE

TED
. TIE

 

229

Table 7.3d OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.120, W,.2_3= 1.14,l.12,l.12
SPECIMEN ID: OCT -4, REPAIRED WITH (3X6) DENSITY STEEL THREAD

ROTA (°) REFERENCE REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)
1 AILURE BEGAN BY
CRACKING

TIE MATRIX ALONG TIE

TIEN PROP TED
. TIE

TO
SEVERE DELAMINATION OF
THE MATRDI. TIE

FAILED
INTIESAME

 

230

Table 7.3e ° OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.120, w,.,,,= 1.20,1.19,1.20
SPECIMEN ID: OCT-6, REPAIRED WITH (4X7) DENSITY STEEL THREAD

ATION (°) REFERENCE REPAIRED COMMENTS
TORQUE (in-lbs) TORQUE (in-lbs)
5 35 ' FAILURE BEGAN BY
CRACKING

1 OFTIEMATRDI TIE

OF TIE .
TIEN PROP TED ACROSS
THE . TIE

 

231

Table 7.31 OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.120, w,.,,,= 1.27,1.22,1.24
SPECIMEN ID: OCT-7, REPAIRED WITH (2X4) DENSITY STEEL THREAD

ROTATION (°) REFERENCE REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)
5 FAILURE BEGAN BY
CRACKING

.5 TIE MATRIX

THEN PROPAGATED
TIE

TO
DELAMINATION OF
TIE MATRIX. TIE

FAILED
IN TIE SAME ZONE.

 

232

Table 7.3g OWENS CORNING: TORSIONAL REPAIR STUDY, STITCHING REPAIR
REFERENCE DIMENSIONS: THICKNESS = 0.120, W,.2.,= l.2l,1.19,1.25
SPECIMEN ID: OCT-8, REPAIRED WITH 4/WIRE GRID STEEL STRANDS PER SIDE

ROT A (°) REPAIRED
TORQUE (in-lbs) TORQUE (in-lbs)
21 FAILURE BEGAN BY
CRACKING

DELAMINATION OF
TIE MATRIX. TIE

FAILED
INTIESAME

 

233

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 7.4 Cumulative S-Joint Results for Three-Point Bending Repair
SPECIMEN WIDTH THICK MAX. MAX. FLEX COMMENTS
ID (mm) (mm) STRESS LOAD (mm)
(MPa) (KN)
OCSJ-l 51.44 3.1!) 358.4 1.106 O-DEGREES
OCSJ-2 51.68 3.02 417.0L 1.310
OCSJ-3 51.14 3.02 423.0 ‘ 1.315
OCSJ-4 50.62 2.89 427.6 1.205 16 45-DEGREES
OCSJ-5 50.46 3.10 352.1 1.138 11
OCSJ-6 51.25 3.06 313.9 1.1134 10
OCSJ-7 98.62 3.06 285.3 1.756 12 90-DEGREES
OCSJ—B 99.18 3.06 268.5 1.662 12
OCSJ-9 99.26 3.06 260.1 1.611 13
24FSJ-4 51.70 2.87 533.4 1.514 11 O-DEGREES
24FSJ-5 51.77 2.88 456.0 1.305 9.5
8681-10 54.47 3.09 412.0 1.428 9.8
86SJ-11 52.70 3.11 427.4 1.452 11
1 86SJ-15 51.65 3.08 467.9 1.528 12
24WSJ-1 48.92 2.70 587.3 1.396 13
24WSJ-2 48.38 2.70 515.2 1.211 12
24WSJ-3 48.23 2.70 529.1 1.240 12
24WSJ-4 48.36 2.70 450.3 1.058 1 1 45-DEGREES
24WSJ-5 49.55 2.70 418.3 1.007 13
24WSJ-6 49.41 2.70 374.6 .8993 16

 

 

 

 

 

 

 

 

 

234

I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 7.5a Tension Layer Study for S-, Butt-, and WW-Joints

SPECIMEN WIDTH' THICK‘ WIDTH2 THICK2 STRENGTH MAX NOTE
ID (mm) (mm) (mm) (mm) (MPa) LOAD
(KN)

OCBJL-l 51.37 2.90 52.43 4.39 160.5 24.40 4 PLY

PER

SIDE

OCBIL-Z 50.66 3.04 50.68 3.73 73.88 11.38 2 PLY
OCB1L-3 51.38 3.06 51.58 3.53 76.11 12.01

OCBJL-4 51.35 3.04 51.71 3.79 98.94 15.55 3 PLY
OCBJL-5 50.74 3.04 50.69 4.20 129.3 19.92

OCBJL-6 51.32 3.09 52.29 4.06 104.4 16.87 GAP

3686

LINE

OCB1L-7 50.69 3.06 51.56 4.08 129.2 20.38 4 PLY
OCBJL-8 51.56 3.03 51.84 4.57 157.4 24.72
OCBJL-9 51.35 3.08 51.51 4.45 122.9 19.50

OCSJL-IS 50.88 2.94 52.65 3.64 75.99 11.76 2 PLY

PER

SIDE
OCSJL-18 50.58 2.92 51.76 3.41 80.08 12.10
OCSJL-19 50.46 2.97 53.21 3.63 76.14 12.03
OCSJL-l 51.47 3.07 52.29 3.77 87.86 14.10

OCSIL-A 51.42 3.02 52.67 4.24 107.8 17.15 3 PLY
OCSIL-B 51.24 3.06 52.87 3.81 113.0 18.27
OCSJND15 49.87 2.95 51.84 3.96 124.0 18.96

OCSJL-7 51.52 3.04 52.32 4.64 135.3 21.52 4 PLY
OCSJL-8 51.94 3.04 53.73 4.26 114.4 18.68

OCSIL-9 51.05 3.08 52.47 4.09 138.6 22.40 CAT.

FAIL

 

 

 

 

 

 

 

 

 

NOTE: ALL BUTT-JOINTS AND WW-JOINTS EXIBITED CATASTROPHIC FAILURE, TIE
S-JOINTS WERE NON-CATSTROPIIIC EXCEPT FOR SOME OF THE 4 PLY SPECIMEN

 

235

 

 

 

 

 

 

 

 

 

Table 7.5b Tension Layer Study for S-, Butt-, and WW-Joints
SPECIMEN WIDTHl THICK' WIDTH2 THICK2 STRENGTH MAX NOTE
ID (mm) (mm) (mm) (mm) (MPa) LOAD
(KN)
OCW1L-2 50.27 3.02 51.05 3.84 115.9 22.71 2 PLY
PER
» SIDE
OCW1L-3 51.53 3.06 52.60 3.59 128.1 20.62
OCWIL-4 51.00 3.04 51.90 3.47 147.1 23.21
OCWIL-S 50.87 3.04 51.96 4.27 146.3 23.10 3 PLY
OCWIL-6 51.06 3.09 52.60 4.40 192.8 31.33
OCWJL-7 50.67 3.02 51.61 3.87 163 25.40
OCWIL-S 51.44 3.06 52.22 4.08 211.1 33.73 4 PLY
OCW1L-9 51.06 3.06 52.62 4.48 226.4 36.44

 

 

 

 

 

 

 

 

 

NOTE: ALL BUTT-JOINTS AND WW-JOINTS EXIBITED CATASTROPHIC FAILURE, TIE S-

JOINTS WERE NON-CATSTROPHIC EXCEPT FOR SOME OF TIE 4 PLY SPECIMEN

 

236

All of the Values in the Table are Average Values.

Table 7.6a Comparison of All Materials for Each Testing Procedure.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OWENS EXCEL EXCEL 2415 EXCEL 2415
CORNING 8610 FILL- WARP-
DIRECI'ION DIRECTION
TESTING MAX MAX MAX MAX MAX MAX MAX MAX
ID 0 Load 0 Load 0 Load 0 Load
(MPa) (KN) (MP8) (KN) (MPa) (KN) (MPa) (KN)
TENSION
NO DAMAGE 131.1 20.35 159.5 25.14 178 24.39 245 35.49
BUTT-JOINT 80.12 12.30 74.50 11.64 88.5 12.03 87.5 12.45
S-JOINT 75.75 11.88 72.5 11.86 84.2 11.26 75.8 11.36
COMPRESSION
NO DAMAGE - 24.83 - 27.4 - 28.93 - 32.49
BUTT-JOINT - 25.81 - 24.9 - 23.14 - 23.32
S-JOINT - - - - - - - -
BUCKLING

NO DAMAGE 54.40 7.83 53 7.80 52 74.05 69.2 9.39
BUTT-JOINT 87.42 13.26 121.5 20.5 98.4 12.94 99 12.37
S-JOINT 134.7 20.47 108.5 15.49 108 14.99 130.3 17.76
TORSION TORQ TORQ TORQ TRQ

L in-lbs in-lbs in-lbs in-#
NO DAMAGE - 125.5 - 232 - 169 - 203.5
REPAIRED — 127.5 - 115 - 103.5 - 124.5
BUTT-JOINT - 141.5 164.5 - 150.5 - 160
S-JOINT 210.5 182 ‘ - 175 - 153.5

 

 

Table 7 .6b

237

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Comparison of All Materials for Each Testing Procedure.
All of the Values in the Table are Average Values.
OWENS EXCEL EXCEL 2415 EXCEL 2415
CORNING 8610 FILL- WARP-
DIRECTION DIRECTION
TESTING MAX MAX MAX MAX MAX MAX MAX MAX
1]) 0 Load 0 Load 0 Load 0 Load
(MPa) (KN) (MP3) (KN) (MPa) (KN) (MPa) (KN)
FATIGUE
CYCLES
NO DAMAGE
100 278.6 0.671 299.1 1.001 362 1.494 484.1 1.272
1,000 232.2 0.695 302.7 1.032 377 0.976 501.3 1.356
10,000 247.9 0.751 293.7 1.002 372 1.003 531 1.408
BUTT-JOINT
100 300.9 0.885 281.6 0.872 350 0.863 335 0.710
1,000 294.6 0.899 271 0.862 332 0.859 287.1 0.652
10,000 156.7 0.493 237.4 0.768 300 0.804 264 0.681
S-JOINT
100 416.8 1.285 438.1 1.457 402 1.176 568 1.451
1,000 459.8 1.477 310.1 1.076 468 1.234 434 1.179
10,000 416.1 1.340 441.9 1.548 349 0.992 437 1.027

 

 

 

 

 

 

 

 

 

 

 

LIST OF REFERENCES

LIST OF REFERENCES

M155
1. Engineered Materials Handbook, Vol. #1, Composites, ASM Inter., August 1989.

2. Composite Repairs, ed. Henry Brown, SAMPE Monograph, No. 1, 1985.

3. Baker, A.A. and Jones, R., Bonded Repair of Aircraft Structures, Martinus Nijhoff
Publishers, 1988.

4. Bonding and Repair of Composites, ed. John Herriot, Butterworths, 1989.

5. ”Joining of Composite Materials,” Kedward, K.T., Ed., ASTM STP 749, PCN =
04-749000-33, April, 1980.

6. "Composites," The International Journal of the Science and
Technology of Reinforced Materials, Butterworth-Heinemann, Vol. 22, No. 2, March
1991.

m

7. ”Field-Level Repair Materials and Processes," Stone, Robert H., 28th Nat. SAMPE
Symp., 1983.

8. ”Repair of Damaged Composite Materials,” Zimmerman, K., Liu, D., Proc. 1992
Joint ASM/ASC Inter. Conf. on Composites, with Japan, Orlando, FL, June 1992.

9. ”Further Developments in the Design and Analysis of Adhesively-Bonded Structural
Joints,” Hart-Smith, L.J., Joining of Composite Materials, ASTM STP 749, K.T.
Kedward, Ed., American Society for Testing and Materials, 1981, pp. 3-31.

10. "From 49-1 to HX1567-Development of a Low Energy Composite Repair
System," Lee, E, Brinkerhoff, S., McKinney, 8., Hexcel Corp., 35th Inter. SAMPE
Symp., April, 1990.

11. ”Repair Adhesives: Development Criteria for Field Level Repair, " Cichon,
,M.J.,Hysol Aerospace Products, 34th Inter. SAMPE Symp., May, 1989.

12. ”Rapid Low-Temperature Cure Patching System for Field Repair,” Sivy, G.T. , ITW
Adhesive Systems, 34th Inter. SAMPE Symp., May, 1989. .

238

239

13. "Composite Repair Concepts for Depot Level Use," Labor, James D., 26th Nat.
SAMPE Symp., 1981.

14. " Field-Level Repair Materials and Processes, ” Stone, Robert H. , 28th Nat. SAMPE
Symp., April 12-14, 1983.

15. " Field Repairs For The AH-l Composite Main Rotor Blade," Falcone, Alfred S.,
28th Nat. SAMPE Symp., April 12-14, 1983.

16. " Repair Procedures For Composite Sinewave Substructure, " Labor, James D. , 16th
Nat. SAMPE Tech. Conf., Oct. 9-11, 1984.

17. "Bolted Field Repair of Graphite/Epoxy Wing Skin laminates,” Bohlmann, R.E.,
Renieri, G.D., Libeskind, M., Joining of Composite Materials, ASTM STP 749, 1980.

18. " Acousto—Ultrasonic Evaluation of Impact-Damaged Graphite/ Epoxy Composites, "
Green, James E., Rodgers, John, 27th Nat. SAMPE Symp., May 4-6, 1982.

19. ”Nondestructive Testing of Composite Bonds,” Cawley, P, Bonding and Repair of
Composites, 1989.

20. ”A Nondestructive Method of Evaluating Adhesive Bond Strength in Fiberglass
Reinforced Plastic Assemblies, " Chapmen II, G.B. , Joining of Composite Materials,
ASTM STP 749, 1980.

21. Effect of Manufacturing Defects and Service-Induced Damage on the Strength of
Aircraft Composite Structures," Garrett, R.A., 28th Nat. SAMPE Symp. April 1983.

22. "A Review of Defect Types and Nondestructive Testing Techniques for Composites
and Bonded Joints,” Cawley, P., Adams, R.D., Bonding and Repair of Composites,
1989.

23. "The Study on the Composite-Patching Repairs for Metallic Aircraft
Structures,” Ong, C.L., Shen, S.B., Aeronautical Research Lab/AIDC, 36th Inter.
SAMPE Symp., April 1991. '

24. ”The Competitive Position of Selected Composite Fabrication Technologies for
Automotive Applications,” Krolewski, S., Busch, J ., IBIS Associates, Inc., 35th Inter.
SAMPE Symp., April 1990.

25. ”Adhesive-Bonded Composite-Patching Repair of Cracked Aircraft Structures, "
Ong., C.L., Shen. S., 34th Inter. SAMPE Symp., May 1989.

26. ”Repairs to Damage Tolerant Aircraft,” Swift, T. , Federal Aviation Administration,

240
Inter. Symp. on Structural Integrity of Aging Airplanes, Atlanta GA, March 1990.

27. ”Development of Repair Procedures for Graphite/Epoxy Structures on Commercial
Transports,” Stone, Robert H., 26th Nat. SAMPE Symp., April 28-30, 1981.

28. "Demonstration of Repairability and Repair Quality on Graphite/Epoxy Structural
Subelements," Knauss, James F. , Stone, Robert H. , 27th Nat. SAMPE Symp., May 4-6,
1982.

29. ”Repairing The Damage," Armstrong, K.B., Bonding and Repair of Composites,
1989.

30. ”Analysis of the Joggle-Lap Joint for Automotive Applications,” Givler, R.C.,
Pipes, R.B., Joining of Composite Materials, ASTM STP 749, 1980.

31. "Crack Patching: Design Aspects,” Jones, R. , Bonded Repair of Aircraft Structures,
pgs 49-76, 1988.

32. ”Theoretical Analysis of Crack Patching," Rose, L.R.F., Bonded Repair of Aircraft
Structures, pgs 77-106, 1988.

33. ”Crack Patching: Experimental Studies, Practical Applications," Baker, A.A.,
Bonded Repair of Aircraft Structures, pgs 107-174, 1988.

34. ”High Performance Epoxy Adhesives," Behm, D.T., Clark, E., Ciba-Geigy Corp.,
34th Inter. SAMPE Symp., May 1989.

35. ”Damage Tolerance of Laminated Composites Containing A Hole and Subjected to
Tension or Compression,” Chang, F.K., Lessard, L., Chang, K.Y., Liu, S., 34th Inter.
SAMPE Symp., May, 1989.

36. ”Viscoelastic Relaxation in Bolted Thermoplastic Composite Joints," Schmitt, R.R. ,
Boeing Military Airplanes, Horn, W.J., 35th Inter. SAMPE Symp., April 1990.

37. ”Induction Heating Development for Aircraft Repair,” Border, J. , Salas, R., Black,
M., PDA Engineering, 35th Inter. SAMPE Symp., April 1990.

38. ”An Apparatus to Prepare Composites for Repair,” Westerman, E.A., Roll, P.E.,
Boeing Advanced Systems, 34th Inter. SAMPE Symp., May 1989.

39. ”The Repair of Thermoplastic Composites After Impact,” Ong, C.L., Sheu, M.F.,
Liou, Y.Y., 34th Inter. SAMPE Symp., May 1989.

40. "Repair Technology for Thermoplastic Aircraft Structures (REPTAS),"

241
Heimerdinger, M.W., Northrop Corp., Aircraft Division, Hawthorn, CA 90250-3277.

41. ”Analysis and Repair of Damaged Composite Laminates," Bair, D.L., Hudson,
P.O., 36th Inter. SAMPE Symp., April 1991.

42. "Resin Requirements for Successful Repair of Delaminations," Russell, A.J.,
Bowers, C.P., Defense Research Establishment Pacific, 36th Inter. SAMPE Symp., April
1991.

43. "Tensile Strength of Stitching Joint in Woven Glass Fabrics,” Lee C., Liu, D.,
ASME, May 1989.

44. ”Impact Energy Absorption of Continuous Fiber Composite Tubes, " Wickliffe,
L.E., Schmueser, D.W., General Motors Research Lab, ASME Annual Meeting, Paper
no. 85-WA/Mats-11, November 1985.

45. "Advanced Composite Repair—-Recent Developments and Some Problems,” Myhre,
S.M., 26th Nat. SAMPE Symp., April 28-30, 1981.

46. " Constant Temperature Heaters For the Repair of Composite Structures, " Walty,
Robert J ., 30th Nat. SAMPE Symp., March 19-21, 1985.

47. " Elevated Temperature Repairs of Advanced Composite Structures, " Roberts,
Larry, 30th Nat. SAMPE Symp., March 19-21, 1985.

48. " Improved Resins For Wet Layup Repair of Advanced Composite Structure, " Scott,
James R., Mashiba, Charlene K., 16th Nat. SAMPE Tech. Conf., Oct. 9-11, 1984.

49. " Repair Investigation of Bismaleimide Structure,” Knight, Ronald C., 28th Nat.
SAMPE Symp., April 12-14, 1983.

50. " Preliminary Evaluation Of Large Area Bonding Processes For Repair of
Graphite/Polyamide Composites,” Deaton, Jerry W., Mosso, Nancy A., 28th Nat.
SAMPE Symp., April 12-14, 1983.

51. " Structural Repair Systems For Thermoplastic Composites,” Weider, S.M., Lause,
H.J., Fountain, R., 16th Nat. SAMPE Tech. Conf., Oct. 9-11, 1984.

52. " Development of Large-Area Damage Repair Procedures For The F-14A Horizontal
Stabilizer Phase I-Repair Development, " Mahon, John, Candella, James, 26th Nat.
SAMPE Symp., April 28-30, 1981.

53. " Repair Of Post-Buckled Advanced Composite Fuselage Structure,” 28th Nat.
SAMPE Symp.,‘April 12-14, 1983.

242

54. " Surface Treatments for Bonded Repair of Metals, " Reinhart, Theodore, Bonded
Repair of Aircraft Structures, pgs 19-30, 1988.

55. ”Design and Analysis of Bonded Repairs for Metal Aircraft Structures, " Hart-Smith,
L.J., Bonded Repair of Aircraft Structures, pgs 31—48, 1988.

56. "Repair of Composite Aircraft,” Williams, J .G., Donnellan, T.M., Traboeco, R.B.,
Repair of Aircraft Structures, pgs 175-211, 1988.

57. ”Adhesives for Composite Joints, " Lees, W.A. , Bonding and Repair of Composites,
1989.

58. ”Frictional Coupling for Bonding of Composites,” Gradin, P.A. , Boning and Repair
of Composites, 1989.

59. "Bondability of Carbon Fibre Composites," Parker, B.M., Bonding and Repair of
Composites, 1989.

60. "Structural Adhesive Bonding of Thermoplastic Fibre Composites,” Kinloch, A.J. ,
Kodokian, G.K.A., Bonding and Repair of Composites, 1989.

61. ”Durability of Composite Joints Using Cold-Setting Adhesives,” Matthews, F.L.,
Kinloch, A.J., John, S.J., Bonding and Repair of Composites, 1989.

62. ”Low Energy Cured Composite Repair Systems,” Lee, R, Brinkerhoff, S,
McKinney, S., Bonding and Repair of Composites, 1989.

63. "A New Low-Temperature Rapid Curing Composite Material for Structural Repair, "
Subrahmanian, K.P., Davis, J.W., Marteness, E.A., Bonding and Repair of Composites,
1989.

64. "Feasibility of Welding Thermoplastic Composite Materials, " Taylor, N .S. , Jones,
S.B., Bonding and Repair of Composites, 1989.

65. "Thermoplastic Interlayer Bonding of Aromatic Polymer Composites, " Cogswell,
F.N., Meakin, P.J., Smiley, A.J., Harvey, M.T., Booth, C., Bonding and Repair of
Composites, 1989.

66. "Adhesive Bonded Joint Between a Fibre Composite Wing Blade and the Steel Hub
of a Wind Turbine," Lystrup, A., Anderson, S.I., Bonding and Repair of Composites,
1989.

67. ”FDEMS Sensors:In-Situ Monitoring of Adhesive Bondline and Resin Cure During
Repair," Kranbuehl, D.E., Eichinger, D., Hamilton, T., Clark, R., Koury, J ., Freeman,

243

T., Bonding and Repair of Composites, 1989.

68. "Further Developments in the Design and Analysis of Adhesively-Bonded Structural
Joints," Hart-Smith, L.J., Joining of Composite Materials, ASTM STP 749, 1980.

69. "Improving Composite Bolted Joint Efficiency by Laminate Tailoring, " Eisenmann,
J .R., Leonhardt, J.L., Joining of Composite Materials, ASTM STP 749, 1980.

70. "Bolt Bearing Fatigue of a Graphite/Epoxy Laminate,” Crews, J .H., Joining of
Composite Materials, ASTM STP 749, 1980.

71. "Failure Analysis of Composite Laminates with a Fastener Hole," Soni, S.R. ,
Joining of Composite Materials, ASTM STP 749, 1980.

72. "Static and Fatigue Strength of High Load Intensity Steel Hybrid Composite Bonded
Scarf Joints," Macander, A.B., Joining of Composite Materials, ASTM STP 749, 1980.

73. "Design Data for Graphite Cloth Epoxy Bolted Joints at Temperatures up to 450K, "

Bailie, J.A., Duggan, M.F., Bradshaw, N.C., McKenzie, T.G., Joining of Composite
Materials, ASTM STP 749, 1980.

REPAIR M QAQ

74. Boeing 767-200 Structural Repair Manual, Boeing Commercial Aircraft Co., P.O.
Box 3707, Ms73-43, Seattle, WA, U.S.A.

75. Boeing 737-300 Structural Repair Manual, Boeing Commercial Aircraft Co., P.O.
Box 3707, MS73-43, Seattle, WA, U.S.A.

76. "FRP Truck—Trailer Repair, " Owens/Coming Fiberglas Repair Manual, 1975.

77 . "Fiber Glass Repairability," A Guide and Directory to the Repair of Fiber Glass
Commercial Fishing Boats,” Owens/Coming Fiberglas, 1972. '

78. ”Repair of Fiberglas Boats,” Owens/Corning Fiberglas, 1972.

W

79. ”Method of Repairing Discontinuity in Fiberglass Structures," Fletcher, J .C. , US
Patent, Patent # 3,814,645, February 1974.

80. "Process for Repairing Body Parts on Vehicles or the Like," Griffith, F.L.F., U.S.

244
Patent, Patent # 4.430,133, February 1984.

81. "Method of Repairing Ballistic Damage," Miller, R.W., U.S. Patent, Patent #
4,517,038, May, 1985.

82. "Method for Repairing Glass Fiber Reinforced Plastic Parts and Product, " Faber,
D.D., Holmes, R.G., Varano, J.J., U.S. Patent, Patent # 4,409,270, October 1983.

83. "Repair Laminate Sheet for Acrylic and Fiberglass Tubs, " Thomas, W. , Lucatuorto,
P., U.S. Patent, Patent # 4,511,621, April 1985.

84. "Conductive Joint Seals for Composite Aircraft," Bannink Jr., E.T., U.S. Patent,
Patent # 4,556,592, December 1985.

85. "Multiple Lap-Joint for Thermoplastic Laminates," Ritter, J .D., U.S. Patent, Patent
# 4,564,543, January 1986.

86. "Aerospace Structure Having a Cast—In-Place N oncompressible Void Filler, "
Scollard, C.Y., U.S. Patent, Patent # 4,861,643, August 1989.

W

87. "Static and Vibration Analysis of Laminated Composite Beams with an Interlaminar
Shear Stress Continuity Theory," Lee, C.Y., Liu, D., Lu, X, Inter. Journal for
Numerical Methods in Engineering, vol. 33, 409-424, 1992.

88. "The Crush Behavior of Pultruded Nbes at High Strain Rates," Thornton, P.H.,
Ford Motor Company, Journal of Composite Materials, Vol. 24, June 1990.

89. "The Strength of Composite Repair Patches: A Laminate Analysis Approach,"
Matthews, F.L., Robson, J.E., Kinloch, A.J., Journal of Reinforced Plastics and
Composites, Vol. 11, July 1992.

90. "Energy Absorption of Composite Materials, " Farley, G.L. , Journal of Composite
Materials, Vol. 17, May 1983.

91. "Energy Absorption in Composite Tubes," Thornton, P.H., Ford Motor Company,
Journal of Composite Materials, Vol. 16, pp 521-545, November 1982.

92. "Energy Absorption in Composite Structures," Thornton, P.H., Ford Motor
Company, Journal of Composite Materials, Vol. 13, pp. 247-260, July 1979.

245
NASA IEEIQAL MEMOS

93. "Adhesively Bonded Double-Lap Joints," Hart-Smith L.J., Nasa Langley Research
Centre, Rept. Nasa CR-112235, January 1973. ‘

94. "Simplified Procedures For Designing Adhesively Bonded Composite Joints,"
Chamis, C.C., Murthy, P.L.N., Lewis Research Center, Tech. Memo #102120, 1989.

95. "Preliminary Design of Adhesively Bonded Composite Joints," Chamis, C.C.,
Murthy P.L.N., Nasa Lewis Research Center, Nasa Tech Briefs, October 1992.

96. "Damage-Tolerant Composites Made by Stitching Carbon Fabrics," Dow, M.B. ,
Nasa Langley Research Center, Smith, D.L., Planning Research Corp., Nasa Tech
Briefs, November 1992.

97. "Stiff, Strong Splice for a Composite Sandwich Structure," Nasa Ames Research
Center, Nasa Tech Briefs, ARC-11743.

98. "The effects of Crushing Speed on the Energy-Absorption Capability of Composite
Material," Nasa Langley Research Center, Nasa Technical Memo 89122, March 1987.

W

99. RE. Robertson, "A Repair Methodology for Fiber-Reinforced Composite
Structures, " Final Report submitted to Michigan Materials and Processing Institute, July
1992.

100. Dahsin Liu, C.Y. Lee, X. Lu, "Repairability of Damaged Structural Composites,"
an Annual Report submitted to Michigan Materials and Processing Institute, September
1991.

101. K. Zimmerman and D.Liu, "An Experimental Investigation on the Repairability

of Damaged Structural Composites, an Annual Report submitted to Michigan Materials
Processing Institute, September 1992.

ADDITIONAL PAM

102. D. Liu, "Delamination Resistance in Stitched and Unstitched Composite Plates
Subjected to Impact Loading," Journal of Reinforced Plastics and Composites, 9(1),
pp59-69, 1990.

103. Lee, C.Y., and Liu, D., "An Interlaminar Stress Continuity Theory for Laminated

246
Composite Analysis," Computers and Structures, 42(1), pp 68-78, 1992.

it: