an)? .
a Jan , .z ,.
.wwmmfi. by}

Q..."

wk.

.
Y!

«w.

1w?
> 5
2i). .
c ‘13.:2
2...... .

‘5»... 1v .0
Quflh

..

. .

: :2 .
:4: i. . 2

1- I 5.

3.... ”Wu «an ,

:5..- 1.135;

:vM-i
’0‘
.!

 

 

r... ... 3.13.? , 1.15...

 

 

.LIBRARY
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

MANUFACTURING COMPOSITE ARCHES AND DOMES WITH
VARTM FOR HIGH ENERGY ABSORPTION

presented by

PATRICK ADAM ROPP

has been accepted towards fuifillment
of the requirements for the

Master of Science degree in Mechanical EnmeerinacL

 

 

@4495:

Major Professor/’3’ Signature
5/Z( /o 7

Date

 

MSU is an aflinnative-aca'on, equal-opportunity employer

4 _.—.-.--._ 4 —

ll-D-I-l-I-t-I-I-O-1-0-0-I-t-o-l-o-n-n--.—.—t-<-.-v—t- -- —

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/C|RClDateDue.indd-p.1

MANUFACTURING COMPOSITE ARCHES AND DOMES WITH VARTM FOR

HIGH ENERGY ABSORPTION

By

Patrick Adam Ropp

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Mechanical Engineering

2007

ton]
of \'
don]

fiber

ABSTRACT
MANUFACTURING COMPOSITE ARCHES AND DOMES WITH VARTM FOR
HIGH ENERGY ABSORPTION
By

Patrick Adam Ropp

Advanced composites allow engineers the benefits of creating low density, high
stiffiiess, high strength and high energy absorption capability in various structural
components in today’s military and civilian vehicles. This study focused on the process
of Vacuum Assisted Resin Transfer Molding (VARTM) for manufacturing arched and
domed composite structures. The composite laminates consisted of continuous E-glass
fibers and an epoxy matrix combining a thermosetting Bisphenol-A/Bisphenol-F resin
with a modified aliphatic polyamine hardener. An optimal manufacturing process was
established. The flow fronts of resin through various fiber geometries and structural
geometries were also examined. The spherical dome and cylindrical arch structures were
then subjected to low-velocity impact loading and their hi gh-energy absorption

capabilities were found to out-perform the conventional flat panels.

Jacob

DEDICATION

This work is dedicated to my wife, Jill, and my children, Emily, Nathan, Connor and

Jacob for their continual and loving support and prayers throughout graduate school.

iii

ACKNOWLEDGEMENTS

I express sincere thanks and gratitude to my advisor, Dr. Dahsin Liu, for his guidance,
teaching and friendship throughout this academic journey. I also want to thank Dr.
Alfred Loos and Dr. Andre Lee for taking time out of their schedules to help me fulfill

this academic achievement by serving as members of my defense committee.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ........................................................................................................... ix
1. INTRODUCTION ......................................................................................................... l
1.1 Composites Background ........................................................................................... 1
1.2 Literature Survey ...................................................................................................... 4
1.3 Scope of Study .......................................................................................................... 6
1.4 Organization .............................................................................................................. 7

2. COMPOSITE MATERIAL MANUFACTURING ....................................................... 9
2.1 Manufacturing Methods ............................................................................................ 9
2.1.1 Hand Wet-Out lamination .................................................................................. 9
2.1.2 Chopper-Gun Application ................................................................................ 10
2.1.3 Vacuum Assisted Hand Lamination ................................................................ l 1
2.1.4 Resin Transfer Molding (RTM) ....................................................................... 11
2.1.5 Autoclave Molding .......................................................................................... 12
2.1.6 Filament Winding ............................................................................................ 13

2.2 Vacuum Assisted Resin Transfer Molding (VARTM) ........................................... 14
2.2.1 Basic Technique of VARTM ........................................................................... 14
2.2.2 VARTM Components ...................................................................................... 16
2.2.3 Mold System .................................................................................................... 17
2.2.3.1 Mold Preparation ...................................................................................... 19
2.2.3.2 Reinforcing Fabric .................................................................................... 22
2.2.3.3 Release Fabric ........................................................................................... 26
2.2.3.4 Surface Distribution/Infusion Media ........................................................ 28
2.2.3.5 Core Distribution/Infusion Media and General Sandwich Construction. 29
2.2.3.6 Vacuum Bag .............................................................................................. 33

2.2.4 Resin System .................................................................................................... 36
2.2.4.1 Resin Inlet Piping ...................................................................................... 36
2.2.4.2 Resin Inlet Manifolds ................................................................................ 38
2.2.4.3 Infusion Resin ........................................................................................... 40

2.2.5 Vacuum System ............................................................................................... 46
2.2.5.1 Vacuum Outlet Piping ............................................................................... 46
2.2.5.2 Vacuum Outlet Manifolds ......................................................................... 46
2.2.5.3 Piping (in general) ..................................................................................... 47
2.2.5.4 Vacuum Gauge .......................................................................................... 48
2.2.5.5 Vacuum Reservoir/Resin Trap .................................................................. 48
2.2.5.6 Vacuum Pump ........................................................................................... 50
2.2.5.7 General VARTM Guidelines .................................................................... 51

a. . a. .— 1 ~. s. g C 53.. 8‘. \3 3‘. ‘3. 53

 

2.3 Problems Associated with VARTM ....................................................................... 53

2.3.1 Vacuum Leaks ................................................................................................. 53
2.3.2 Race-Tracking .................................................................................................. 55
2.3.3 VARTM’s Involved Set-up .............................................................................. 57
2.4 Safety ...................................................................................................................... 59
2.5 Processing Composite Laminated Parts Using VARTM ........................................ 60
2.5.1 Step 1: Mold Preparation ................................................................................. 61
2.5.2 Step 2: Laminate Preform Preparation ............................................................. 63
2.5.3 Step 3: Consumables Preparation .................................................................... 64
2.5.3.1 Release Fabric ........................................................................................... 65
2.5.3.2 Resin Infusion/Distribution Media ........................................................... 66
2.5.3.3 Resin Inlet and Vacuum Outlet Manifolds ............................................... 67
2.5.3.4 Composite Monitoring Equipment ........................................................... 70
2.5.3.5 Resin and Vacuum Piping ......................................................................... 70
2.5.3.6 Vacuum Bag .............................................................................................. 71
2.5.3.7 Vacuum Reservoir and Pump ................................................................... 75
2.5.4 Step 4: Vacuum Compaction ........................................................................... 75
2.5.5 Step 5: Matrix Impregnation by Infusion ......................................................... 77
2.5.6 Step 6: De-Molding of Laminate ..................................................................... 79
2.6 Domed and Arched Laminates ................................................................................ 80
2.6.1 Unique Attributes to Spherical Domed Laminates .......................................... 80
2.6.2 Unique Attributes to Cylindrical Arched Laminates ....................................... 82
2.7 Conclusions on Manufacturing ............................................................................... 84
3. COMPOSITE MATERIAL PERMEABILTY AND POROSITY STUDY ................. 86
3.1 Permeability Testing ............................................................................................... 86
3.1.1 Experimental Set-up ......................................................................................... 87
3.1.2 Permeability Determined by Darcy’s Law ...................................................... 87
3.1.3 Test Results - Flow Diagrams and Permeability .............................................. 89
3.2 Effects due to Fiber Geometry ................................................................................ 94
3.2.1 Fiber Geometry ................................................................................................ 94
3.3 Effects due to Fiber Orientation .............................................................................. 98
3.3.1 F inc-Woven E-glass Fiber Orientation ............................................................ 98
3.3.2 Coarse-Woven E-glass Fiber Orientation ........................................................ 99
3.3.3 Unidirectional E-glass Fiber Orientation ......................................................... 99
3.4 Effects Due to Structural Geometry ...................................................................... 100
3.5 Laminate Void Content and Porosity .................................................................... 101
3.5.1 Experimental Set-Up ...................................................................................... 101
3.5.2 Determining Composite Density and Volume ............................................... 103
3.5.3 Determining Fiber Volume ............................................................................ 103
3.5.4 Composite Volume ........................................................................................ 104
3.5.5 Calculating Void Volume .............................................................................. 105
3.5.6 Calculating Void Percent — Porosity .............................................................. 105
3.5.7 Data Analysis on Laminate Porosity .............................................................. 105
3.5.8 Results of Laminate Porosity ......................................................................... 106
3.6 Testing Conclusions .............................................................................................. 110

vi

ICC

 

 

 

 

 

4. COMPOSITE MATERIAL IMPACT TESTING AND RESULTS .......................... 1 ll

' 4.1 Determining Energy Absorption ........................................................................... 111
4.1.1 Low-velocity drop-weight impact test ........................................................... 111
4.1.2 Impact data ..................................................................................................... 114
4.1.3 Operating Procedure ...................................................................................... 114

4.1.3.1 Pre-impact test adjustments .................................................................... 114
4.1.3.2 Impact test procedure .............................................................................. 115
4.1.3.3 Rebounding and perforation ................................................................... 115
4.1.3.4 Data Acquisition ..................................................................................... 115

4.2 Data Analysis ........................................................................................................ 116
4.2.1 Load-Deflection Relation ............................................................................... 117
4.2.2 Impact Stiffness ............................................................................................. 120
4.2.3 Peak Load ....................................................................................................... 121
4.3 Energy Absorption ................................................................................................ 122
4.3.1 Integration Method ......................................................................................... 122
4.3.2 Extension Method and Integration ................................................................. 123
4.3.3 Energy Profile ................................................................................................ 124
4.4 Flat Composites Experimental Results ................................................................. 126
4.4.1 Load Deflection ............................................................................................. 126
4.4.2 Energy Absorption ......................................................................................... 127
4.4.2 Failure Results ............................................................................................... 128
4.5 Arched Composites Experimental Results ............................................................ 129
4.5.1 Load Deflection ............................................................................................. 130
4.5.2 Energy Absorption ......................................................................................... 131
4.5.3 Failure Results ............................................................................................... 132
4.6 Domed Composites Experimental Results ............................................................ 134
4.6.1 Load Deflection ............................................................................................. 134
4.6.2 Energy Absorption ......................................................................................... 135
4.6.3 Failure Results ............................................................................................... 136
4.7 Impact Test Conclusions ....................................................................................... 138
5. CONCLUSIONS, RECOMMENDATIONS AND FUTURE STUDY .................... 141
5.1 Conclusions ........................................................................................................... 141
5.2 Recommendations ................................................................................................. 143
5.2.1 VARTM ......................................................................................................... 143
5.2.2 Porosity Calculations ..................................................................................... 144
5.3 Future Study .......................................................................................................... 144
APPENDIX ..................................................................................................................... 145
REFERENCES ............................................................................................................... 152

vii

“I

 

—

 

Table
Table
Table
Table
Table
Table
Table
Table
Table

Table

LIST OF TABLES

Table 1.2.1 Summary of Literature on Domed Composites ................................................ 6
Table 2.2.1 Different types of mold releases available ...................................................... 21
Table 2.2.2 Relative thickness, rigidity and weight of sandwich composites ................... 33
Table 2.2.1 Gougeon Brother’s Inc., PRO-SET Epoxy Technical/Handling Data ............ 43
Table 2.7.1 VARTM Components used in this research ................................................... 85
Table 3.2.1 Characteristics of E—glass used in this study ................................................... 94
Table 3.5.1 Composite characteristics ............................................................................. 109
Table 4.4.1 Flat Composite Experimental Results Summary .......................................... 126
Table 4.5.1 Arched Composite Experimental Results Summary .................................... 130
Table 4.6.1 Domed Composite Experimental Results Summary .................................... 134

viii

 

 

F1;

LIST OF FIGURES

Figure 2.1.1 Chopper Gun Application ............................................................................. 10
Figure 2.1.2 RTM Mold system ........................................................................................ 12
Figure 2.1.3 Autoclave pressure vessel ............................................................................. 13
Figure 2.1.4 Filament winding schematic ......................................................................... 13
Figure 2.2.1 VARTM being used to produce a hull of a recreational boat ....................... 14
Figure 2.2.2 Typical VARTM Set-up ................................................................................ 16
Figure 2.2.3 Mold and its components in VARTM ........................................................... 18
Figure 2.2.4 Top View of mold and its components in VARTM ....................................... 18
Figure 2.2.4 Various mold releases used to keep parts from sticking to mold .................. 19
Figure 2.2.5 Relief angle to demold parts .......................................................................... 21
Figure 2.2.6 Mastic sealant tape ........................................................................................ 22
Figure 2.2.7 Reinforcing Fabrics ....................................................................................... 23
Figure 2.2.8 Cutting Tools ................................................................................................. 24
Figure 2.2.9 Bridging effect of vacuum bag ...................................................................... 25
Figure 2.2.10 Vacuum bag pleats ..................................................................................... 25
Figure 2.2.11 Schematic with laminate preform in place ................................................. 26
Figure 2.2.12 Schematic with release fabric placed over the laminate preform ................ 26
Figure 2.2.13 Effects Of Stretchy vacuum bag covering helix tubing ............................... 27
Figure 2.2.14 Schematic with distribution/infusion media installed ................................. 28
Figure 2.2.15 Types of distribution media ......................................................................... 28
Figure 2.2.16 Effects of 3-point bending applied on neutral axis ..................................... 30

ix

Figure 2.2.17 Compression and tension on upper and lower skins ................................... 30

Figure 2.2.18 Effect of core on relative flexural rigidity/stiffness and weight .................. 32
Figure 2.2.19 Vacuum bag installed on mold ................................................................... 34
Figure 2.2.20 Resin Piping and Manifold .......................................................................... 36
Figure 2.2.21 Resin piping Clamping Methods ................................................................. 37
Figure 2.2.22 Manifold placement and effect on laminate impregnation .......................... 39
Figure 2.2.23 PRO-SET 117 Resin and 226 Hardener ...................................................... 42
Figure 2.2.24 Measured, dispensed and mixed resin ........................................................ 44
Figure 2.2.25 Flow of Resin through the system ............................................................... 45
Figure 2.2.26 Vacuum flow in a laminate .......................................................................... 47
Figure 2.2.27 Vacuum reservoir/resin trap ........................................................................ 49
Figure 2.2.28 Quick Release Fitting .................................................................................. 50
Figure 2.2.29 Vacuum Pumps ............................................................................................ 51
Figure 2.2.30 General gridelines to consider when setting up VARTM ........................... 52
Figure 2.3.1 Ultrasonic leak detectors .............................................................................. 53
Figure 2.3.2 Examples of race-tracking ............................................................................. 56
Figure 2.3.3 Bridging examples ......................................................................................... 56
Figure 2.3.4 VARTM set-up ready to be infused .............................................................. 57
Figure 2.5.1 Completely waxed mold ............................................................................... 61
Figure 2.5.2 Spherical domes on the glass to produce domed composites ........................ 61
Figure 2.5.3 Step 1 Mold Preparation ................................................................................ 62
Figure 2.5.4 Steel cylindrical arch mold ............................................................................ 62
Figure 2.5.5 Mold for cylindrical arches ........................................................................... 63

Figure 2.5.6 Step 2 Laminate preparation .......................................................................... 64

Figure 2.5.7 Step 3 Release fabric lay-up .......................................................................... 66
Figure 2.5.8 Step 3 Distribution/Infusion media lay-up .................................................... 67
Figure 2.5.9 Step 3 Resin and vacuum manifolds ............................................................. 69
Figure 2.5.10 Step 3 Resin and vacuum piping ................................................................. 71
Figure 2.5.11 Step 3 Vacuum bag installation ................................................................... 73
Figure 2.5.12 Inserting a pleat into a vacuum bag ............................................................ 74
Figure 2.5.13 Step 3 Connect vacuum and vacuum reservoir to the mold ........................ 75
Figure 2.5.14 Step 4 Laminate compaction ....................................................................... 76
Figure 2.5.15 Step 4 Laminate compaction and satisfactory leak test ............................... 77
Figure 2.5.16 Step 5 Matrix impregnation by vacuum infusion ........................................ 78
Figure 2.5.17 Step 5 Resin infusion via VARTM ............................................................. 79
Figure 2.5.18 Step 6 De-molding of cured composite laminates ....................................... 80
Figure 2.6.1 Mold for manufacturing arched laminates .................................................... 83
Figure 3.1.1 Flow Area vs. Time curves for all flat and domed laminates ....................... 90

Figure 3.1.2 Flow Area vs. Time curves for unidirectional flat and domed laminates ....91

Figure 3.1.3 Permeability of flat and domed laminates ..................................................... 92
Figure 3.1.4 Pictorial definition of permeability unit of measure, Darcy .......................... 93
Figure 3.2.1 Different fiber geometries used in this study ................................................ 95
Figure 3.2.2 Woven E-glass’ Unit Cell .............................................................................. 96
Figure 3.2.3 Dry areas of unidirectional preform ............................................................. 97
Figure 3.5.1 Experimental set-up to measure density ...................................................... 102
Figure 3.5.2 Pre-bum out specimens ............................................................................... 107

xi

Figure 3.5 .3 Post-bum out specimens .............................................................................. 107

Figure 3.5.4 Calculated percentage of laminate porosity ................................................ 108
Figure 4.1.1 Instron’s Dynatup Low-Velocity Impact Tester .......................................... 113
Figure 4.2.1 Types of curves for penetrated laminate and rebounded laminate .............. 118
Figure 4.2.2 Load-Deflection curves for domed laminates ............................................. 119
Figure 4.2.3 Load-Deflection curves for flat, penetrated laminates ................................ 120
Figure 4.2.4 Parts of a Flat Composite Load-Deflection Curve ...................................... 121
Figure 4.2.5 Parts of an Arched or Domed Composite Load-Deflection Curve .............. 122
Figure 4.3.1 Area under the load-deflection curve ......................................................... 123
Figure 4.3.2 Energy profile of flat composites ................................................................ 125
Figure 4.4.1 Load-Deflection curves for flat composites ................................................ 127
Figure 4.4.2 Absorbed energies of flat composites .......................................................... 128
Figure 4.4.3 Typical flat composite damage by penetrated tup tip .................................. 129
Figure 4.5.1 Load-Deflection curves for arched composites ........................................... 130
Figure 4.5.2 Absorbed energies of arched composites .................................................... 132
Figure 4.5.3 Damaged arched composite ......................................................................... 133
Figure 4.5.4 Perforated arched composite ...................................................................... 133
Figure 4.6.1 Load-Deflection curves for arched composites ........................................... 135
Figure 4.6.2 Absorbed energies of Domed composites ................................................... 136
Figure 4.6.3 Penetrated Domed laminate ........................................................................ 137
Figure 4.6.4 Rebounded Domed laminate ...................................................................... 138
Figure 4.7.1 Absorbed energies Of 3 ply coarse woven composites ............................... 139

xii

 

 

1.1

to st
in to
Drea

With

1. INTRODUCTION
1.1 Composites Background

Interest in advanced structural composites has grown exponentially over the past half
century as they have evolved from resorcinol glue-laminated birch wood used in Howard
Hughes’ SPRUCE GOOSE and laminated plywood used in the PT boats of World War II
to specialized infusion resins married with carbon or aramid reinforcing fabric laminates
in today’s most sophisticated structures. Some of these behemoths include Boeing’s 787
Dreamliner (made with 50% composite material by weight) and Airbus’s A380 (made
with 40% composite material by weight), as well as various armored tanks and ground
support vehicles and the Royal Swedish Navy’s 76 meter (236 foot) Visby-class corvette.
Although contemporary materials such as steel, aluminum and titanium are still being
used in today’s sophisticated structures, advanced structural composites are quickly
replacing these homogeneous alloys because they can be tailored by aligning fibers to
properly meet the engineering requirements of each application to maximize
characteristics such as high stiffness, high-strength and low-density while improving
energy absorption capability. It is in this improved energy absorbing characteristic that
this research focuses; to understand, utilize and improve an established and efficient
manufacturing process to produce composite armor that may be used in personal and
vehicular protection.

A composite is a material consisting of multiple components whose mechanical
performance and properties are designed to be superior to the individual components

acting by themselves [1]. Composites have been and continue to play a large role in

 

industries such as aeronautics, aerospace, automotive, marine, medical, military,
recreational — just to mention a few.

As noted above, thin shell structures, such as those found in the structures of air, land
and sea vehicles, have been employed in various civilian and military uses, including
various types of armor. Although there have been many studies on composite cylindrical
shell and metallic shells of various shapes and sizes, there has been little research written
on spherical or other geometrically complex shaped shells, using continuous fiber
reinforcement and utilizing the vacuum assisted resin transfer molding (V ARTM)
manufacturing process.

The Armor Research Lab [2] compared the effectiveness of glass-reinforced plastics
to conventional steel. They found that steel has less damaged areas and increased
residual integrity as compared to glass-reinforced plastics. Because the composite
structures have less residual integrity, manufacturers of armor systems must use designs
with cellular patterns to ensure that damage of one cell does not affect the integrity and
effectiveness of those cells adjacent to them [3]. During research for armor and impact
tests completed on various structures, many authors have focused on the cylindrical arch.
Most of the composites are not only reinforced in the in-plane directions, but through the
thickness direction, too. Therefore the total structure strength is improved [4].

The key to any armor system or structure that needs to withstand impact from
multiple sorts is to design a composite that has a low density, to decrease the weight of
the material, and to increase the energy absorption characteristics to allow the structure to
withstand single and multiple impacts and continue to maintain structural integrity. By

using complex shapes, such as domes, and altering fiber orientation, the stresses

 

prodt

energ

comp
Sllllllr
ascor
energj
hghbx
Vehk
€fiéeh
abSOrl'
Trudi:
lighter
Tl
T116351]
Val’lnu;
Hers/[j
Compt)
Adum
base 01.
onlbti
Conduit

require

produced by impact may be better distributed and absorbed. This allows more efficient,
energy—absorbing structures to be designed.

A couple of parameters must be considered when designing energy-absorbing
composites. Two of the most important parameters are strength-to-weight ratio and
stiffiiess—to-weight ratio. For structural applications, the higher the strength and stiffness
as compared to weight, the better. For personal body armor, low stiffness and high
energy absorption is needed. Similarly, many vehicles used in the theaters of war are
light-weight vehicles, such as AM General’s M998 High Mobility Multipurpose Wheeled
Vehicle (HMMWV) commonly called the “Humvee.” These light vehicles cannot be as
effective to our military fighting forces if they are laden with heavy armor. Therefore, the
absorbed energy-to-weight ratio or the strength-to-weight ratio must be maximized.
Traditional materials, such as steel and other metals cannot meet these demands, whereas
lighter composites can adequately fill this niche.

There have been many studies completed on arched laminated composites that
measure the impact response, characterization of damage, distribution of stress and
various forms of buckling, such as those found by Ambur et al [5], Chun and Lamb [6],
Herszber and Weller [7], Kistler [8, 9 and 10] and others. However, the domed
composites have not been evaluated or studied as extensively as the arched composites.
A domed shape differs from an arched shape in that the curvature forms a circle at the
base of the laminate, forming a complete boundary. Arches, on the other hand, are open
on two sides and are therefore not constrained to move in these directions. The boundary
conditions, that domed specimens naturally provide, help in their energy absorption and

require increased load to induce deformations to the specimens.

 

 

qu;

Th:

frat

1.2 Literature Survey

Gupta and Prasad [ l 1] studied the collapse mechanisms and energy absorption using
quasi-static and drop hammer tests of random glass mat and foam filled spherical domes.
They found that progressive crushing was due to the formation of successive zones of
fracture. They were able to develop an analytical model to predict load deformation and
energy compression curves. Although they used spherical domes, their material was
random glass mat, which does not have the structural integrity of continuous-stranded
fabric.

Gupta and Velmurugan [12] subjected random glass mat and foam filled parabolic
domes to axial compression, and studied the mechanics of deformation and energy
absorption. They found analytical expressions of average crush loads and total strain
energies through experimental observations. Again, they developed analytical models
that gave average crush loads and total strain energies, which matched well with
experimental results.

Cui, Moltschaniwskyj and Bhattacharyya [13] studied domes using a thermoforrned
twill woven electrical-grade fiberglass (E-glass) and knitted E-glass weft fabric in a
polypropylene thermoplastic matrix. They studied and described the large deformation
behavior of transversely loaded domes between rigid platens. They found that domes of
various radius-to-thickness ratios demonstrated an initial bifiircation buckling and a
rolling—plastic hinge post-buckling response. They found that material type, ply number
and stacking sequence were major factors in the deformation behavior.

Prasad and Gupta [14] studied aluminum spherical domes and conical fi'usta. They

researched similar collapse mechanisms and energy absorption to studies of random

fiberglass mat in [11]. However, with a homogenous material, such as aluminum, they
were able to describe the crush behavior. They found that the collapse starts with initial
flattening, and then a rolling plastic hinge is formed. They also found that the modes are
sensitive to rate of compression of the quasi-static and drop hammer.

Ganapathy and Rao [15] studied the failure analysis of laminated fiberglass spherical
domes and cylindrical shells subjected to low-velocity impact. They found that the major
mode of damage was matrix cracking, which was always along the fiber direction. They
also found that the dome specimens were stiffer and experienced more damage than the
cylindrical shells.

Her and Liang [16] explored spherical shell made of epoxy and graphite fibers. They
used finite element analysis (FEA) and subjected shells to low-velocity impacts. The
impact-induced damage, including matrix cracking and delamination were predicted by
the various failure criteria and the damaged areas were plotted. The FEA results proved
to be in synchronization with experimental results found in published literature, such as
those published by Chandrashekhara K. and Schroeder T [17].

In summary of the literature mentioned on complex geometry domed composites,
Table 1.2.1 was created to quickly ascertain the information noted above. It covers the
shape, material, radius, thickness, shell depth, and manufacturing process used in the
studies. The loading type studies ranged from low-velocity to quasi-static. The studies
were mostly experimental; however, one used finite element analysis to obtain results.
The materials were mostly random fiberglass mat, however, one was continuous
fiberglass strands, another was continuous graphite strands and one was aluminum. The

manufacturing processes varied from hand-lay-up to thermoformed domes. None of

 

ther
Inul

Tabl

’3 SC01

technlqu
CUrnpl

Wk;

Inlugmn

them used an automated process such as vacuum assisted resin transfer molding utilizing
multiple layers of continuous strands of reinforcing fabric or woven cloth.

Table 2.2.1 Summary of Literature on Domed Composites
Shell Radius/

Ref Shape Material Radius Thickness Depth Thickness Mfg
mm mm . process
mm ratio
Random
11 Spherical Glass Mat 53-106 1.1-2.07 52-100 25-96 “and
empty & Lay-up
foam filled
Random Hand
12 Parabolic Glass Mat 49-54 2.3-3 0.2-1.2 N/A La -u
foam filled y p
Towflex 2/2
twill woven e-
glass & e-
. glass knitted Therrno-
l3 Spherical Milano weft 100 1.9-4.5 N/A 22.2-52.6 formed
fabric in a
polypropylene
matrix
Spherical 15 3 _
l4 & conical Aluminum 39-126 0.5-2.8 33-110 ' Stamp
240.9
frusta
Spherical [0’99]
& graphite/ 3 8 1 _
15 . . bismalemide 1-2 mm N/A 190.5 N/A
cylindrical . 508
shells continuous
fiber
T300/976 Arc =
16 Spherical graphite 25.4 1:421 20'31 6 N/A N/A N/A
epoxy mm '
1.3 Scope of Study

The scope of this study was two-fold. The first was to understand and develop
techniques utilizing vacuum assisted resin transfer molding (VARTM) to manufacture
complex geometrical specimens such as arches and domes. Various forms of vacuum

infusion, including VARTM, have been used for the past 40 years in different capacities,

 

however, recent developments in resins, fabrics and distribution media have made it
advantageous for manufacturing of parts of all sizes. The Objective was to study a resin’s
flow characteristics in a flat panel, changing the unit cell size and fiber orientation, and
study the difference these changes made on the flow of resin. Then geometric shapes
were introduced into the mold, which changed the resin flow front, but how much it
changes was not fully understood. The infusion rate of flat panels would be compared to
the domed panels.

The second objective was to investigate the relationship of geometry of flat, arched
and domed composites and absorption energy for low-velocity impact. All specimens
were manufactured with the VARTM manufacturing process.

The laminates consisted of continuous fiber-reinforced polymer-matrix composites,
specifically, electrical-grade fiberglass (E-glass) and thermosetting epoxy resin with a
modified aliphatic polyamine hardener. The E-glass used was (a) fine, plain woven (b)
coarse, plain woven and (c) unidirectional fibers bound with non-structural fill in various
fiber orientations. Analysis of the load-deflection relation, the energy profile and the
damage process were of primary interest because they provided insight into the impact
behavior of the laminate specimens such as peak load, deflection at the peak load,
specimen stiffness, maximum specimen deflection, contact duration, energy absorption
and various damage modes.

1.4 Organization

This thesis is organized into five chapters. Chapter 1 is an introduction into

composites, which includes a background history, armor and structure design, and domed

composites. Chapter 2 discusses the VARTM manufacturing process in detail. Since

VARTM
details on
future res
from suec
med in V
technique
porosity 1
focuses 0
discusses

research 2

VARTM is useful to a wide range of composite sizes, it is important to lay the specific
details out so others may utilize this document to produce satisfactory composites in
future research. The tips and techniques used were gained from hands-on experience and
from successes and failures of many VARTM manufacturing sessions. The specific steps
used in VARTM are then discussed in detail, to emphasize the materials, tips and
techniques discussed earlier in the chapter. Chapter 3 details the permeability and
porosity testing of specimens and contains the results of these experiments. Chapter 4
focuses on the impact testing, equipment operational procedure, data acquisition and
discusses the results of the impact testing. Chapter 5 contains the conclusions from this

research and makes recommendations for future study.

1.. COMM
llilanula

\acuur
detelopmer
widely user
finnbuuor
processrs

0 lid

'Cl

One
IS “'61 0‘
bleak U]
311 \de
becaUSC

ofneat.

2. COMPOSITE MATERIAL MANUFACTURING
2.1 Manufacturing Methods

Vacuum Assisted Resin Transfer Molding (V ARTM) has been under continuous
development since the 1940’s. However, not until the last two decades has it become
widely used because of advancements in resins, reinforcing fabrics and availability of
distribution or infusion media. The hierarchy of resin—fiber laminate manufacturing
process is as follows:

0 Hand Wet-Out Lamination

o Chopper-Gun Application

0 Vacuum Assisted Hand Lamination

o Resin Transfer Molding (RTM)

0 Autoclave Molding

o Filament Winding

0 Vacuum Assisted Resin Transfer Molding (VARTM)
2.1.1 Hand Wet-Out lamination

One of the most traditional methods is the hand wet-out method, where the laminate
is wet out by resin using brushes. Air buster rollers and squeegees are then employed to
break up the air bubbles and voids in the laminate and force resin into the dry laminate’s
air voids. This process is time consuming, and for a multiple ply laminate, can be tricky
because if too much resin is used, the plies will “float” in the excess resin, causing layers
of neat resin, which create stress concentrators between the plies of reinforcing fabric
when stressed. If too little resin is used, the plies may not adhere to each other, creating

large air pockets, which makes causes the part to premature delamination and fail if

stressed.
fiber mtic
2.1.2 C110
Anotl
fibers are
taken off ;

into taryr'

[hel' are m
meld‘ the
“lids Th.
the pan IS
addllion. tl
for Super“
COHQQmmL

Aside Iron

stressed. Some professional hand laminators can laminate a part of nearly 50:50 resin-to-
fiber ratio, which is fairly well considering the circumstances.
2.1.2 Chopper-Gun Application

Another kind of early manufacturing processes is where the resin and reinforcing
fibers are applied to a mold using a “chopper gun.” Here, a strand of fiberglass tow is
taken off a roll of fiberglass and sent through a chopper, which cuts the fiberglass tow

into varying random lengths. As the non-continuous strands are ejected from the gun,

  

Figure 2.1.1 Chopper Gun Application [18]

they are mixed with a stream of catalyzed resin and blown onto the mold. Once on the
mold, the wet resin and fiber mixture is then rolled and squeegeed by hand to remove air
voids. This process causes non-uniforrnity of thickness and consistency in the part and
the part is relatively weak because of the non-continuous reinforcing material used. In
addition, there is normally a high resin-to-fiber ratio, nearing 65:35, which doesn’t allow
for superior composite benefits, which includes areas of high resin content and stress
concentrators in the part. The human factor becomes a large concern of this process.

Aside from this somewhat messy application process, this is the most common form of

10

lamination
process. H
utilized.
2.1.3 \‘acr
Hand
The lamrr
Then. a sc
applied 0'
resin is re
results; h
mfllfil‘lal:
10~frber r
ItSln UCE
the \‘acu

V 3C UUm

pm
inn
Ihei

the

ap-

lamination today. Most recreational fiberglass boats are still manufactured with this
process. However, in more advanced structures, the chopper-gun technique is not
utilized.
2.1.3 Vacuum Assisted Hand Lamination

Hand lamination was improved by applying a vacuum over the wet-laminated part.
The laminate is wet-out by hand as noted in the Hand Wet-Out Lamination section.
Then, a series of layers, consisting of release fabric, bleeder/breather and vacuum bag are
applied over the wet laminate. A vacuum is then applied on the laminate and the excess
resin is removed and forced into the bleeder layer. This process produces very good
results; however, it has a potential of being fairly messy because of applying the
materials over a wet laminate. A vacuum assisted hand-laid laminate can obtain a resin-
to-fiber ratio of nearly 40:60. A precaution that must be taken for larger parts is that the
resin needs to have a long enough open time to still be wet when the vacuum is pulled on
the vacuum bag. If the resin gels before the vacuum is applied, all the advantages of
vacuum bagging will be lost because excess resin would not be removed from the
laminate stack and a resin-rich laminate would be produced.
2.1.4 Resin Transfer Molding (RTM)

There are a few processes that use vacuum to draw resin through the laminate. One

process, RTM, uses a closed, matched-mold concept. Two rigid matching molds,

inner and outer molds are used. When they come together, there is a space between

them that the laminate preform fits into. Vacuum not only draws the resin through

the laminate, but resin is pressurized and forced into the laminate stack by means of

a pump. Here, high pressures can be obtained and the laminate can be quickly

ll

lamir
quml
hour
separ
procr

make

2J~5:\u|

AUIO
Pieprcgs‘
high pres:
the resin i
tunaddn

required C

laminated. In addition, the mold can be heated to promote the resin to be cured
quicker, thus reducing the cycle time. This process produces excellent parts;
however, the mold can be very expensive, being that it is matched with a uniform
separation. Both surfaces are also finished, so aesthetics may be better than other
processes. In some cases, the comers or perimeter areas can be resin-rich, which can

make them somewhat brittle.

 

2.1.5 Autoclave Molding

Autoclave Molding uses laminates that are pre-impregnated with resin, called
prepregs. The laminate is placed into a large pressure vessel. Once closed, high heat and
high pressures are applied to liquefy the resin and condense or debulk the laminate. Once
the resin is liquefied, the pressure compacts laminate and the laminate continues to cure
with additional heat. Most prepregs require a specific cure schedule to obtain the

required engineering properties. This process is costly because of the manufacturing tools

mama

lL6F

F I
contm
Wmn:
Home

needed. Pressure vessels range from smaller ones of modest sizes (40 gallons) to large

ones that entire boats or airplane fuselages can fit into.

  

Figure 2.1.3 Autoclave pressure vessel [20]

2.1.6 Filament Winding

Filament winding processes use continuous strands of reinforcing fabric that are
continually wet out in a resin bath and spun on to a rotating mandrel. The shapes are
normally confined to cylindrical shapes because of the limitations on the mandrel.
However, fiber orientation can be controlled to maximize the strengths of the aligning

fibers to meet engineering requirements.

 

2.2 Vac
Th:
usuaH}'
because
surface.
inexpen
annospb

boat ben

12J Ba:

time In
mdnjduE
18351 TWO
rmnfificn
5pm). r65

2.2 Vacuum Assisted Resin Transfer Molding (V ARTM)

The infusion process used in this study is Vacuum Assisted Resin Transfer Molding,
usually referred to as “VARTM.” VARTM is a much less expensive process than RTM
because it uses an open, single-sided mold, which means that the mold is only one
surface, not a matched mold like RTM. Very large parts can be molded fairly
inexpensively with this process because of the low resin injection pressures, only that of
atmospheric pressure, and a one-sided mold. Figure 2.2.1 is a picture of a recreational

boat being manufactured by a one-man shop.

2.2.] Basic Technique of VART M

As noted in Chapter 1, a composite is a material consisting of multiple components
whose mechanical performance and properties are designed to be superior to the
individual components acting by themselves. Fiber composites are usually made up of at
least two types of components, the reinforcing fibers and the matrix. In this study the
reinforcing fibers were electrical grade fiberglass (E-glass) and Bisphenol-A/Bisphenol-F
epoxy resin blend combined with a room-temperature cured modified aliphatic polyamine

hardener.

14

and (3-
M.
reinfort
compo:
next. A
resin th.
to be pl.
Res
The pip;
Vat
creates a
bag aehi
Checked
the resin
Resi
highEI PT
13 DUShed
reinforC i r.
Once all f
compat‘lm

The ..

There are three basic parts to a VARTM set-up: (l) Mold System, (2) Resin System
and (3) Vacuum System.

Mold System: A non-porous mold is used to form the composite part. The
reinforcing fibers are cut and laid upon the mold. A release fabric is applied next so the
composite part won’t stick to the resin distribution media, which is placed on the mold
next. A vacuum manifold and resin manifold are placed to maximize the distribution of
resin through the distribution media and laminate. A vacuum bag is the last component
to be placed on the mold and sealed with sealant mastic tape.

Resin System: Resin lines are attached to the resin manifold inside the vacuum bag.
The piping is clamped off until the resin is ready to be infused into the laminate.

Vacuum System: The vacuum manifold is connected to a vacuum pump, which
creates a vacuum inside a vacuum bag; creating a low pressure within. Once the vacuum
bag achieves 1 atmosphere or 101 kPa of pressure, the vacuum bag’s perimeter is
checked to ensure there are no air leaks. Once this is confirmed, the resin is mixed and
the resin clamp is removed.

Resin Impregnation: The low viscosity, reactive thermosetting resin flows from a
higher pressure reservoir (atmospheric pressure) to the lower pressure vacuum. The resin
is pushed through the high permeable distribution media and is forced into the dry
reinforcing laminate. The resin displaces the air in the laminate and wets out the fabric.
Once all fabric is impregnated with resin, the flow is stifled and the laminate is then
compacted under the vacuum pressure.

The “head” or weight of air made by the column of air above the mold creates a

pressure on the mold and laminate. The pressure helps to drive the resin to fill the Open

15

spaces in a
pump sucl
above the
laminate 5
process. I
Figure 2.2

Vacuum
P'BSSJre

 

 

M ‘33 :

bin.

spaces in and among the reinforcing fibers. A common misconception is that the vacuum
pump sucks resin through the laminate, however, in actuality it is the pressure of air
above the laminate that pushes the resin out of the resin reservoir into the lower pressure
laminate stack. Equalizing from a higher pressure to a lower pressure is a natural
process. It is also important to note that the resin will take the path of least resistance.

Figure 2.2.2 schematically depicts the VARTM Process.

  

Ma
Resin Manifold — Sealant
Vacuum Helix tu - in

Pressure

  
 

Vacuum Piping

  
  
 

  

Vacuum
Reservoir/
Resin Trap

 
   
  
 
 
 

Vacuum Manifold —
. . . Helix tubing
Distribution
Media Release
Fabnc

 

Laminate Preform

Figure 2.2.2 Typical VARTM Set-up

2.2.2 VARTM Components

Now that the basics of VARTM are addressed, each processing step will be studied
in more details. Although there are many steps to the VARTM process, each one must be
fairly exercised. Overlooking one step can ruin a part and give it a quick trip to the trash

bin.

2.2.3 .\1
Fig
3 top-vi

on to th

The
a PTOCed

mCIUSH c

2.2.3 Mold System

Figure 2.2.3 and 2.2.4, below, show a cross-section of one variation of VARTM and

a top-view, respectively, of the mold system. It includes the various parts that actually fit

on to the mold during the manufacturing process.

Mold or T 001 — a cavity or surface in which the fiber preform and matrix is shaped
into a desired finished composite laminate

Laminate or Preform — Reinforcing fabrics & cores used to make part

Infiasion Resin — Low viscosity resin that fills air voids in part and keeps
composite’s shape

Release Fabric — Separates laminate and infusion media from sticking with each
other

Vacuum Bag — Creates an isolated area and vacuum pressure on the system
Distribution/Infision Media — Creates low resistance openings to distribute resin
Vacuum Tubing — Evacuates air from the mold to create vacuum pressure on mold
Resin Tubing — Channels resin from reservoir to infusion media

Mold Release — Keeps part from sticking to mold

Vacuum Pump — Evacuates air from the system and maintains a vacuum on the
system until the resin cures

The specific parts utilized in VARTM are described below. This discussion includes

a procedure and tips to create a satisfactory part using VARTM. Although not all-

inclusive, it contains precautions that should be taken for each step.

17

Vacuum Piping —
Resin Inlet Vacuum Bag 1° vacuum Pump
reservoir ManlIOId Vacuum Outlet
Manifold

Resin Piping —

     

Plate Glass Mold fiberglass Cloth]

 

 

Tape Laminate preform
Release Fabric
Figure 2.2.3 Mold and its components in VARTM
Mastic
Resin Manifold - Sealant Mold
Helix tu-ing tape

Vacuum Manifold -
. Helix tubing
Distribution .
Media ‘ . ' Release
Laminate Preform Fabric

Figure 2.2.4 Top view of mold and its components in VARTM

18

 

2.2.3.1 .\
The
part pullr
Scratche:
mold. T1
and effor
surfaces.
an except
process rr
T0 prex'er
Surfaces a
trapped in
excellent 1
imperam.
Slick. Fig

mold for l.-

2.2.3.1 Mold Preparation

The key to any lamination process is the mold, or sometimes referred to as tool. The
part pulled from the mold will take on any characteristics or flaws that are on the mold.
Scratches, scrapes, dents, bumps or other irregularities will be directly transferred to the
mold. The more time and attention to detail spent on the mold, the less finishing time
and effort will be needed on the part. The mold must be smooth and clean. For flat
surfaces, tempered plate glass or Formica can be used with good success. Plate glass is
an exceptional mold because it is sturdy and naturally smooth. In addition, the laminate
process may be seen from the underside, where voids and resin flow may be monitored.
To prevent parts from sticking to the mold, it must be waxed or otherwise treated. Most
surfaces are porous and a wax needs to be used to fill the pores so the epoxy won’t get
trapped inside. Since most surfaces have porosity to some varying degree, wax is an
excellent material to prepare the mold. When the mold includes complex shapes, it is
imperative that the mold is well waxed or otherwise protected so as the epoxy will not
stick. Figure 2.2.4 shows some mold release waxes and other agents used to prepare a

mold for lamination.

 

Tab

at an
wax
relea:
these j
tested
many r.
matena.
mold rel
and main
are spray:
mllSl thicl
must be re;
Stoner men
iceler gangt

There are various types Of mold release waxes on the market, as shown below in
Table 2.2.1. Some resist temperature better than others, so if the part will be post cured
at an elevated temperature, then a high-temperature wax should be used. Otherwise, the
wax may melt out of the pores and cause parts to stick. In addition to waxing, other mold
release products may be used to help parts de-mold, also noted in Table 2.2.1. Some of
these products may leave residue on the part, so it is imperative that mold releases are
tested or researched to ensure what performance can be expected out of them. There are
many mold releases available to the laminator. Locktite’s FreCote is a low-viscosity
material that wipes or brushes on. The solvent vehicle evaporates and leaves an excellent
mold release. Stoner produces various materials that are in convenient aerosol containers
and makes application fairly simple. Other mold releases, like polyvinyl alcohol (PVA)
are sprayed or brushed on and creates a physical barrier or skin of 0.5 to 1 millimeter (2-4
mils) thick between the laminate and part, which normally de-molds with the part and
must be removed with warm water. PVA is a thicker mold release than FreCote or
Stoner mentioned above, so tighter tolerances of 0.5 to 1 mm need to be evaluated using a
feeler gauge or by other thickness measurement devices. However, when sprayed, it may

cover some minor mold surface irregularities in the part.

20

 

Shape.
the part. A
[Li Shown ir
Parallel am-

mold tun -~

Table 2.2.1 Different types of mold releases available

 

 

Mold Release
W§_x Non-wax
TR Reg
Temperature Polyvinyl Alcohol (PVA)
TR High
Temperature Silicone Mold Releases
Honey Camuba Wax Urethane Mold release
Meguiar's Mirror Zinc Stearate Mold
Glaze Release
Naphtha-based Releases

 

 

 

 

Shaped areas are more prone to stick parts, especially if there is little-to-no relief in
the part. A relief is a slight angle, from 1° to 10°, in a mold that has nearly parallel sides,
as shown in Figure 2.2.5. In this case, the relief prevents the sides from being exactly
parallel and thus, causing difficult de-molding. As more parts are pulled from a mold, the

mold will “season” and it will not be necessary to wax every time a part is laminated.

  
   

  

 
     
 

=... : -'“~:".fi:5,525g.=...:
: a: : 5gfi'mi' :.--...=- :--~.
"Tap-fa” -i'§in::i§fi-:’:miii:iii

'3! samuueeis%:::::

 

 

 

/ Relief Angle—> ('—

 

 

 

7

Figure 2.2.5 Relief angle to demold parts.

Mold

Once the mold is prepared, mastic sealant tape, as shown below in Figure 2.2.6, is
then applied around the perimeter of the mold. The sealant tape has one side with paper

on it. The paper is kept on the sealant until the vacuum bag is applied; otherwise, dirt or

21

other materi
stick the var
draw the res
preparation
vacuum pip
other tubing
system. It i

formed.

21.32 Rl‘li
The lat
IS plElCQd UT

Pl}? may C01

311 electr
oh-“
e 35). Carb.
KEVLAR

man}. I\‘p

other material will fall on it and reduces its effectiveness. The sealant tape is used to
stick the vacuum bag onto the mold and create the air-tight seal to allow the vacuum to
draw the resin through the laminate pre-form. Sealant tape has many uses during the
preparation of a mold. It may be used to cover and seal piping joints in both resin and
vacuum piping, seal small holes in the vacuum bag and keep omery piping manifolds or
other tubing, distribution media or fabrics in place before a vacuum is applied to the
system. It is also useful to cover sharp comers under the vacuum bag so holes are not

formed.

’ 33%,?

I"
,/ :;

  
   

Mastic ”A

Sealant

Tape Protective
(yellow) ./ Paper

. 9
Figure 2.2.6 Mastic sealant tape

2.2.3.2 Reinforcing Fabric

The laminate contains all reinforcing fibers and core material. The order in which it
is placed or stacked is called the “schedule.” Each layer of reinforcing fabric is 3 ply. A
ply may consist of different material and different type of weave. Reinforcing fabric may
be an electrical-grade fiberglass (E-glass) as in this study, structural-grade fiberglass (S-
glass), carbon or graphite fiber or any number of aramid fibers, such as Dupont’s
KEVLAR. The reinforcing fabric may be knitted together to form a weave. There are

many types of weaves, such as plain, twill, herringbone, etc. In some cases it is not

22

 

woven. b
other case
-45 445 b
the el‘fecti

composite

woven, but the strands are all parallel with each other, and considered unidirectional. In
other cases, unidirectional fabric may be stitched together at various angles, such as a
+45/-45 bias ply. In any case, the laminate should be engineered such that it maximizes
the effectiveness of each ply and utilizes the alignment of the fibers to create strong

composites. Samples of E-glass used in this research are found in Fig. 2.2.7, below.

 

Figure 2.2.7 Reinforcing Fabrics, L to R,
unidirectional E-glass, coarse woven E-glass
and fine woven E-glass

Careful preparation ensures that the laminate will perform as desired. Although
scissors are widely used to cut plies to shape, some special scissors are needed for the
aramid fabrics. These scissors use serrations to grip the fabric so that it doesn’t slip out
of the scissor blades. All reinforcing fabrics can be somewhat tough to cut with scissors
due to them being seemingly slippery because of fiber treatment or sizing. One easy fix
is to use rotary cutters, such as those used in the regular textile industry. The rotary
cutter’s blade is basically a round razor blade. Rotary cutters, as shown below in Fig.
2.2.8, cut most fabrics well, especially E-glass, the most commonly used reinforcing
fabric. Paired with a good self-healing backing board, constructed of laminated PVC

plastic of different grades/hardnesses, the rotary cutter can be an indispensable tool for

23

 

 

prepann

fiber are

In cas
room 10 fit
area. bridg:
COHtour Of‘
Which may.
“it rein ion
thick resin T
needs to be ;
minimjze [in

Ft

'3

-19

(.10

preparing the laminate schedule. Cutting tools commonly used to prepare reinforcing

fiber are found above in Fig. 2.2.8.

    

Figure 2.2.8 Reinforcing Fabric Cutting
Tools, scissors on left, rotary cutter on right

In cases where complex geometry is involved, it is imperative that there is plenty of
room to fit into a concave shape or around a convex one. If care is not exercised in this
area, bridging may occur. Bridging is when the reinforcing fabric does not follow the
contour of the mold, but rather skips from one surface to another, leaving an air pocket,
which may, or may not fill up with resin, both resulting in inadequate composite parts.
The reinforcing fabric spanning the two surfaces acts like a bridge. This air pocket or
thick resin pocket can significantly affect the quality of laminate. The reinforcing fabric
needs to be able to maintain contact with the concave or convex surface to eliminate or
minimize this bridging effect. A schematic of the bridging effect can be seen below in

Fig. 2.2.9.

24

 

Not or
pl}: infusio
may create
inserted inn
and allovv s t

Fig. 2.2.10.

FIgUi’e 2n

tape placed on

Vacuum bag in
contact with
laminate & mold

  
   
   

Figure 2.2.9 Bridging effect of vacuum bag,
or any other laminate part, i.e. laminate,
reinforcing fabric, etc.

Not only does the reinforcing fabric need to be carefully prepared, so does the peel
ply, infusion media and vacuum bagging material. All of these, if not prepared properly,
may create the bridge effect and ruin a laminate. The vacuum bag may have pleats
inserted into it to allow extra room in the bag. A pleat takes up room on the perimeter
and allows the extra bag to conform where needed. An example of a pleat can be seen in

Fig. 2.2.10.

Figure 2.2.10 Vacuum bag pleats made to
take up extra room needed within mold

Figure 2.2.11 depicts a schematic of the mold with the laminate and mastic sealant

tape placed on the surface.

25

 

Mastc Seaan'.
Taoe

Figure 2.2.1

 

2.2.3.3 Rel
The la:
that allows
l5 Shown '1,
release fab
doesn‘t all.
laminate.

 

 

ar

E:
l—
f
L

V menu

 

Mastic Sealant Plate Glass Mold Fiberglass Cloth/
Tape Laminate preform

Figure 2.2.11 Schematic with laminate preform in place

2.2.3.3 Release Fabric

The laminate preform is then covered by a release fabric, which is a porous material
that allows resin to flow through it, but does not allow the resin to adhere to it. This layer
is shown in Figure 2.2.12. A release fabric or peel ply is used for this purpose. This
release fabric is normally a finely-knit nylon or other material that repels resin and
doesn’t allow the resin to adhere to it. The release fabric should completely cover all the
laminate. When using a stretchy-type vacuum bag, such as Stretch Vac 350, the release
fabric needs to cover some of the piping components as well as the infusion media.

In some areas along the vacuum manifold, release fabric is used as a resin flow
throttle. Without being saturated with resin, it allows the free-flow of air through the
laminate. However, once it is filled with resin at the vacuum outlet side, it stifles the

resin flow and allows the laminate consolidation process to begin.

 

 

 

 

 

Mastic Sealant Plate Glass Mold Fiberglass Cloth/
Tape Laminate preform
Release Fabric

Figure 2.2.12 Schematic with release fabric placed over the laminate preform

26

 

During
using stretc
stretchy var
the point of
to enter the
mmmafix
was insertet
vacuum bag
3353' flow 0
the helix tul

excellent p3

During this study, an innovative concept was employed for the release fabric. When
using stretchy vacuum bagging material over helix resin inlet piping, it was found that the
stretchy vacuum bag would bridge across the openings in the helix tubing and stretch to
the point of “popping” into the tubing and creating detrimental holes, which allowed air
to enter the system, as shown in Fig. 2.2.13. This phenomenon ruined at least one part
before a fix was found. Release fabric was then sewn into sleeves and the helix tubing
was inserted into them. The reinforcing strength of the release fabric ensured the stretchy
vacuum bag would not deform to failure, yet the porosity of the release fabric allowed the
easy flow of rein to the distribution media. The release fabric sleeves bridged the gaps in
the helix tubing and didn’t allow the stretchy vacuum bag to fail, whereby producing

excellent parts.

may ,

-..u

Helix tubing Helix tubing

’ac. Bag
failure

wit) release
fabric

w release
fabric

 

Figure 2.2.13 Stretchy vacuum bag covering
helix tubing, release fabric supported bag within
the spaces

27

 

2.2.3.4 Si

lna:

laminate .

Ill

Masts?

K
V!!!

Figure j
The i
resin man
easily flm
dISIFlbUIiC

thrOUgh it

2.2.3.4 Surface Distribution/Infusion Media
In a solid laminate, the distribution or infusion media is the next layer on the

laminate stack, as depicted in Fig. 2.2.14.

 

 

 

 

 

 

 

l t , . .,
t . 1.x. . x. , ,‘ci
Distribution/Infusion \
Mastic Sealant media Plate Glass Mold Fiberglass Cloth]
Tape Laminate preform

Release Fabric
Figure 2.2.14 Schematic with distribution/infusion media installed

The infusion/distribution media evenly distributes and delivers the resin from the
resin manifold to the laminate. It acts as a low-resistance piping system for the resin to
easily flow to the laminate preform. Thinner laminates may incorporate a surface
distribution media, such as the three on the right side in Fig. 2.2.15. As the resin flows

through the distribution media, resin is allowed to wet-out the laminate plies below it.

 

Figure 2.2.15 Types of distribution media. Two types of core
distribution media on the lefi. Three types of surface
distribution media in the center and to the right.

28

 

The d
unhindere
media sho
allow the l
composite
media tour
time to inf
touch the \
infusion m

Slstem.

The distribution media should be placed beneath the resin manifold to ensure
unhindered flow of resin from resin reservoir to laminate. However, the distribution
media should not touch the vacuum outlet because the resin may infuse too fast and not
allow the laminate to totally impregnate the laminate preform, causing dry spots in the
composite part. Resin will naturally follow the path of least resistance, if the infusion
media touches the vacuum outlet, then the resin will flow out of the system before having
time to infuse into the laminate. Therefore, it is imperative that the infusion media not
touch the vacuum outlet piping, a layer of release fabric should span a gap between the
infusion media and the vacuum outlet piping, which creates a natural throttle for the
system.
2.2.3.5 Core Distribution/Infusion Media and General Sandwich Construction

However, for some thicker laminates, the infusion media may be placed in the core
of the laminate, where a surface media would deem insufficient and not wet-out the entire
laminate preform. The core infusion media may be made of foam with channeling slits or
other low density material that would allow the flow of resin to the laminate plies on
either side of the core. Examples of core distribution media include DIAB Corporation’s
Divinycell foam distribution media shown at far left in Fig. 2.2.15, above, with the pre-
cut slits, whereas the Lantor Soric’s (white) non—woven distribution media is shown
second from the left in the same figure. As in any sandwich composite, the core is made
up of a low density material, as compared to that of the skins.

The core is designed to be at the neutral center of the laminate, where it doesn’t deter
the strength of the laminate as a whole, but it spreads the strength plies (laminate skins)

apart, creating an “I-beam” effect and increasing flexural stiffness of the laminate. This

29

can be seen in Fig. 2.2.16 and Fig. 2.2.17. Figure 2.2.18 shows the effect of increased

thickness on the stiffness or flexural rigidity, as well as the relative weight added.

Force

Stress

Compression

 

V2 h
Neutral Axis -— -

1/2h

 

 

Figure 2.2. 16 When 3-point bending applied, neutral axis has
no stress, compression on upper face, tension on lower face

Force

 

Figure 2.2.17 Compression and tension on upper and lower
skins, respectively (picture exaggerated for effect)

Since the flexural stiffness of any panel or beam is proportional to the cube of
its thickness, the purpose of the core is to increase its stiffness by effectively
spreading the two faces apart and thickening the beam with a lower density
core. The low density core acts like the shear web in a steel I-beam and the skins

are similar to the I-beam’s flanges. When loaded, such as in a three point bending

30

situatior

tension.

flexural 1

D:

“he

 

for a bear

1:-
I

And 4
Hour
flexing] n,
as the din“;

bemmes [

situation, the upper skin is put into a compression, whereas, the lower skin is placed into
tension. The core is placed into a shear situation. This can best be shown by using the
flexural rigidity, D:

D 2 E1 (2.2. 1)

Where E is Elastic or Young’s modulus and I is moment of inertia of a beam, where,

for a beam, the moment of inertia is normally:

1 = 112-!) . h3 (2.2.2)

And b is the base distance, h is the height of the beam.
However, since there are at least two different materials in a sandwich composite, the
flexural rigidity needs to take into account the thicknesses of the skins and core, as well

as the different elastic or Young’s modulus of the different materials, the flexural rigidity

 

 

becomes [23]:
DzEf b ’+Ef"’"'d2+Ec'b"'3 (2.2.3)
6 2 12
Ef = Elastic/Young’s Modulus of the skins d = Distance between skins’ centroids
Ec = Elastic/Young’s Modulus of the core t = Thickness of one skin
b = Width of beam c = Core thickness

31

Relative

Relative Rigidity
Relative Rigidity 43.1
Rigidity 13.0 t J— ’L

1

m t ‘
T Relative 1“ Relative T if Relative T

   

Weight Weight Weight
1 1.04 1.11
Ef = 29.7 GPa pf =2.2 g/ml b = 10 cm

Ec = 77 MPa pc =0.08 g/ml t = 0.2 cm
Figure 2.2.18 Effect of core on relative flexural rigidity/
stiffness and weight

An example of flexural rigidity can be made using the modulus model in equation
2.2.3, which is schematically depicted in Fig. 2.2.18. For Fig. 2.2.18, these theoretical
laminates are made of woven E-glass and epoxy, where the modulus (E0 is 29.7 GPa.
The total skin thickness remained the same throughout this example. A single skin
laminate of thickness 0.4 cm has a relative stiffness of 1 and a relative weight of 1. If the
skin is split in to two equal parts so that each skin is 0.2 cm and then separated by a 0.4
cm foam core with a modulus of 77 MPa. This gives the laminate a total thickness of
twice its original thickness, 0.8 cm, resulting in a relative rigidity of 13.0 and a relative
weight of 1.04. If the skin is further separated to make a laminate of four times its
original thickness, 1.6 cm, where the core is 1.2 cm, the relative rigidity increases to 43.1
and the relative weight only increases to 1.11. Table 2.2.2 summarizes the increased

thickness, increased flexural rigidity compared to the slight increase of weight.

32

 

Kno
core and
Ion-dens

detracts f1

Table 2.2.2 Relative thickness, rigidity and weight of
sandwich composite

 

 

Relative Relative Relative
Thickness Flex Rigfly Weight
1 1 .00 1 .00
2 13.00 1 .04
4 43.07 1.11
8 97.32 1.18

 

 

 

 

 

Knowing what was discussed above, if a surface distribution media is used as the
core and is filled with only resin, because it has a lower elastic modulus than that of a
low-density core, it will add unnecessary weight. Therefore, the denser epoxy core
detracts from the use of structural composites and should be avoided.

The infusion media is available in various shapes and sizes, depending on the
thickness of the laminate and the porosity and permeability of the laminate plies.
Because there are so many combinations of resins, laminates, and cores to manufacture a
composite, it is advantageous to conduct testing on all composites, especially thicker
ones, to marry the proper distribution media with the laminate. Testing specific
components with VARTM will go a long ways to ensure that the final parts are
satisfactory.
2.2.3.6 Vacuum Bag

The vacuum bag is the last component to be placed on the mold. The vacuum bag
maintains the pressure gradient within the system and ensures that the flow of resin is
constricted from the resin inlet to the vacuum outlet. Without a properly sealed vacuum
bag, the part may encounter excessive voids, air pockets or dry spots. Air pockets and

dry spots are parts of the laminate that are not entirely wet-out or impregnated with resin.

33

 

The vac

outlet rr

sealant I

:37
'{Q

3‘)

[J

The vacuum bag covers the laminate preform, release fabric, infusion media, vacuum
outlet manifold and resin inlet manifold. It is adhered to the mold by means of the mastic

sealant tape. The schematic of the vacuum bag’s location on the laminate is shown in

 

   

 

 

 

I
Flg. 2.2.19.
Resin Inlet
2 Vac um 8
Manifold U 39 Vacuum Outlet
Ma lfold
"2" ‘-.- $fi$¥fii€~ita§$¢bfifififflitlifi‘é-T-Z'.‘:=".“'.' .1 .‘;..-.- -‘ a -_ -1- _' .. . _ ...._... ,
".-:r‘--'2. . i in .
' ' ' ;. >1. .r' re .1". ' ' ' ”.J ' . » .2 _ .
94,2 5. ,2::: 2.},,‘.;:,.:‘_‘ .: _.:.i . . m ' “Le-2'31 a"

 

 

Distribution/Infusion

Mastic Sealant media Plate Glass Mold Fiberglass Cloth]
Tape _ Laminate preform
Release Fabric

Figure 2.2.19 Vacuum bag installed on the laminate preform, release fabric,
distribution media and resin/vacuum manifolds.

The vacuum bag is a tough, non-porous material that will hold a vacuum and
chemically resist the infusion resin. Although there are many vacuum bag materials
available, there are two employed in this study. One is a nylon-based bag that is 2 mils
thick. It is relatively non-flexible, so it is imperative that excess bag is made available to
the mold so that bridging is minimized. Extra vacuum bag material is placed inside a
mold by means of making pleats in the sealant tape, as noted above and depicted in Fig.
2.2.10. This process is tedious and in some cases, tricky. However, it is an important step
to ensure that the part is a success. Another type of vacuum bag that is used is a stretchy
material, called Stretch Vac 350. It will elongate nearly eight hundred percent. It is an
excellent material employed in composites with complex geometry because it will form
to the mold much better than a non-stretchy type. In most cases, pleats are not required

and save time during set-up. However, being flexible comes with its own set of

34

 

problerr
various
filament
infusion
systems
hole. call
stretchy l
cases oft

preduced

problems. Since this type of vacuum bag conforms to shapes easily, it will also shape to
various flow media. In some cases where the flow media consists of only a mono-
filament in thickness, this stretchy vacuum bag can depress into the openings in the
infusion media and cut off the free-flow of the resin. In addition, openings in piping
systems or sharp edges may cause this vacuum bag to deform to the extent of producing a
hole, causing air intrusion in the laminate, trashing the laminate. As noted before, the
stretchy vacuum bag may be protected from damage by another layer of release fabric. In
cases of the helix/spiral tubing used in the resin and vacuum manifolds, a sleeve was

produced to house the tubing.

35

 

2.2.4 Resi
2.2.4.1 RE

Resin
flow medi
of simple I
atmospher
Piping she
process. P
criteria. pll
man}; hard

this Study‘ '

Initiall}
the prefOnn
method-5 “Cl
aorle inch bl
the Piping p-

2.2.4 Resin System
2.2.4.1 Resin Inlet Piping

Resin inlet piping is used to transfer infirsion resin from the reservoir of epoxy to the
flow media which distributes it to the laminate preform. Resin inlet piping can be made
of simple tubing, but the material needs to be stiff enough so as to not collapse when a 1
atmosphere or 101 kPa of pressure is inwardly applied in the form of a vacuum. The
piping should also be transparent so that the resin flow may be monitored throughout the
process. Polyethylene tubing worked great in this research in that it met both of these
criteria, plus it was relatively inexpensive and, being common material, was found in
many hardware stores. Figure 2.2.20, below, shows an example of resin piping used in

this study, along with resin manifold tubing, which will be discussed later.

Resin
). .
l lplng

Resin
Manifolds"
Helix tubing

 

Figure 2.2.20 Resin Piping and Manifold
Initially, the piping or tubing needs to be clamped while the vacuum is compressing
the preform. This may be done by means of clamps. A few resin piping clamping
methods were used in this study, which are shown below in Fig. 2.2.21. This study used
a one inch brass gate valve that was slipped over the tubing; the gate was turned down on
the piping, pinching it off to stop the flow of resin in the line. Another method was a set

of locking pliers that was clamped over the tubing. The pinching action of the gate valve

36

 

and the j
metacu
pmuaHy
resin 00‘
faster res
theresni
preform

flousthn

The Op
ieI'L’i in [he
SFSIE‘m if ii)

and the pliers’ locking jaws pinched the tubing closed and allowed leaks to be detected in
the vacuum bag. The brass valve could also serve a dual-purpose if needed; it could be
partially closed once the resin was flowing to help throttle the resin. In some cases, if the
resin flows too fast, it encourages air bubbles which results in laminate porosity. The
faster resin does not allow all the air to be displaced in the laminate preform. Throttling
the resin flow gives the resin time to properly and thoroughly wet-out the laminate
preform. Since the gate just pinches the tube to clamp it closed, there is no resin that

flows through the valve, which allows it to be used indefinitely.

    

,1.”
l l

 

Figure 2.2.21 Resin piping Clamping Methods, Brass gate
valve & locking pliers

The open end of the resin inlet piping was set into the reservoir of resin. The epoxy
level in the reservoir was continuously monitored so that air is not allowed to enter the
system if the level drops below the tube inlet. If air is allowed into the system, it will

introduce air bubbles into the laminate and cause porosity.

37

 

2.2.4.2 l
The
media. ‘
manifolt‘
been slie
tubing. 1
vacuum 1
resin fror
helical tu
larger are
Situated d
because [I
resistance
thinner di-
quCker 1h
h€llcal [Ul‘

2.2.4.2 Resin Inlet Manifolds

The resin inlet manifold distributes the resin from the inlet piping to the infusion
media. The solid polyethylene resin inlet tubing connects the resin reservoir to the resin
manifold inside the vacuum bag. The manifold consists of polypropylene tubing that has
been sliced in a helical or spiral fashion, thereby being called helical tubing or spiral
tubing. Figure 2.2.20, above, shows an example of the helix manifold tubing used for the
vacuum manifold. The sliced openings in the tube prove to be effective in distributing
resin from the solid tubing to the infusion media. For larger or thicker applications, the
helical tubing may be laid on top of the infiision media to distribute more resin over a
larger area, as shown in the center picture in Fig. 2.2.22. One downside to an inlet being
situated directly over a laminate is that it may leave a slight impression on the laminate
because the tubing acts like a point load. Because resin flows in the path of least
resistance, resin will flow quicker through the half-inch tubing faster than through the
thinner distribution media. The quicker the resin is distributed to the infusion media, the
quicker the resin will impregnate the thicker reinforcing laminate preform. Since the
helical tubing is fully open in the middle, it restricts little resin flow and quickly delivers
resin to the distribution media anywhere it is placed in the system. The effect the helical
tubing has on resin flow and consequential laminate impregnation can be seen in Fig.
2.2.22. Within the vacuum bag, the flow begins at each manifold and evenly flows
outwards and perpendicular to the manifold, as noted by the arrows for each type of
manifold. Figure 2.2.22’s right set-up shows the manifold is set along one of the ends
and the resin infuses the total length of the laminate. In the center manifold set-up, the

manifold is aligned lengthwise on the laminate, the resin quickly transverses the length of

38

 

the lam
manifol
circular
used to

laminatc

gensral indu

the laminate and then irnpregnates the resin outward towards the sides. In the left
manifold set-up, a center gate or port is used, since it is only one port, the resin flow is
circular outwards towards the edges. All set-ups work and a combination of them may be
used to effectively impregnate all areas of the laminate, especially in odd—shaped

laminates, where only one manifold type would not be effective.

Effect of different
positions of distribution
media

Laminate Preforms ET
/. l . . \

  

   

Center Line Center Gate

End Line Resin Manifold Manifold

Manifold

Figure 2.2.22 Manifold placement and effect on
laminate impregnation

Other manifolds may consist of additional layers of infusion media or one layer of a
thicker infusion media at the onset. There is also tubing that is formed in the shape of an
omega, (2, that is also used as a resin manifold, where the small opening at the bottom of
the channel allows the resin to free-flow into the distribution media. The resin inlet
system must be carefully placed to ensure that there is an even resin flow throughout the
preform. Depending on the size of the preform, many resin inlets may be necessary. A

general industry rule is that there should be no more than 75 cm between resin inlets.

39

 

This ge
infusior
remaini
ability t.
research
distribut
there are

such as r

This general rule, as well as others, is noted at the end of Section 2.2. Over 75 cm, the
infusion process may slow down too much and stifle proper infusion of resin into the
remaining laminate in that area. As the resin travels across the preform, it loses the
ability to flow freely at one atmosphere of pressure. This distance needs to be tested and
researched for every infusion configuration, as it will depend greatly on types of
distribution/infusion media, laminate preform fabrics and thickness of laminate. Because
there are many factors that affect the flow of resin in a laminate preform during infusion,
such as resin viscosity, preform permeability, distribution media thickness and vacuum
pressure, it is difficult to prescribe just one type of distribution/infusion media for each
specific VARTM manufacturing situation. As noted before, each VARTM configuration
needs to be evaluated with the various types of distribution media and resin manifolds to
determine what is best for the kind and thickness of laminate preform and resin chosen
for the composite part. The resin manifold should be no farther than 1.25 cm from the
laminate and should be situated on top of the distribution media.
2.2.4.3 Infusion Resin

Because VARTM has many advantages over other forms of lamination processes,
resin manufacturers are stretching the technology to obtain the perfect resin for infusion.
There are many aspects of the resin that the operator must take into consideration before
choosing an infusion resin. Some of the most important aspects are resin viscosity, type,
and gel/cure time.

Viscosity is one of the key attributes of an infusion resin. The viscosity of a resin or
any liquid is how much it resists flow. A resin with high viscosity will want to flow less

than a resin of low viscosity. Since the resin in VARTM is displacing the air in the

40

 

additic
lines it
inlet pi
the vac
With a \
create a
piping F
the diagj
Ill-”initiale-
included
Thcl

The mosl
5879116 a]
[he StETcn
maWrath
IO Shrink;
”Wiener
safe Use in

normally it

laminate, it is important that it flows well so all air voids are removed. Common resin
viscosities range from 300 centepoise (CF) to 900 CR Although resins of higher
viscosities may be used, additional resin inlet manifolds may be needed, as well as
additional flow media. Additionally, higher viscosity resins may not produce even flows
lines in a part. This becomes a problem when a large part is infused and additional resin
inlet piping is needed to adequately cover the entire part. The more inlets and breaks in
the vacuum bag are made, the more the system will be susceptible to air leaks. Resins
with a viscosity between 300 cP and 900 cP will flow well into a laminate preform and
create a uniform flow front. As noted above, the maximum distance between resin inlet
piping ports for resins in this viscosity range is 75 cm, which is discussed and noted in
the diagram at the end of Section 2.2. Beyond this distance, additional resin inlet
manifolds should be incorporated into the system. Some simple guidelines will be
included at the end of this chapter.

There are a few types of resins that may be considered for successful resin infusion.
The most common resin types are vinyl ester and epoxy resins. Vinylester resins contain
styrene and other strong aromatic chemicals, but have proven very successful. However,
the styrene in this resin system can cause minor shrinkage of parts due to the styrene
evaporating once the part is out of the mold. As the evaporating process causes the resin
to shrink, it may also cause inner laminar stresses. Although, this may not be true to all
Vinylester resins, styrene also has a strong odor and special ventilation is needed for its
safe use in the workplace or laboratory. On the positive side, Vinylester resins are

normally less expensive than other resin systems, namely epoxy-based resins. Since resin

41

 

chemist
\‘inylesl
Ept
reactive
as comp
stresses
hardener
a smaller
needed d
iOWer 0d,
Study. G.
Hardener.
reSin in fu;
aliphatic l
HOled abut

quality as,

chemistry is beyond the scope of this paper; that is all to be mentioned regarding
Vinylester resin.

Epoxy resins are commonly used for infusion. They obtain their low viscosity by
reactive diluents, not volatile solvents. Because of this, they infuse well and shrink little
as compared to other resin types used. The low-shrinkage allows for less inter-laminar
stresses in a part. Many epoxy manufacturers produce one infusion resin with various
hardeners to adjust for gel and cure times. This flexibility allows the operator to maintain
a smaller inventory and only change the hardener when a faster (or slower) gel time is
needed depending on size of laminate. Because of the versatility of epoxy resin and
lower odor compared to Vinylester resins, it was chosen as the infusion resin for this
study. Gougeon Brother’s PRO-SET Epoxy 117LV resin was used, along with 226
Hardener, as shown in Fig. 2.2.23. This resin system was formulated specifically for
resin infusion and is a blend of Bisphenol-A/Bisphenol-F epoxy resins and a modified
aliphatic polyamine hardener system. It has a viscosity of 350 cP, well within the range
noted above. Gougeon Brothers is known for their manufacturing quality control and

quality assurance, which are in keeping with ISO 9000/9001 quality standards.

 

Figure 2.2.23 PRO-SET 117 Resin and
226 Hardener

42

 

Table 2.1

?

 

Progen‘v
DenSlty
VlSCOSil}
Mix Rati
Mix Ratl
Pot Life

Working

Gel Tlmi

' Requir

 

Goug
resins. fro
limited [0
handling C
One Inqul
COmbinati
HOWCVeL
Chara-men:
are djf‘f‘erc
thereft'ire‘

An0th
resin. The
to its ”lick.
"aCUum m;

Table 2.2.1 Gougeon Brother’s Inc., PRO-SET Epoxy Technical and Handling Data

 

 

 

 

 

PRO—SET Epox Technical and Handling Da_ta
many 117LV/224 117LV/226 117LV/229 117LV/237* 117LV/239*
Density g/ml 1.114 1.09 1.114 1.078 1.102
Viscosity cP 800 350 310 360 290
Mix Ratio Weight 100233 100:30 100:31 100:30 100231
Mix Ratio Volume 100:34 100:35 100237 100:36 100:35
Pot Life 1009 Min
65° F 29 85 110 281 465
72" F 21 44 61 190 360
80" F 15 25 40 105 219
Working Time @ 72° F 20 min 1 hr 2 hrs 4 hrs 7 hrs
Gel Time @72° F 2.2_5 hrs 3 - 3.5 hrs 5 hrs 7 - 8 r_ir_s 10 - 12 hrs

 

 

 

 

 

 

" Requires at least 125° F Post Cure

Gougeon Brother’s Inc., supplies many types of laminating, infusion and adhesive
resins, from low to high viscosities, as well as other products. Since this research is
limited to VARTM and vacuum infirsion, Table 2.2.1, below, describes some of the
handling characteristics associated with the PRO-SET Epoxy infusion resins. There is
one infusion resin and five different hardeners available for the VARTM process. Three
combinations do not require a higher-than-room- temperature post-cure, and two do.
However, it is noted that all the resin/hardener combinations have increased physical
characteristics by a higher post cure. These physical properties are lengthy because there
are different cure schedules each combination may have to increase its properties;
therefore, all of the infusion resin/hardener combinations are noted in Appendix A.

Another issue that should be taken into consideration is gel time and cure time of the
resin. The gel time of a resin is how long it takes to change from the low-viscosity liquid
to its thicker gel stage, where the consolidation of the laminate is complete and the
vacuum may be removed from the system. If a resin gels before the laminate preform is

firlly infused, the part will be destroyed. It is imperative that the resin remains in the

43

 

liquid E
were c
336 H;
Techn
manui
Weigh
meast
minut
the let
not ca
part.

allow

measl

liquid stage throughout the entire infiision process. In this study, the gel and cure time
were controlled by the hardener used. Since all parts were relatively small, PRO-SET
226 Hardener was used because it gave ample working time and satisfactory gel time.

Technical data on resin and hardener used may be found in the Appendix.

Another key to a good infused part is properly dispensed and mixed resin. The
manufacture’s mixing instructions should be carefully followed, whether mixing by
weight or volume fraction. The resin-hardener ratio range is usually narrow, so a good
measurement system is imperative. The resin should be mixed for a minimum of two
minutes in the pot, carefully scraping the sides and bottom of the mixing container to mix
the resin and hardener fully together. If the sides and bottom of the mixing container are
not carefully scraped, un-mixed resin may enter the system, causing sticky areas in the
part. Completely mixing the resin and hardeners together will alleviate this problem and
allow for a complete and thorough cure. Figure 2.2.24 shows a carefully dispensed,

measured and mixed resin.

 

Figure 2.2.24 Properly measured,
dispensed and mixed resin is key to
a successful composite laminate

Mixil
allow the
However.
the menu
from the l
bubbles c
mixing p(

Then

:7
(go
I.)
[J
I.)

Creates a
resin con:
media. C

Atrr

Ci

Mixing normally produces air bubbles, and the resin should be given some time to
allow the bubbles to float to the top, which in some cases can be a long process.
However, the process may be accelerated by placing the resin into a vacuum chamber or
the vacuum reservoir, where the vacuum will naturally cause the bubbles to evacuate
from the liquid mixture. This process can be a messy adventure if not careful, as the
bubbles can pop and create a thin film of resin at the top of the chamber and around the
mixing pot.

There are many parts to the VARTM system, the flow of the resin is shown below in
Fig. 2.2.25. It is pushed from a high pressure to a low pressure. The vacuum pump
creates a low pressure in the piping and under the vacuum bag. The resin flows from the
resin container into the resin piping through the resin manifold and into the distribution
media. Once there, the distribution media disperses it to the laminates and the liquid

resin displaces the air in the laminate preform.

   
 
  

Mastic
- Vacuum Space
Atmospheric Re’s‘lnrManlfold— Sealant Bag
Pressure 9' u'mg tape

   
   
    
  
 
 
    

 

  

Vacuum
Reservoir]
Resin Trap

Container

Vacuum Manifold —
. . . Helix tubing
Distribution
' Release
Laminate Preform Fabric
Figure 2.2.25 Flow of Resin through the system, from the resin container,
resin piping, resin manifold, distribution media and into the laminate

preform

45

..,_...__—4

N
!V)
{II
’

N
.9
Di
:—

In t
the lami
of air to
the vacu
2.2.5.2 \

The
resin inle
critical as
manifold
System, 0
media m:
Possible I
51”Mar to
“thing,

ThE‘
PTCfOrm‘ i
A layer 0:
vacuUm 0
manifold
lhrOulc m‘

TECeiVe [h
x

2.2.5 Vacuum System
2.2.5.1 Vacuum Outlet Piping

In the same manner that resin is piped into the laminate, air must be evacuated out of
the laminate by similar piping. The vacuum outlet piping allows a steady and open flow
of air to evacuate the system. The vacuum outlet piping connects the vacuum manifold to
the vacuum reservoir/resin trap. It is also made of solid polyethylene tubing.
2.2.5.2 Vacuum Outlet Manifolds

The vacuum outlet manifold is constructed from helix/spiral tubing, similar to the
resin inlet manifold. However, the specifics of the vacuum outlet manifold are not as
critical as the resin inlet manifold. The only fluid that needs to travel through this
manifold is air. Although the helix tubing provides a good flow of air evacuation of the
system, other methods may be used as well. Thick layers of breather cloth or distribution
media may be used as the vacuum manifold. However, to keep parameters as even as
possible throughout this research, the resin inlet and vacuum outlet manifolds were kept
similar to each other. All manifolds were constructed of half-inch polyethylene spiral
tubing.

The vacuum outlet manifold is placed about one to 5 cm away from the laminate
preform, which is included in some general VARTM guidelines at the end of this section.
A layer of release fabric should span the gap between the laminate preform and the
vacuum outlet manifold. This gap allows the flow of air from the laminate to the
manifold, but when the resin reaches the single layer of release fabric, it will stifle and
throttle the resin flow and allow other areas that have not been impregnated with resin

receive the resin. Additionally, if the resin was allowed to free-flow out the vacuum

46

manifol
spots in
' means 0
trap. Tl:
not eniei
consol id

DECCS 53 l'

AMDI
Pressi

 

RQSi
C°ntai

manifold, it could cause the resin to prematurely exit the system and possibly create dry
spots in the laminate. The manifold is then connected to solid polyethylene tubing by
means of “T” fittings. The vacuum flow then continues on to the vacuum reservoir/resin
trap. The vacuum system flow can be seen below in Fig. 2.2.26. Ideally, the resin will
not enter the vacuum outlet manifold and vacuum piping. However, during the
consolidation process, it is best if excess resin flows into this piping system and, if

necessary, continues into the resin trap.

 
    

 
       
    

Mastic
Atmospheric ResinrM‘ani‘fold ‘ Sealant Vchzgm Mold
Pressure e 'x u ""9 tape

   
 

  

  
  
  
 

Vacuum Piping

  
  

Vacuum
Pump

Vacuum
Reservmn‘
Resm Trap

  
    
 
   
 

Container

Vacuum Manifold -
Helix tubin
Distribu ion 9

Media Release
Laminate Preform Fabric

 

Figure 2.2.26 Vacuum flow in a laminate. Once vacuum is pulled to the maximum of
vacuum pump, the flow is stifled until the resin piping is opened.

2.2.5.3 Piping (in general)

Now that both the resin inlet and vacuum outlet piping have been discussed, this
section covers some general guidance on piping systems.

It is imperative that the solid tubing is sealed well against the vacuum bag. These
areas are common areas for vacuum leaks. Sealant tape may be wrapped around the tube

and then stuck to the perimeter sealant tape. This helps reduce openings in the outer seal,

47

which cut
introducti
laminate.
needs to t
otherwise

The l
clamps tl
Sometimi
It is lmpc
the creep

2.2.5.4 \

which could result in unwanted air flow inside the vacuum bag. As noted before, any
introduction of air inside the vacuum bag could lead to air bubbles and porosity in the
laminate. As noted above, if a stretchy vacuum bag is used, then the helix/spiral tubing
needs to be covered with release fabric sleeves or an additional layer of release fabric,
otherwise the stretchy vacuum bag could pop where there is an opening.

The resin inlet piping is to be clamped off with a clamp or valve. Two types of
clamps that were used in this study are noted above, the gate valve and locking pliers.
Sometimes the weight of these clamps may produce an undue strain on the sealant tape.
It is imperative that the clamps are properly supported so that a leak is not generated by
the creep of the mastic sealant caused by the strain fiom the weight of the clamp.
2.2.5.4 Vacuum Gauge

A vacuum gauge can be inserted at one edge of the mold. This will allow constant
observation of the laminate vacuum pressure. The tubing for the vacuum gauge in the
laminate should be sealed as to not allow air into the vacuum bag. Monitoring the
vacuum gauge throughout the process will help ensure that leaks can be detected and
dealt with appropriately.
2.2.5.5 Vacuum Reservoir/Resin Trap

A resin trap is needed to protect the vacuum pump. It ensures that any resin that
transfers to the vacuum outlet manifold and starts to travel in the vacuum piping doesn’t
reach the vacuum pump. If resin reaches the vacuum pump, it can ruin a pump in little
time. The resin trap used in this application doubles as a vacuum reservoir, too. If a
sealed vacuum is obtained on the mold and no leaks are present, the vacuum pump may

be turned off and the vacuum reservoir will maintain the vacuum on the system.

48

 

Howet
bag. (3
many (
iacuun
some V
outfitte
Shows 1

P

 

However, in most cases, there may be some slight leaks due to (1) porosity in the vacuum
bag, (2) vacuum bag not adhering to the sealant tape, (3) joints in the piping system and
many others; a vacuum pump is needed to overcome these minor leaks, whereas the
vacuum reservoir will eventually leak down to atmospheric pressure. Although there are
some vacuum reservoirs/resin traps made for this purpose, a simple pressure paint tank,
outfitted with inlet/outlet nipples and a vacuum gauge, is very effective. Figure 2.2.27

shows the vacuum reservoir/resin trap, both the external and internal views.

    

, ‘L

Figure 2.2.27 Vacuum reservoir/resin trap. Left is an external view. Center is
the opened reservoir containing a tub of excess epoxy that was removed from
numerous VARTM laminate processes. At right is the reservoir/trap’s lid with a
rubber gasket around its perimeter that maintains a head of vacuum pressure on
the laminate.

A quick-disconnect fitting, as shown in Fig. 2.2.28, below, is placed on the vacumn
outlet side of the reservoir helps detect leaks in the system. This is done by
disconnecting the vacuum pump from the system and monitoring the vacuum gauge. If a
drop in vacuum is noted on the gauge, then there is a leak in the system that needs to be
addressed. In addition, the quick-disconnect can be used to release the vacuum on the
system and minor adjustments may be made to the laminate, as some components move

during the consolidation process.

49

factor tha
pressure
piston. 5h
Piston \m
COoler the

Both
enabled t]
the DOISQ
and infllgj
maintain“
lamjnale (

Slack rem

    

, ~~u
Figure 2.2.28 Quick Release Fitting
2.2.5.6 Vacuum Pump

There are many options for vacuum pumps on the market. However, each case
needs to be evaluated and the proper vacuum pump determined. In this study, the main
factor that was considered was that the pump needed to pull one atmosphere or 101 kPa
pressure. Two types of vacuum pumps were employed to do this, a rotary vane and dual
piston, shown in Fig. 2.2.29, below. Although both reach the required pressure, the
piston vacuum pump was quieter and required no lubrication oil. In addition, it ran
cooler than the rotary vane pump.

Both pumps were outfitted to a PVC manifold that spanned across the room. This
enabled the vacuum pumps and the infiision system to be spread out. Being separated,
the noise from the vacuum pumps didn’t interfere with detecting leaks in the vacuum bag
and infusion system as a whole. The vacuum pumps were kept running while the resin
maintained a liquid stage, however, when the resin gelled and no more compaction of
laminate could take place, the pumps were turned off. This ensured that the laminate

stack remained consolidated, the resin content was minimized, the fiber-to-resin ratio

50

 

maximiz

and void

maximized and that the results produced light, strong composites with minimal porosity

and voids.

         

. . !
a?% X

 

Figure 2.2.29 Vacuum Pumps, rotary vane on left, dual

piston on right

2.2.5.7 General VARTM Guidelines

As noted many times in this document so far, there are many, many parameters that
are involved with VARTM manufacturing and testing must be completed prior to any
large-scale project goes forth. However, Figure 2.2.30 shows some general guidelines
that have been noted earlier in this chapter that should be considered to ensure the most
successful infusion possible. Guidelines noted in separate sections are grouped together
and noted here:

0 Resin inlet ports should not be more than 75 cm from another resin port or

vacuum outlet. This distance should be tested before incorporating on a thick

part.

51

Release fabric should extend 2.5 — 5 cm before laminate (resin side)

Release fabric should extend 5 — 7.5 cm past the laminate (vacuum side)

Distribution media should extend 2.5 cm before laminate (resin side)

Distribution should NOT extend past laminate on vacuum side

Resin manifold should be 0 — 1.25 cm away from laminate preform and lay

directly on the distribution media

Vacuum manifold should be no closer than 5 cm from the end of the laminate

Use pleats on edges of vacuum to ensure enough vacuum bag is available to the

laminate and eliminate/minimize bridging

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Resin Manifold - Distribution Media
Helix tubing short
Waxed
= 0.5 cm M .d 5 cm Space
0 between laminate &
< 75 cm J \ vacuum manifold
Release ‘ fi
fabric 2.5 - 5 \ :
V
cm past \ = , -
laminate 4» :
—: e
g. E \ Release fabric 5 —
5." E \ 7.5 cm past
5} ii 5 laminate
—- e— Z :— d-
: g Mastic Sealant tape around
i E i) : “ perimeter of mold and
: 2 wrapped around any tubing
__5 g or other appendages
Extend 5; g.\
Distribution H" (i :
Media / = : \
= 2.5 cm / T'Y ' Vacuum Manifold —
\ \ Helix tubing
0 - 1 .25 cm . Release
space between Laminate Preform Fabric
laminate & resin °°"e”"9

manifold

all
laminate

Figure 2.2.30 Schematic of general guidelines to consider
when setting up VARTM

52

2.3 Prol
Alti
process
V’ARTA‘
fatal to t
with 0th:
2.3.1 Va
met
to enter t
not detec
“11171me
COUid be
Once 3 pr
0f hating
laminate :
resin. its ;
Vet}. 10w ;
Once
Check the
VaCUUm ht
disconnCCI
Ibought i0

2.3 Problems Associated with VARTM

Although there are many benefits associated with VARTM, this manufacturing
process is not without its drawbacks. Even though there is an unlimited set-up time for
VARTM, once the resin is infused into the laminate, any leaks in the system will prove
fatal to the part. Because of this, vacuum leaks will be discussed in detail below, along
with other problems associated with VARTM.

2.3.1 Vacuum Leaks

Inevitably there were leaks in the vacuum bag. Leaks in the vacuum bag allowed air
to enter the system and will not allow a full vacuum to be achieved. If these leaks were
not detected and removed before infusing the resin, they would be a major source of air
entrapment, causing voids and porosity in the infiised parts. In some cases, the porosity
could be so extreme that parts would become damaged beyond repair. Unfortunately,
once a preform is infused with resin in the VARTM process and has the unfortunate fate
of having air entrapment, there is little to be done to remove bubbles or consolidate the
laminate any further without aid of other equipment. When the preform is filled with
resin, its ability to draw non-porous resin through to displace the air-entrapped resin is
very low and this damage is irreversible.

Once a vacuum is pulled on the laminate, the first thing that needs to be done is to
check the bag for tightness. The best way to check if the vacuum bag is holding a
vacuum head is to disconnect it from the vacuum source. As noted above, a quick
disconnect is used on the vacuum reservoir. When the bag is in place and all leaks are
thought to be eliminated, the quick disconnect fitting from the vacuum pump to the

vacuum reservoir may be taken off. The vacuum reservoir will have a head of vacuum

53

still i

Ln
- (4
r/‘r
H
(D

VBCU'

elimi

SOUI

dete

hun

3V3

the

ant

iii

{/1
’?

\\'l

he

still pulling on the vacuum bag. If there are any leaks in the bag or anywhere along the
system, including piping, fittings, holes in bag itself, etc, the vacuum gauge on the
vacuum reservoir will drop. If the vacuum gauge drops, then the leak must be found and
eliminated.

Because of the irreversibility of porosity in the laminate, it is important to remove all
sources of leaks and air entrapment in the vacuum bag. There are many different ways to
detect leaks in a vacuum bag. However, the most common way to detect leaks is with the
human acoustic sensor — the car. When air leaks by a small opening in the vacuum bag, a
high pitched, low frequency “whistling” noise is made. The perimeter of the bag may be
evaluated by placing an ear near the bag, listening for this whistle. When the source of
the noise is found, usually, the vacuum bag may be pushed, pulled or otherwise molded
and formed into the vacuum bag sealant. Forming the vacuum bag against the sealant
will eliminate the gap and cause the leak to disappear. Sometimes, however, additional
sealant may be needed to stop the leak. In some cases, the gap is too large and a
whistling sound may not be made, or the whistling noise is too high a frequency to be
heard by the human ear. The human ear can only detect noises in the frequency range of
20 Hz to 20 kHz of sound wave lengths of 1.9 cm to 17 meters. If the noises are above or
below this range, the human ear will not pick up or be able to audibly detect or pinpoint
the noise’s source.

In these cases, an ultrasonic leak detector can be used, like one shown in Fig. 2.3. 1.
An ultrasonic leak detector can detect wave lengths ranging from 0.3 cm to 1.6 cm long.

Not only can these detectors “hear” the noise, they can also locate them. Once located,

54

the some

tested ag:

2.3.2 Race

Becau
tracking. \i
bridging. i]
be {gum o
vacuum bit
release fab]
the laminar
around [he .'
lili’ pen-met,
Inimé’lliarql'

the source of the leak can be dealt with as noted above. The vacuum system can then be

tested again to ensure all leaks are eliminated and the infusion process may begin.

    

Figure 2.3.1 Ultrasonic lake deetctorsncan be used
to find small leaks undetected by the human ear

2.3.2 Race-Tracking

Because no system is perfect, imperfection in the mold set-up can cause race-
tracking, which is depicted below in Fig. 2.3.2, and caused by bridging. As noted above,
bridging, in any degree, can cause problems in a laminate. Bridging is most notorious to
be found on the sides of a laminate. The bridging can be caused not only by not enough
vacuum bag, but it may also be caused by overhanging distribution media or excess
release fabric draping and/or spanning over the sides of the laminate. In some cases, if
the laminates are not perfectly aligned, the laminate itself can cause a bridging effect
around the perimeter. This is why it is common to experience a “race-tracking” around
the perimeter of the laminate. This bridging, as depicted below in Fig. 2.3.3, creates a
triangular-like shape, if in a 90° sharp comer or an, otherwise, open void along the

laminate’s surface transitions, all which produce an open “pipe” for resin to flow through.

55

As mentit
laminate.
top of the
of the lan
On a lami
Race-trac

bridging i

As mentioned above, bridging can also cause race-tracking if it happens at the edges of a
laminate, where the thickness of the laminate causes a large vertical distance between the
top of the laminate preform and the mold surface. Because of this race-tracking, the sides
of the laminate will wet-out quicker than the areas in the middle of the laminate stack.

On a laminate, the wet-out can form a parabola shape if both sides are race-tracking.
Race-tracking is more prevalent in thicker laminates than thinner ones because potential

bridging is larger.

 

 

 

Types of Bridging on a laminate —
open channels for resin to flow through

Figure 2.3.3 Bridging examples

56

2.3.3 VA

Evei
breach tl
gauge ”it
introduce
part. In 2
sealant am
for air se.
ahhough
the lantin
problem 1‘

made froi

2.3.3 VARTM’s Involved Set-up

Even though VARTM uses an open mold system, there are many components that
breach the vacuum bag, such as resin inlet piping, vacuum outlet piping and vacuum
gauge monitoring. All of these systems must be perfectly sealed to not allow air to be
introduced into the system. Any leak even so slight, will cause air entrapment into the
part. In addition to the systems that breach the vacuum bag, there is also the mastic
sealant around the perimeter of the mold. The perimeter seal is normally a large potential
for air seepage. This is especially true for areas where pleats are needed. The pleats,
although necessary to ensure plenty of vacuum bag material is available for contours in
the laminate, they also have great potential for leaks. The vacuum bag itself can pose a
problem if it is not carefully handled during the process. Unannounced pin holes can be
made fiom many different sharp objects commonly found in the shop. It is imperative
that scissors, knives and other cutting materials are carefully used around the vacuum

bag. Figure 2.3.4 shows a VARTM set-up ready to be infused.

  
   
 
 
 
 
  
 
 

Vacuum Reservoir/

Resin Trap Sealant

Vacuum Piping

& Manifold
MOId & MOId Distribution Media
Frame
Resin Piping &
Manifold
Vacuum Bag var -‘ if ‘ Resin Clamp

“12;. 3*.
Figure 2.3.4 VARTM set-up ready to be infused

57

The
potential

laminatio

The benefits greatly outweigh the problems associated with VARTM. If the
potential pitfalls are known, then they may be carefully avoided throughout this

lamination process.

58

2.4 Safer)

Safetj
noted abo
during the
during eac
laboratory

Eyes:
against ch:
dE-moldin,
glasses or .
machine 31

Hands
Chemicals,
mOldlng pa
hafdeners (
hand Clean.
eI‘OXy fro”.

RESP”;
l0 minimiz‘

Lag”):
Chemicals tl

labeled and

2.4 Safety

Safety in any laboratory or manufacturing situation cannot be over emphasized. As
noted above, there are many types of materials and processing tools that are encountered
during the VARTM process. It is imperative that all safety precautions be observed
during each processing step. The following should be protected while working in the
laboratory:

Eyes: ANSI-rated safety glasses or goggles should be worn at all times to protect
against chemicals, such as resin, hardeners and liquid mold release agents. During the
de-molding process, hard chunks of cured resin can pop off in any direction, safety
glasses or goggles will help prevent eye injury. When using drills, laminate saw, impact
machine and other machinery, safety glasses or goggles must be worn.

Hands: Nitrile or latex gloves must be worn when working with solvents and
chemicals, including resins and hardeners. Leather gloves should be worn when de-
molding parts and cutting specimens, as the sharp edges may cut. If liquid epoxy resins or
hardeners come into contact with skin, the epoxy should be washed off with waterless
hand cleaner and thoroughly rinsed with water. Do not use solvent to remove liquid
epoxy from the skin, as it may cause skin sensitization.

Respiratory: Dust masks should be wom when machining or sanding composite parts
to minimize airborne fibers and epoxy dust.

Lastly, material safety data sheets (MSDS) should be available for all solvents and
chemicals that enter the laboratory or workshop. All chemicals should be properly

labeled and stored so as to not violently react with each other.

59

2.5 Proce

Now

with \'AF

i discussed
l

l. M

fx)

'1
L

Cc

b.)

Ur
j/
p—n

dOmed [an
lammate i:
manufaclu
laminates 1
Willow, I

discussed 5

2.5 Processing Composite Laminated Parts Using VARTM
Now that all the components of VARTM are understood and the problems associated
with VARTM are known, the actual steps to produce a composite laminated part can be
discussed. The basic six steps are:
l. Mold Preparation
2. Laminate Preform (fibers) Cutting
3. Consumables Preparation
A. Release Fabric Lay-up
B. Infusion/Distribution Media
C. Bagging & Sealing
4. Vacuum Compaction
5. Matrix impregnation by Infusion
6. De-molding of Laminate
The next sections will discuss manufacturing flat, cylindrical arched and spherical
domed laminates, all of which require unique changes to the mold to ensure that the
laminate is shaped properly. The flat laminates and the domed laminates can be
manufactured on the same flat mold. The only difference is that the spherical domed
laminates have plugs inserted on to the flat mold surface to obtain the spherical dome’s
contour. The cylindrical arches were manufactured on a separate mold and will be

discussed separately.

60

2.5.] Step 1

As note
adhere to its
excellent ad
mold where
It is easily c

domes stratt

2.5.1 Step 1: Mold Preparation

As noted before, the mold should be waxed and prepared so that the resin will not
adhere to its surface. Epoxy resins are not only good for laminating parts, they are also
excellent adhesives. The key is to make the resin stick where it needs to stick, and de-
mold where it needs to de—mold. Figure 2.5. l , below, is an example of a flat glass mold.
It is easily converted to a domed mold, as shown in Fig. 2.5.2, by carefully placing waxed

domes strategically apart.

  
  

 

Figure 2.5.1 Completely wax the mold to
ensure the resin does not adhere to it

 

66/ N
Figure 2.5.2 Placing spherical domes on the
glass to produce domed composites

6]

Figi

the top V

Figure 2.5.3, below, is a schematic representation of mold preparation step, it shows

the top view of the mold (top) and a cross-section of the mold (bottom).

. Waxed
Mastic Mold

 

 

N

Mast Sealant
Tape Plate Glass Mold Mastic Sealant
Tape

Figure 2.5.3 Step 1 Mold Preparation, both a top view and a cross-section
view of the mold

The cylindrical arches were manufactured on a different mold, consisting of multiple

sections; one such section is shown below in Fig. 2.5.4.

 

Figure 2.5.4 Steel cyiihancéi aic'ii mold

62

Mult
together.
be compli
surfaces.

which C01

Compo
Q

2.5.2 Step

The n.
Prefom I
of lite pmp
with differ.
the remit)“
Piles that a!

in~
“Gimme.

Multiple arched laminates could be manufactured at once by butting mold sections
together, with the longer edges of each section next to each other. Since the mold had to
be completely air tight, mastic sealant tape was inserted between each of the mating
surfaces. Figure 2.5 .5 shows how the molds were mated together to form a longer mold,

which could then be infused.

  
  
 
 

Steel Mold

Foot Valve

 
 
  

Composite Strip

Support Plate

Figure 2.5.5 Mold for cylindrical arches

2.5.2 Step 2: Laminate Preform Preparation

The next step is to prepare the reinforcing fibers that will make up the laminate
preform. Each layer must be carefirlly cut and placed to ensure that the laminate consists
of the proper laminate sequence or schedule. This especially holds true for laminates
with different fiber orientation. When making laminates with different fiber orientation,
the reinforcing fabrics must be cut at perfect angles and transferred to the mold. Any
plies that are skewed off-angle could ruin an experiment and produce faulty results.
Although this study did not encompass any types of cores, if there are cores to be

incorporated into the laminate, the cores would be prepared and stacked accordingly at

63

this time.

the sister

Mast}; SEE
Tape

Fig

CFO:

2.5.3 Sic
Coilsu
bag. Sealing
manifolds a
are 3150 cor.

The VaCUun

this time. Figure 2.5.6, below shows step two pictorially. The top View is a schematic of

the system, and the bottom view is a cross-section of the mold and its contents.

Mastic
Sealant Mold
tape

 

 

 

Laminate Preforms

 

 

~1..-
I

 

 

E Auud

 

m I :21
Ex . ..... l. .l »i
Mastic Sealant
Tape Plate Glass Mold Fiberglass Cloth] Mastic Sealant
Laminate preform Tape

Figure 2.5.6 Step 2 Laminate preparation, both a top view and a
cross section view of the mold

2.5.3 Step 3: Consumables Preparation

Consumables preparation involves preparing the infusion/distribution media, vacuum
bag, sealing system and release fabric preparation. Although the resin and vacuum
manifolds are part of the resin and vacuum system, respectively, the piping and manifolds

are also considered consumables because, for the most part, they are a one-shot item.

The vacuum piping and manifold may be able to be used additional times, however, if

resin reac
prepared "
vacuum 13
2.5.3.1 R:

The l
the lamin.
approxim
laminate ;
at the vac
escape (h.

release {3

resin reaches it, they are no longer considered useable. The manifolds need to be
prepared before the vacuum bag is set in place because they are situated under the
vacuum bag.
2.5.3.] Release Fabric

The first part of step three is to cut the release fabric to shape, completely covering
the laminate and extending past the vacuum end by 5 to 7.5 cm and the resin end by
approximately 2.5 to 5 cm, as shown below in Fig. 2.5.7. All critical distances for
laminate and consumables’ placement are noted above in Section 2.2, Fig. 2.2.30. Again,
at the vacuum outlet side, the release fabric acts as a throttle, it allows free flow of air to
escape the laminate so long as it is not saturated with resin. Once resin reaches the

release fabric, the free flow is throttled and slowed down.

65

Extend Release Mastic Extend Release
Fabric Sealant M°'d Fabric

(04.5 cm) \ tape \ / (5-7.5 cm)

\A .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\\

. Release
Laminate Preforms Fabric

 

 

 

 

 

0—2.5

 

    

Sealant
Tape Plate Glass Mold Fiberglass Cloth! Mastic Sealant

Laminate reform Ta
Release Fabric p pe

Figure 2.5.7 Step 3 Consumables preparation, release fabric lay-up.

2.5.3.2 Resin Infusion/Distribution Media

The infusion/distribution media is also cut to shape and placed on the laminate
preform. It completely covers the laminate, too. However, it extends beyond the
laminate only on the resin inlet side. The distribution media needs to stop short of the

laminate’s vacuum outlet end and laminate edges about a half of a centimeter so as to

66

minimize effects of race-tracking. Figure 2.5.8 shows the addition of the

distribution/infusion media on top of the laminate and release fabric.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Extend Distribution Mastic
Media Sealant Mold
(:7- 2.5 cm) \ ape
I
1
r : it
i
V : n
z 0
l
/% ' i i
/ ..
Distribution Rel
ease
Media Laminate Preform Fabric
0 — 2.5
cm

 

 

 

 

/No Overlap

]

 

I “‘1‘71-373‘.

i w—
—r

Mastic Sealant Distribution/intusmn

Tape media Plate Glass Mold Fiberglass Cloth/ Mastic Sealant
Laminate preform Tape
Release Fabric

Figure 2.5.8 Step 3 Consumables preparation, Distribution/Infusion media
lay-up.

2.5.3.3 Resin Inlet and Vacuum Outlet Manifolds
The resin inlet and vacuum outlet manifolds, made of spiral tubing, need to then be
assembled and placed on the mold; Fig. 2.5.9 shows the schematic their placement in the

system, along with the cross-section of the mold. The resin and vacuum piping are

67

attached to the manifolds. Where the resin and vacuum piping crosses the sealant tape, a
layer of sealant tape is wrapped around the tubing to seal it off when the vacuum bag is
applied.

The resin manifold sits atop of the distribution media so as to allow free flow of resin
from the resin inlet piping to the distribution media. The resin inlet manifold should be
no farther than 1.25 cm from the laminate to advocate the free flow of resin to the
distribution media and on to the laminate.

The vacuum manifold sits atop of one layer of release fabric, about 5-8 cm from the

laminate. Again, once the resin impregnates the release fabric, the flow of resin is stifled.

68

. _ Mastic 5 cm Space
Resrn Manifold Sealant bewveen laminate &

— Helix “’99 vacuum manifold

 

Vacuum

Distribution Manifold —
Media 0 — 1.25 cm , P Release Helix tubing

space between Laminate reform Fabnc

laminate & resin
manifold
0 _ 125 Vacuum Outlet
Resin Inlet cm M ifold
Manifold 1 | z 5 cm

.‘En‘m‘rvt‘mm-nat-zxz-zseeem.. .. - , u; - -~... .5 .15.... .1. "a .

       

Distribution/Infusion
Mastic Sealant media

T Plate Glass Mold Fiberglass CW Mastic Sealant
39° Laminate preform Tape
Release Fabric

Figure 2.5.9 Step 3 Consumables preparation, Installation of the resin and vacuum
manifolds.

69

2.5.3.4 Composite Monitoring Equipment

If a vacuum gauge or a digital thermometer is to be used to monitor the laminate,
these need to be installed on to the mold surface at this time. Any appendages, like these,
also need to be wrapped in sealant tape where they penetrate the mastic sealant tape, too.
2.5.3.5 Resin and Vacuum Piping

The next part to install is the resin/vacuum piping. As noted above, the resin piping
will receive a clamp to temporarily stop the flow of resin until the laminates are ready to

infuse. Figure 2.5.10 shows the piping system being placed on to the mold.

7O

Mastic Space between laminate

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Resin Manifold - Sealant Mold 8. vac manifold
Helix NYE tape I (5 ClTi)
}

‘13 : Vac um Piping
. : l .
Resin ;: i ;
Pipi 9 -§ 5
a
:I 4 :
4p '3'“ = l =

— i ;I :fl
3 :
3' :
:t ‘t :
I l e
Resin __g :
Container 9 i
.I '
v :i ‘i ?\
/ EA : \
/ 5\ ' Vacuum Manifold —
. . . \ ' Helix tubing
Dtstnbutton
Media Space between _ Release
laminate & resin Laminate Preform Fabric
manifold
to — 1.25 cm)
Resin Inlet

   

Resin Piping Vacuum Outlet

1

  
  
  

Vacuum Piping

 
  
 

Distribution/Infusion

media Plate Glass Mold Fiberglass Cloth/ Mastic Sealant
Laminate preform Tape

 

Mastic Sealant
Tape

Release Fabric

Figure 2.5.10 Step 3 Consumables preparation, Installation of the resin and
vacuum piping, and resin clamp. The cross-section of the mold has not changed.

2.5.3.6 Vacuum Bag
The vacuum bag may now be cut and applied to the mold, since the laminate and all
of the consumables have been applied. Figure 2.5.11 shows the addition of the vacuum

bag on to the system. Until this point, the sealant tape should have had its paper

71

protective strip intact. When the vacuum bag is ready to be installed, this paper
protective strip may be removed. It is advisable to only seal the vacuum bag onto the
mold one side at a time. Otherwise, the vacuum bag could get stuck on the sealant tape
inadvertently while its sticky surface was exposed. As noted above, normally pleats are
needed to ensure plenty of vacuum bag material is available to the contours of the mold.
A pleat is created during the vacuum bagging process; there are many different methods
to install a pleat. While the vacuum bag is being stuck to the mastic sealant tape, it is
stopped short where a pleat is to be inserted. A separate piece of mastic sealant tape (still
with one side with its protective paper on) of about 5 to 10 cm in length is cut from the
sealant roll. It is laid, paper side down, on top of the mastic already on the mold. The
paper will prevent it from sticking to each other. Next, the vacuum bag is then stuck only
to this new section of mastic sealant. Then the vacuum bag is doubled back on to itself,
taking the small section of mastic with it. The protective paper is removed from the
section. About 1 cm of the end is rolled over onto itself. The vacuum bag is then folded
over the other side of the mastic sealant tape and continued on to the mastic already on
the mold. The vacuum bag can then continue to be applied around the mold’s perimeter,
ensuring the resin and vacuum piping, as well as the monitoring appendages, are
completely sealed. Figure 2.5.12 pictorially shows the steps in making a pleat in a

vacuum bag.

72

Mastic
Resin Manifold — Sealant VaCuum
Helix tu - ing tape B

   
 
     
 
   

Vacuum Piping

Resin
Container

Vacuum Manifold —

. . . Helix tubing
Dtstnbutton .
Media , Release
Laminate Preform Fabric
Resin Inlet
Resin Piping Manifold Vacuum 339 Vacuum Outlet
Ma . Id

    

Mast-ifaielant media Plate Glass Mold Fiberglass Cloth] Mastic Sealant
Ta

Laminate preform pe

Release Fabric
Figure 2.5.11 Step 3 Consumables preparation, Vacuum bag installation and
sealing the mold and all its components under the vacuum bag.

73

 

 

 

 

 

 

 

 

 

 

 

  
   

 

 

U l D
i P Protective I -
Mastic Sealant aper rin Plate Glass Mold Mastic Sealant
Tape 00"" 9 Tape
(I l
P r Protective .
Mastic Sealant ”:0an Plate Glass Mold MasthlcaSealant
Tape Removed pe
Vacuum Bag / ' "TEN-a... ,5
Mastic Sealant Plate Glass Mold Mastic Sealant
Tape Tape
" . _ Mastic-to-become—
"" * - ... . Pleat \
Vacuum Bag/ me/
an? I; - ...~ ,,_,-1_‘_," '. _ ..... . ' '2
Mastic Sealant Plate Glass Mold Mastic Sealant
Tape Tape
J9.
. .0”. 3
Vacuum Bag \ /
CI v
Mastic Sealant Plate Glass Mold Masti'c Sealant
Tape ape
Vacuum Bag \\
Mastic Sealant Plate Glass Mold M3590 5933'“
Tape Tape

Remove
protectiIe paper

   

 

.\
c
“\1.

Th.

Vacuum Bag\

 

 

_..____.E

 

Mastic Sealant

 

Plate Glass Mold M35“; Sealant
Tape ape
Finished Pleat
Vacuum Bag
(4 —“-.‘" j)

 

 

Mastic Sealant

Plate Glass Mold
Tape

Mastic Sealant
Tape
Figure 2.5.12 Inserting a pleat into a vacuum bag to gain extra vacuum bag within

the mold to avoid bridging of vacuum bag

74

2.5.3.7 Vacuum Reservoir and Pump

The resin inlet piping should also be clamped off with locking pliers or other means
mentioned above. The vacuum piping should be connected to the vacuum reservoir. The
vacuum reservoir should also be connected to the vacuum pump. The vacuum reservoir
then needs to be connected to the vacuum pump, again, checking all piping joints and
wrapping with sealant tape to eliminate any possible leaks. The schematic should look
similar to the diagrams in Fig. 2.5.13.

Mastic
Resin Manifold — Sealant

Vacuum Manifold -
Helix tubing

    

Distribution

edla Release

Laminate Preform Fabric

Figure 2.5.13 Step 3 Last step before vacuum compaction — connect vacuum
and vacuum reservoir to the mold, via vacuum piping.

2.5.4 Step 4: Vacuum Compaction

After the entire system is ready to be compacted, the vacuum pump can be turned on
and the vacuum bag is drawn down to the mold, compacting and squeezing all
components inside the vacuum bag. It is imperative that while the vacuum bag is being

drawn down, all geometric shapes are massaged to ensure that the laminate, release

75

fabric, distribution media and vacuum bag all nestle closely to the contours. Transitions
from the mold to geometric shape is where bridging can and will occur if not carefully
worked. Since the pressure in VARTM is limited to one atmosphere (101 kPa),
therefore, the steep transitions need to be worked to minimize bridging. Figure 2.5. 14
shows the schematic of the applied vacuum pressure.

Mastic
Resin Manifold — Sealant
Helix tu -l tape

    

Atmospheric
Pressure

    
   
 
  
  

   
  

Container

Vacuum Manifold —

. , Helix tubing

Distribution
is Release

Fabnc

  

Laminate Preform
Figure 2.5.14 Step 4 Laminate compaction, time to check for leaks.
Figure 2.5.15 shows an actual domed laminate ready to be checked for vacuum
leaks. Once the compaction of the vacuum bag on the laminate is satisfactory, the system
needs to be tested for leaks with the various methods mentioned earlier in this chapter.

Afier all vacuum leaks are addressed; the resin may be infiised into the laminate.

76

      

41

Figure 2.5.15 Step 4 Laminate
compaction and satisfactory leak
test, ready for resin.

   

2.5.5 Step 5: Matrix Impregnation by Infusion

The resin is now ready to be dispensed into a container at the proper ratio and
amount and then thoroughly mixed. The resin inlet tube is placed inside a resin reservoir.
It is imperative that the resin reservoir is not higher than the mold. If this happens, then a
siphon effect could be produced and the resin may fill the bag at a faster pace than the
vacuum pressure can push it through. This could lead to excess resin in the mold,
causing resin—rich pockets and may even cause the laminate preform, release fabric and
distribution media to float inside the bag and become misaligned. Once the resin
reservoir is filled with the laminating resin, the clamp may be removed and the infusion
process begins. This process is depicted below in Fig. 2.5 .16, which shows the top view
of resin flow, as well as the cross-section of the laminate, also showing the flow of resin.

After the resin impregnates the entire laminate preform, the resin piping clamp may
be applied to the resin inlet piping and further laminate compaction can take place.
During this process, any excess resin is drawn out of the system into the vacuum outlet

lines and into the resin trap.

77

The vacuum is maintained until the resin gels, and then it may be shut-off and
removed. The resin is then allowed to continue to cure to a hard solid. Once cured, the

de-molding process may begin.

  
   
   
  
 
   

Mastic
. . v on s 06
Atmospheric RezlnrManlfoid - Sealant aBagm betwpa
Pressure 0 'x tu ' lng tape '

R
Container

Vacuum Manifold —
Helix tubing

Distribution
Media Release
Laminate Preform Fabric
Resin inlet
Resin Piping Manif°ld Vacuum Bag Vacuum Outlet
Ma ' Id

1 l Vacuun<iping

 

 

 

 

" \\ W'.

”as": sea'an‘ media Plate Glass Mold Fiberglass Cloth] Mastic Sealant
ape Laminate preform

Release Fabric

 

Figure 2.5.16 Step 5 Matrix impregnation by vacuum infusion. (red = resin flow)

Figure 2.2.17 shows a series of pictures of an actual VARTM manufacturing process
happening. The E-glass laminates (left to right) are [90](, unidirectional, [0](,

unidirectional, [0/90]; coarse weave, [0/90]5 fine

78

weave .

.....

    

   

     

1 ' ‘\ i \ ', - .
.Has ~i - _ l . ‘2 “‘r’w. ..‘ , . H}. - ', .. “1"“

 

Figure 2.5.17 Step 5 Resin infusion via VARTM.“FF-vglass-Iaminates on each mold. L t
R: [90k Uni, [0]6 Uni, [0/90]; Coarse weave, [0/90]5 Fine weave

0

2.5.6 Step 6: De-Molding of Laminate

Afier the resin is cured, the consumable layers may be carefully removed from the
mold and laminate. Care should be taken as to not over-stress the laminate and mold,
especially in the case of a glass mold. The release fabric, distribution media and vacuum
bag can normally be peeled off with minimal effort.

Normally, cured resin is hard and sometimes produces a sharp edge, due to tapering
of vacuum bag from thickness of laminate to nothing. Gloves should be worn to protect
hands from the sharp laminate edges. In addition, protective eyewear should be used as
small resin particles could snap off in any direction. Figure 2.5.18 shows the laminate
with all of the consumables removed. The laminate can now be cut into samples and

prepared to be tested and evaluated.

79

Vacuum Piping

    
 

     

a u
I I

«a»
Vacuum
Reservoir]
Resin Trap
Resin
Container

 

 

 

Composite Laminates

Figure 2.5.18 Step 6 De—molding of cured composite laminates

2.6 Domed and Arched Laminates

As compared to flat laminates, spherically domed and cylindrical arched laminates
pose some unique challenges in the manufacturing process. Although some have already
been discussed, this section will go into a little more depth of the particulars in each one.
2.6.1 Unique Attributes to Spherical Domed Laminates

Spherical domed laminates take some extra steps in the manufacturing process. The
first step, as noted above is the mold itself. The flat mold must be changed by adding
domed plugs to it. These plugs need to be made of a material strong enough to not
collapse from the pressure of one atmosphere; the plugs were constructed of epoxy,
thickened with thermoplastic micro-balloons, as described below. Once the plugs are
manufactured, they need to be prepared in the same manner as the rest of the mold. Each
plug needs to be well waxed, especially because the material used will normally be more

porous than that of glass.

80

In this study, a ball was used as the initial plug. It was waxed and set into a mound
of thickened epoxy. Once cured, the ball was removed and a perfect concave
hemispherical void or mold was produced. This concave shape was then waxed. A
mixture of epoxy resin, hardener, and low-density thermoplastic micro balloons were
formed into a catsup-like consistency and poured into the cavity. The thickened resin
was then allowed to cure. Once cured, the plug was removed from the mold and the flat
surface machined until the desired height was achieved. The epoxy/thermoplastic micro-
balloons plug allowed easy shaping, yet would not fail when one atmosphere pressure
was applied. Multiple spherical domes were created to conduct VARTM experiments.

Extra reinforcing fabric also had to be taken into account for domed laminate
preforms. Since the width of the mold now incorporated additional surface area of the
dome, extra fabric was needed to ensure the final laminate was not too narrow.
Sometimes, the fabric would wrinkle and distort around the transition edge from flat
panel to the base of the dome. To make this transition smoother, wax or thickened epoxy
could be filleted in this corner. Sometimes there is not enough pressure to draw the
laminate preform, release fabric, and distribution media tight against the mold in these
tighter comers. In these cases, the vacuum had to be relieved from under the vacuum bag
and the preform manipulated by hand to allow the vacuum bag to reach these transitional
corners. If the voids are not manipulated out, it is a source for bridging to occur.

Once the domes were incorporated into the mold, it was interesting to see how resin
would flow up and over the dome. After the first try, the VARTM proved itself able to
infuse over the domes; no other problems regarding infusion over the domes were

encountered.

81

Removing the domes also took some extra steps. Because the domes were not
adhered to the mold, a thin gap remained between the bottom of the dome and the top of
the plate glass mold. This gap could be filled with wax for future VARTM
manufacturing sessions. However, in some of the samples, when infused, this gap filled
in with resin, causing the plug to de-mold with the laminate. A utility knife was used to
score through the resin skin around the circumference of the dome plug, which allowed
the dome to be removed from the laminate. If the domes were not waxed, they would
have been a permanent part of the laminate.

2.6.2 Unique Attributes to Cylindrical Arched Laminates

As noted above, the cylindrical arches were produced with a completely different
mold than the flat and domed panels. It was important to keep the mold sections tightly
together and sealed with mastic sealant. Without the sealant, the laminate would have
been plagued with air voids and porosity. Three arched molds and one flat mold were
mated together for the VARTM process. Figure 2.6.] shows an actual mold that was

used to manufacture arched composites.

82

 

Figure 2.6.1 Mold for manufacturing arched laminates

The VARTM process ended up infusing slower over the arched laminate because of
the continual ascending and descending flow path and the extra total length of fiber
preform needed to cover the arches.

Before the infusion resin was allowed to impregnate the laminate preform, the
laminate needed to be massaged during vacuum compaction step. Because there were
transitions from the horizontal to the vertical direction, the potential of bridging was
possible. Figure 2.3.3, above, shows the effects of bridging and the related section
describes how it is created and avoided. Because of this, the laminate, release fabric,
distribution media and vacuum had to be carefully designed and checked to ensure no
bridging occurred.

The de-molding process was easier than that of the domed laminate because the steel
arches used in the mold were connected to the horizontal base. When de-molding, there

was very little resistance from the mold and the specimens were removed with case.

83

2.7 Conclusions on Manufacturing

As noted above, there are many key issues involved with vacuum infusion. The tips,
techniques and steps noted above are some of the basic processes needed for quality
vacuum assisted resin transfer molding. It is imperative that testing be conducted to
ensure that all elements are evaluated. One of the most important features would be how
well the laminate preform is infused with the chosen resin. In VARTM systems,
inadequate spacing between multiple resin inlet ports or spacing between a resin manifold
and an adjacent vacuum manifold could prove catastrophic if too much of a gap existed
so that resin was not allowed to saturate the laminate preform before the resin gelled. For
instance, a specific preform, distribution media and resin combination was engineered to
have 60 cm between resin inlet ports to infuse in fifteen minutes because the resin gel
time was only thirty minutes (assuming a safety margin of 2). In one section of the
laminate, spacing was overlooked. There was a gap between a resin port and a vacuum
outlet manifold, say 75 cm. The results could be that he last 15 cm of laminate may not
get saturated with resin before the resin gels and become an air void. If not detected soon
enough in the infusion, the laminate might have to be scrapped.

Table 2.7 lists the specific components that were used in this study. As noted before,
that the unique configuration of each laminate and resin must be tested to ensure that all
components will produce the desired results before a large part is manufactured with

VARTM.

84

Table 2.7.1 VARTM Components used in this research

 

 

VARTM Components Used
Step 1 Mold Prep Mold: Plate Glass
Wax TR High Temp Mold wax

Mastic Sealant Tape

Mold Scraper

Yellow Econotape
Razor Blade scraper

 

Step 2 Laminate

Preform Prep Reinforcing Fibers

Cutters

18 oz/sq yd -Coarse, plain Woven E-glass
12 oz/sq yd - Fine Woven E-giass

11 oz/sq yd - Unidirectional E-glass
Scissors and rotary cutters

 

Step 3 Consumables Release Fabric
Infusion Media

Richmond Air Products Peel Ply
EnkaFusion 7001
Richmond Air Products 350 deg vacuum

 

 

Vacuum Bag bag
Manifolds Airtech spiral tubing
Piping 1/2" OD Polyethylene tubing
Clamps 3/4" brass valve and locking pliers
Step 4 Vacuum
Compacton Vacuum Pump Rotary vane and dual piston vac pumps
Airtech Resin Trap & modified pressure
Vacuum Reservoir paint tank
Step 5 Resin
infusion Resin PRO-SET 117 LV Epoxy Resin

Hardener

PRO-SET 226 Hardener

 

 

Step 6 Demolding Wedges

WEST SYSTEM stir sticks

 

85

 

3. COMPOSITE MATERIAL PERMEABILTY AND POROSITY STUDY

This chapter of research is focused on two studies — calculating the permeability of
various fabrics and calculating the void content of infused composites. The testing
techniques, data analysis and experimental results regarding permeability and porosity
will be discussed. Lastly, a discussion on the effect of fiber orientation, fiber geometry
and structural geometry will be presented.

The permeability research was done by studying the resin flow through various
fabric types and fiber orientations. A non-woven unidirectional fabric has simple fiber
geometry although fiber orientation can be different. A woven fabric, however, changes
its fiber direction each time its weaving tows cross over each other.

The porosity research was completed by determining the theoretical density and
calculating measured density of the material components, whereby the void content could
then be calculated from these two densities. Preform geometry was studied by
incorporating structural shapes into the laminates, specifically a spherical domed shape
and a cylindrical arch shape. Fiber orientation was altered by changing the angle of the
fibers as compared to the flow of the resin. By changing the composite geometry, fiber
geometry and fiber orientation, the flow of resin in the VARTM process could be better
understood.

3.1 Permeability Testing

Permeability is the rate of flow of a liquid or gas through a porous material. Darcy’s
law was used to study the flow of the resin through a porous laminate preform.
Permeability is an important characteristic of a material because it will help determine

how fast, or slow, a resin will impregnate it. Once the permeability of a fabric is

86

determined, it could then be used in various off-the-shelf modeling software packages to
predict flow through more complex laminates, which is extremely useful when large
composites are being manufactured.
3.1.1 Experimental Set-up

The VARTM set-up, as described in Chapter 2, was used during test. The mold was
set into an open-framed table that could pivot 180 degrees. Once a vacuum was pulled
on the laminate, the whole mold system was rotated 180 degrees, where the underside of
the laminate could be clearly viewed through the glass mold. After the resin inlet piping
clamp was removed, the flow front of the resin through the fibers could be captured by a
digital video recorder, which was mounted above the laminate. The fibers changed color
when they were impregnated by the resin; the color change was distinguishable by the
recorder. The vacuum pressure and time for complete infusion was recorded for each
sample.
3.1.2 Permeability Determined by Darcy’s Law

The flow of resin in reinforcements during the processing of composites can be
described by Darcy’s law. This model describes the flow of Newtonian fluids in porous
media and states that the volumetric flow rate (Q, in cm3/sec) through a constant area
specimen is proportional to the permeability (K in cmz), cross-sectional area (A in cmz),
pressure difference over the specimen (AP in Pa) and inversely proportional to the length
in the flow wise direction of the specimen (L in cm) and the viscosity (p. in Centipoise,
Cp) [24]. For one-dimensional flow, Darcy’s law is written:

_K-A-AP

Q #‘L

(3.1.1)

Rearranging Equation (3.2.1), we can directly calculate permeability:

87

 

KZQ'IPL (3.1.2)
A-AP

Since the volumetric flow rate is cross sectional area times the length of the specimen
divided by time, the two areas can be cancelled, leaving the lengths squared.

A-L

 

Since Q = , substituting Q into Equation (3.1.2), the following result is found:

K_ A-L-p-L_y-L2
t-A-AP t-AP

 

(3.1.3)

Lastly, the reduction of units is calculated:

_ Cp - cm2 .0.001Pa - sec
sec- Pa le

K = 0.001-cm2 :>1x10-9m2 (3.1.4)

 

Once the VARTM process was completed, the results from the digital video recorder
were evaluated to determine the resin flow front progression through the laminate. The
time it took the laminate to be totally impregnated by the resin was recorded. The wet-
out area of the laminate could then be plotted against time to obtain a flow curve and the
laminates could be compared.

Laminate permeability is derived from Darcy’s Law and noted above in Equation
(3.1.3). Obtaining permeability for the laminate preforrns indicates how well they are
able to be infused with the chosen resin. By comparing laminates with different fiber
orientations and fiber geometries, engineers could choose the type of laminates to be
included into a composite part. Using numerical analysis and fluid dynamics, a flow
model could then be used to determine how fast a particular laminate would be infused
and if, or where, potential dry spots would develop.

It was theorized that fiber orientation, fiber geometry and geometric shapes

introduced in the laminate would have an impact on a laminate’s permeability. The

88

structural geometries would not only add extra volume to the fiber preform, but also the
changing of direction for the flow front that was thought to have an effect on
permeability.

3.1.3 Test Results - Flow Diagrams and Permeability

Flow diagrams for the laminates can be seen in Fig. 3.1.1 on the next page. The data
is plotted resin-impregnated area (cmz) versus time (sec). The general trend is that the
coarse and fine woven fabrics wet out faster than the unidirectional fabrics, with a couple
exceptions. The domed laminates are noted with dashed lines and the flat laminates are
solid lines.

The fine woven laminate preform was initially the fastest to be impregnated with
resin; however, the coarse laminates overtook wet-out after about 90 seconds. With the
exception of the [i45]3 coarse woven domed laminate, the other three domed laminates
infused at a slow rate as compared to similar-fabric flat laminates. This could be
attributed to the resin having to change flow direction when impregnating the domes,
moving from a horizontal flow to an upward angular flow over the dome shape. Most of
the coarse laminates’ flows started to drop off as the flow front moved away from the
initial condition. The unidirectional laminates, however, show a steadier slope
throughout the experiment. This could be due to effects of micro channels in the woven

fabrics that will be discussed later in this chapter.

89

VARTM Flow Front Trend Lines
Area vs Time

 

    
  
 

Coarse Woven 3“ Fine Woven Fabric

Fabrics

   

Bro-9613 CcErsa '
7 I [0-90]5 Fine Weave

A [90]6 Unidirectional

I [0]6 Unidirectional

Ci [OMS/901$ Unidirectional
" 0 [451-4513 Unidirectional
D [451-4513 Coarse Weave
X Domed [451-4513 Coarse
X Domed [0/9013 Coarse
O Domed (0)6 Uni
-Domed [[4516 Uni

Area cm2

 

 

 

 

0 1 00 200 300 400 500

Time See

Figure 3.1.1 Area vs. Time curves for flat and domed laminates of unidirectional and woven
fabrics

Figure 3.1.2 shows an experiment of all unidirectional fabrics whose orientation
angles are 0°, 30°, 45°, 60° and one specimen whose fibers were stretched to obtain
“straighter” fibers in the 0° fiber orientation. All of the laminates were generally bundled
close to each other in this diagram. However, it is noted that the specimen with the 30°
fiber orientation was the fastest to impregnate with resin through most of the experiment,
Although, the preform with the stretched 0° fibers had the overall fastest infusion. The
preform with the un-stretched 0° fibers was the slowest to be wet—out. The laminate
preforms with 45° and 60° fiber orientation had a similar resin infusion with each other;

however, they were close to that of the un-stretched 0° laminate, too. There are no

90

definite trends that can be specified for the fiber angle, as the slopes are somewhat

sporadic.

However, regarding the comparison between the domed and flat laminates, there is a

definite difference between resin flows. Both of the domed unidirectional laminates

infused at a slower rate than any of their flat-specimen counterparts. As noted above with

the coarse domed and flat specimens, this difference is due to resin having to change

direction from a horizontal flow, transitioning to the vertical direction over the dome.

700

VARTM Flow Front Trend Lines
Unidirectional Flow Fronts

 

600

400

Area cm2

200

100 <—

 

u I ’ — 2 o [3016 Uni 7 A
2 ’ I [45]6 Uni
' , r; ' ’ A [60]6 Uni
. e o [0]6 Uni Unstrelched
n [016 Stretched
, ,, o [016
.- 0 + [90]6 Uni
' o [9016 Uni Dome
% A [016 Domed Uni
o i- - - Poly. ([9016 Uni Dome)
i: - - Pow- «016 Wynn

 

 

100 200 300 400 500
Time Sec

Figure 3.1.2 Area vs. Time curves for flat and domed laminates of unidirectional only

fabrics

The permeability results, as calculated by Equation (3.1.3), from the VARTM

manufactured composites are shown below in Fig. 3.1.3. In this thesis, permeability is

measured in square meters, even though it is not an “area” measurement. Again, the

91

higher the laminate’s permeability, the quicker the resin flows through it and impregnates

it. Lower permeability means that the flow rate is slow.

Laminate Preform Permeability

 

2.47

Permeability x 10'8 m2

 

 

so \
°p\ ‘E’ 6 Op
\9\ \Q \ \

Figure 3.1.3 Permeability of flat and domed laminates

Another unit of permeability is a Darcy (D), named after the French Scientist, Henry
Darcy. This unit is normally used in petroleum engineering and geology, regarding rock
permeability. One Darcy is equivalent to 9.86923><10_l3 In2 or 1 m2 is equal to 1.01325
X 1012 Darcies. This unusual measurement is based on Darcy’s early research on water
flowing through a column of sand. A porous substance has a permeability of l Darcy if,
in 1 second, 1 cubic centimeter of a gas or liquid with a viscosity of 1 centipoise, like that
of water, will flow through a section of l-centimeter thick with a cross section of 1
square centimeter, when the difference between the pressures on the two sides of the

section is 1 atmosphere [25] and [26]. This is best shown pictorially in Fig. 3.1.4, below.

92

Time = 1 Second

 

V

 

 

 

 

 

Definition of unit of

Vol = 1 cm3 Permeabili = 1D , _ .
Vo| = 1 1:ng permeability.
1: =1 cP
Dist = 1 cm Darcy
Apatm=1 atm
Area = 1 cm2

9.86923 x 10-13 m2

Figure 3.1.4 Pictorial definition of permeability unit of measure, Darcy

93

3.2 Effects due to Fiber Geometry

This study focused on three different fiber geometries and up to four different fiber
orientations. Although each type of fiber geometry and fiber orientation produced
different permeability results, there were some similarities within certain groups. Table

3.2.1 shows the various characteristics of the E-glass used in this study.

Table 3.2.1 Characteristics of E-glass used in this study

 

 

 

Fiber Fiber Tow Tow
Type of Density Density Weight Width Thickness
Mfg Calc
Fiber g/cc g/cm"3 glcmAZ mm mm
Coarse Woven E—glass 2.55 2.089 610.292 5.08 0.4318
Fine Woven E-glass 2.54 2.106 406.861 1.016 0.2032
Unidirectional E-giass 2.55 2.318 372.956 2.3622 0.2794

 

 

3.2.1 Fiber Geometry

There were three types of E-glass fiber geometry studied in this research, fine
woven, coarse woven, and unidirectional. The fine woven E-glass was a plain woven
cloth of 12 ounces per square yard. The coarse woven E-glass was also plain woven, but
was 18 ounces per square yard. The fiberglass tow bundles in the coarse woven fabric
were much bigger than that of the fine woven cloth and had a weight of 18 ounce per
square yard. The unidirectional E-glass was a pliable material with little defined tow
bundles, meaning that in most fabrics, the tows are well defined and are able to be
separated easily from the fabric, however, with this particular unidirectional fabric, the
tows run together and there is not much differentiation between the tows and not much
relief across the top or bottom face of the fabric. It was tacked together with a non-

structural adhesive binder about 4 cm apart and weighed 11 ounces per square yard.

94

Figure 3.2.1 shows examples of the three types of E-glass used in this study. Note the

difference between the tows of the fine and coarse fiberglass.

Unidirectional
E—Glass

Fine Woven ,Jfiy 1m,
efflueemrgmr.

.‘i‘ ,
.q ' .g- :aan
E-Glass . f"? wjfwuew $3.1,“

Coarse Woven
E-Glass

 

Figure 3.2.1 Different fiber geometries used in this study

The tows of the coarse woven E-glass are five times as wide and a little over twice
the thickness as the fine woven E-glass. As compared to the unidirectional fabric, the
tows of the coarse woven E—glass are a little over two times the width and 1.5 times the
thickness. The unidirectional E-glass is a little over 2.3 times the thickness as the fine
woven E-glass and a little less than 1.4 times the thickness. Table 3.2.1, above, notes
these dimensions in tabular form. Other composite characteristics as calculated in this
chapter are noted at the end of Section 3.5.

During the permeability testing, the channeling of resin with woven cloth was much
more prevalent than that of the unidirectional cloth. This is because of the unit cell in the
fiber geometry. In a plain woven cloth, a unit cell is made up of four individual fiber
tows, two tows make up the warp and two tows make up the fill, as shown below in Fig.

3.2.2. The open space in the middle of these tows is called the interstices. Because each

95

warp tow must “bend” over the neighboring fill tow, and vise versa, natural resin
channels are created in the fabric. These channels are also best shown in Fig. 3.2.2. If
the fabric is oriented in a 0° or 90° to the flow direction, there is naturally a direct flow
channel from one end of the laminate preform to the other. The larger or coarser the tows
are, the bigger the channels become. However, in smaller woven cloths, the channels are

smaller due to the size of the individual warp and fill tows.

Fiber To“ 5 \ lilICI‘sliccs

Resin /

Channels

 

Figure 3.2.2 Woven E-glass’ Unit Cell

In contrast to the woven cloth, a unidirectional fabric does not have any fill tows.
The unidirectional E-glass used in this study did not have definable tows. Instead, the
tows were somewhat flatter and the bundles of glass filaments ran together. This is
depicted best in Figure 3.2.1, above, containing the three different styles of E-glass. The
fiber tows in the unidirectional E-glass do not form separate bundles of fibers like the
tows in the woven material. Because of this characteristic, when the unidirectional E-

glass was subjected to compaction under a vacuum bag, the laminates were so tightly

96

compacted in some areas; it prevented thorough saturation of resin, which created dry

spots, as evidenced in Fig. 3.2.3, below.

 

Tightly compacted fibers

Figure 3.2.3 Dry areas of unidirectional preform surrounded by
saturated laminate due to tight compaction of pliable fiber bundles

However, if unidirectional fabric had definable tows, the resin channels would
theoretically extend along the entire length of the specimen when the fiber orientation is
0° to the resin flow direction. A unidirectional fiberglass cloth was chosen early in the
testing that had pliable and more un-definable tows as compared to the tows of the woven
specimens. Figure 3.2.2, above, shows close-up pictures of each of the E-glass material
used. Because these tows were not as defined in other types of unidirectional fabrics,
when this material was submitted to vacuum pressure, the fiber bundles squeezed
together and closed any recognizable resin channels. Theoretically, a unidirectional
fabric should have a high permeability because of the longer resin channels. However,
with the pliable tows used in this study, often times the resin was not able to channel
straightforward and had to take a detour. This was evidenced by dry areas contacting the
glass mold surface. During some unidirectional VARTM sessions, a dry area was

surrounded by wet-out fabric, and didn’t become saturated with resin until left to “soak”

97

and wick in resin from the surrounding areas. However, some of these fibers never were
saturated with resin, which become voids and porosity in the laminate, detracting from
the composite’s performance. Figure 3.2.3 depicts this phenomenon of resin starvation or
under-saturation.

Throughout this research, fiber orientation had an effect on all aspects of testing,
especially permeability. This is believed to be mainly because the orientation changed the
nature of channeling.

3.3 Effects due to Fiber Orientation
3.3.1 Fine-Woven E-glass Fiber Orientation

As noted above, there were up to five different fiber orientations used in this study.
The fine woven cloth was only tested in one direction, that being the 0°/90° direction.
Since this cloth showed relatively weaker results in early impact testing, it was only used
in the permeability testing, and not in the low-velocity impact nor composite laminate
void content and porosity tests.

The fine woven fabric preforrns consistently had the highest permeability, nearly 7 x
10—8 m2. However, one of the coarse woven fabrics, when the fiber geometry was
changed to i45°, it, too, nearly matched that of the fine woven preform. The only woven
fabric that didn’t show a higher permeability than the unidirectional preforrns was that of
one of the coarse woven domed laminate preforms. The specimens with the lowest
permeability were those of unidirectional domed preforrns, which had one third the

permeability of the preforrns constructed of fine woven reinforcing fabric.

98

3.3.2 Coarse-Woven E-glass Fiber Orientation

The coarse woven cloth was tested in the 0°/90° direction, as well as the i45°
direction. The laminate preform fibers were cut in an off-axis direction by means of a
cutting board. The off-axis laminates were then layered on top of each other, ensuring
that the edges were perfectly aligned so that all the warp and fill fiber tows were in the
proper i45° orientation.

3.3.3 Unidirectional E-glass Fiber Orientation

The unidirectional E-glass was tested for permeability in five different fiber
orientations, 0°, 30°, 45°, 60° and 90°. The fiber orientation allowed the flow of resin to
be studied on the same cloth thickness and weight. As noted before, this particular cloth
did not have well defined tows and was somewhat pliable. Because of this, the fibers,
when pressurized under the vacuum bag, were tightly compacted and allowed no natural
resin channel, which sometimes caused resin starvation, leading to porosity issues in a
laminate.

The results in Fig. 3.1.3 show the permeability of various types of laminate preforms
and different fiber orientations. The unidirectional laminate preforms, in general, had the
lowest permeability of all the specimens. Of the unidirectional preforms, the laminates
that had the highest permeability were those that had a variation of fiber orientation,
specifically laminates [45/—45]3 and [0/45/90]s. It is believed that because these
preforms’ had fibers that crossed each other that they formed somewhat of a natural
channeling or air space between the plies that allowed the resin to flow through it faster
than if all the fibers were aligned together. This is evidenced by the fact that all the other

unidirectional laminate preforms that had all their plies in alignment infused slower than

99

those that had a different fiber orientation between each of the plies. It is interesting to
note that seven laminates with the lowest permeability all had preforms with constant
fiber orientation throughout the laminate.
3.4 Effects Due to Structural Geometry

This section will discuss the effect of different structural geometries on energy
absorption capability. It was already shown in Section 3.1, as depicted in Fig. 3.1.3 that
adding a spherical dome to a mold lowered a coarse woven laminate preform’s
permeability by nearly 1.5 times that of a flat laminate preform. Although not all fabrics
behaved this dramatically, all of the preforms with a dome inserted into the mold
geometry had a lower permeability. This information is especially important when
designing a mold for production, because extra time will be needed to impregnate the
laminate with the resin, or other parameters might need to be changed, such as resin with
lower viscosity or the addition of extra resin ports or distribution media to properly infuse
the preform. The laminate preforms constructed of [0]6 unidirectional fabric and a dome
inserted had a permeability of 15% less than the same material with a flat laminate
geometry. Similarly, the laminate preform constructed of [90]6 unidirectional fabric and

a dome inserted had a permeability of 39.4% less than the same material with a flat

geometry.

100

3.5 Laminate Void Content and Porosity

To find the density of the composite, Archimedes’ principle is used. Archimedes’
principle is that of “upthrust” or buoyancy. The “principle” basically states that the mass
of fluid that a submerged object displaces is the same amount of buoyancy force applied
to the submerged object.
3.5.1 Experimental Set-Up

The specimens are first cut into small samples and labeled. The samples are then
dried in an oven to remove any entrapped moisture. A balance is then set up to weigh
samples in both air and water. The system uses a measurement frame that is set on the
balance. The bottom of the frame is attached directly to the force transducer on the scale.
The frame supports the specimen dish/basket that allow for dry and wet weighing of the
samples. A bridge is inserted through the frame, but does not come in contact with the
frame or the balance’s force transducer. The bridge’s only purpose is to hold the water
dish for weighing the sample wet. The water dish is filled about 3 cm deep with purified
water and its temperature is measured for water density purposes. The specimen
dish/basket has a dish on the top and is connected, by wire, to a basket that sets into and
is covered by the water in the water beaker. Once the water beaker is in place, the bottom
of the specimen basket is placed into the water and the dish is then secured on the top of
the frame. Once the balance is tared, specimens are ready to be weighed. The balance
used in this research was a Sartorious BP 2218 balance, with the ability to measure in 0.1
milligram increments. A specimen is first weighed in air and the mass is recorded. It is
then set into the water on the basket and weighed. This procedure is repeated for all the

composite specimens.

101

The specimens are then taken to a fume hood and a propane torch is used to burn out
the matrix in the composite sample. Once the entire matrix is burned out of the
specimens, they are allowed to cool and then weighed. The fiber-only specimens are
weighed in air and water and the masses are recorded, as done before with the composite
samples. The water temperature is measured to the nearest tenth of a degree to determine
its density, which is taken off from a water density table. From these four mass
measurements (composite’s wet and dry mass and the fiber-only wet and dry mass) the
volumes and densities can be calculated to determine void content. All calculated and
retrieved data from manufacturers are noted at the end of this section. This process is

shown below in Fig. 3.5.1.

Weigh Composite wet and dry Bum-out Matrix

Density Balance

    
 

 

 

 

 

 

C C
m — m
c _ dry we!
V — water
p
6
pc _ mdry
VC
Calculate Volume. Weigh Fibers
Density dry

Figure 3.5.1 Experimental set-up to measure wet and dry samples to determine density
of materials

102

The following notation will be used in this section:

V c = Composite Volume (cm3) p’" = Matrix density [? 2)
cm

Vf = Volume of fibers (cm3)
mjry = Composite mass weighed dry (g)

V ’" = Volume of matrix (cm3)
me = Composite mass weighed wet (g)

wet

V” = Volume of voids (cm3)

G
h)

min. = Fibers mass weighed dry (g)
p p = Composite density [% m )

f =

wet

m Fibers mass weighed wet (g)

f = . . g
p Mfg Fiber densrty {/szj

mc = Composite mass (g)

PM” = Water density (%m2) mm = Matrix mass (g)

mf = Fiber mass (g)

3.5.2 Determining Composite Density and Volume

Archimedes’ principle of buoyant objects is then used to determine the density and
the volume of the specimens. The volume of the composite can be calculated knowing
the relationship of water density and mass as weighed in both water and in air. The

following equations are used to determine the volume and density of the composite:

 

m‘ , —mc_,
V‘ =——""—— (3.5.1)
,0
mg,
C = "’ 3.5.2

3.5.3 Determining Fiber Volume
To obtain a “fiber only” specimen, the matrix must first be removed out of the

composite. This can be a tricky operation. because a bum-out test must be conducted. A

103

bum-out test incinerates the entire matrix in a composite. Since the fibers are not
flammable, only the matrix is incinerated and only the fibers remain. Precautions must
be taken because fumes and gases from burned resin can be toxic and should not be
inhaled, proper ventilation must be used when conducting this test. After the bum-out test
is conducted on the composite sample, the remaining ashes and residue must be removed
by rinsing and then drying, to remove any residual moisture. This is done so that the
ashes, residue and rinsing water do not alter the fiber mass. Similarly, the density and
fiber volume can be calculated also by using the weighed mass and the known density of

the fabric. The volume can be calculated using Equations (3.5.3).

In .
V1: d” (3.5.3)

 

3.5.4 Composite Volume

Since the mass of the composite and fibers are known from above calculations, the
matrix’s mass in each specimen can be calculated with Equation (3.5.4):

m'" = mc — mf (3.5.4)

Knowing the calculated mass of the matrix, the volume of the matrix can be
calculated using a resin density obtained by the manufacturer, or by using a sample of
neat matrix and calculating it using Equations (3.5.1) and (3.5.2). Neat matrix is cured
sample of epoxy containing no voids or reinforcing fabric. The volume of the matrix in

the composite can be calculated with Equation (3.5.5):

Vm

— (3.5.5)

104

3.5.5 Calculating Void Volume

Knowing information found by Equations (3.5.1), (3.5 .3) and (3.5.5), the
composite’s void volume can be calculated

V" = V6 - V’" — Vf (3.5.6)
3.5.6 Calculating Void Percent - Porosity

Knowing the volume of the composite found in Equations (3.5 .1) and the volume of

the voids in Equation (3.5 .6), the composite’s void content can be calculated:

V

 

%M%=V-m0 can

C

3.5.7 Data Analysis on Laminate Porosity

Void content of a laminate is directly correlated to the quality of the VARTM
manufacturing process. The sealed vacuum bag and resin piping systems are susceptible
to air leaks, if all air leaks are not fully removed, porosity will find its way into the
laminate. Normally, if there is an air leak in the vacuum outlet system, including the
vacuum piping or vacuum reservoir/resin trap, the vacuum pump can overcome the leaks.
As long as a vacuum is maintained on the system until the resin gels, leaks downstream
of the laminate air cannot migrate into the laminate. However, any leaks in any degree,
upstream of the laminate will normally result in porosity, and if great enough, can ruin a
composite. Because of this, steps noted in Chapter 2 must be taken to ensure all vacuum
leaks are detected and eliminated. Although some air bubbles are created in the resin
during mixing, they are normally removed during the vacuum process or migrate to and
get trapped in the distribution media. A laminate with zero porosity will have no air

voids and is considered a perfect laminate. Many manufacturers tolerate up to 1% void

105

content in certain laminates; less than 1% void content for critical composite parts, more
for non-critical composite parts.
3.5.8 Results of Laminate Porosity

Three laminate specimens were tested for laminate porosity. The first specimen lot
was a coarse woven composite with visual indications of porosity, labeled Coarse
Woven. The second specimen lot was a unidirectional composite, labeled Unidirectional,
that was fairly clear but contained a few visual entrapped air bubbles and some lines of
un-saturated fibers. The third specimen lot was a fine woven composite with fewer visual
indications of porosity than the first, labeled Fine Woven. A picture of these specimens
is indicated in Figure 3.5.2 before the bum-out test. The after bum-out specimens are
shown in Figure 3.5.3. Voids are notably visible in the Coarse Woven sample, whereas,
Fine Woven and Unidirectional samples are fairly clear. Although not seen in the picture
of the Unidirectional sample, there were some fine lines of narrow visual fiber bundles
that were not wet out, and were only detected when the sample was seen in the right

reflection of light.

106

 

Figure 3.5.2 Pre-bum out specimens
Although these samples are fairly dark, in actuality they are white to silvery-gray in

color, due to some residual ashes unable to be rinsed out.

Coarse Woven Unidirectional Fine Woven

      

Figure 3.5.3 Post-bum out specimens

Figure 3.5.4, below, graphically compares the calculated porosity of the three
different samples. The Coarse Woven material had over 3.5 times the porosity than the

Fine Woven samples. The Unidirectional had 1.8 times the porosity than the Coarse

107

Woven and 6.5 times that of the Fine Woven. VARTM was designed to minimize
porosity in a laminate; the results shown here show that different laminate preforms
contain various levels of porosity. It is interesting to note that visually, the Coarse
Woven sample looks like it has the most porosity; however, the Unidirectional sample
calculated the highest. This is believed to be due to the fibers that were unsaturated and
visible only in certain lighting. Obtaining porosity in a laminate emphasizes the
importance of leak checking after VARTM is set-up. Once infused, it is virtually
impossible to remove air bubbles entrapped in the liquid resin. If porosity is too severe,
the laminate is surely bound for the scrap bin. Many times visual inspection of a
laminate can determine if its porosity is too severe and should be discarded. If a
laminate’s visual porosity is teetering on the fence of tolerance, then the above

calculations can be completed to determine if it meets the user’s tolerances.

Void Percent of Laminate by Preform Type

 

0.8

.0
ox

Void Percent %
O
a

.0
N

 

 

 

0.0 i

Coarse Woven Fine Woven Unidirectional

Figure 3.5.4 Calculated percentage of laminate porosity

108

Table 3.5.1 below contains the calculated composite density; known fiber density from

manufacturer, calculated and manufacturer’s supplied density of matrix and the

calculated void volume and percent.

Table 3.5.1 Composite characteristics

 

 

 

 

 

 

 

 

Composite Fiber Matrix Void
Density- Density- Density-
Caic Density-Mfg Calc Mfg Volume Percent

Laminate g/cc g/cc g/cc J/CC cc %
Coarse

Woven 1.93 2.54 1.168 1.108 0.0068 0.68
Fine Woven 1.80 2.54 1.168 1.108 0.0019 0.13
Unidirectional 1.86 2.54 1.168 1.108 0.0125 1.06

 

 

It is important to note that there are many factors that can skew results from this
porosity test. Since the porosity and void content values are normally very small, it lends
much for error. Some error can be attributed to the sensitivity of the balance. The higher
the sensitivity, the more accurate the calculations will become. One inaccurate value can
escalate into large calculated error. Another big error factor is during the bum-out test
used in this research. Although it is a fairly crude process, it is limited to the heat of the
propane torch and the operator’s ability to situate the specimen to burn out and remove
the matrix from the entire composite, especially within the specimen’s center layers. In
some cases, there could be ashes or residue trapped between the multiple plies of the
composite or all of the matrix material may not be totally burned out of the specimen. It
is very important that heat be continuously applied to drive the entire matrix out of the
specimen. The rinse cycle may also have a similar effect, due to it not removing the
remaining ashes or residue. Although not available during this experiment, there are
commercially-available crucibles that isolate the specimen in a clean chamber and burn a
specimen up to 600° F. These systems also remove all exhaust fumes out of the chamber

so they are not trapped in the fibers of the laminate. The propane torch method was

109

effective and, as proved in this study, produced results that were in line with visual
examination.
3.6 Testing Conclusions

The methods utilized above helped characterize various aspects of VARTM and the
different reinforcing fibers and matrix used in this research. As noted above,
permeability and porosity are keys to understanding composites and determining how
composites react to VARTM and infusion of a resin matrix. Since void content in any
composite is crucial to the performance of the composite, it is imperative that these
destructive tests be used to help hone the VARTM manufacturing process, creating a
nearly void-free laminate. When the somewhat transparent fiberglass fabric is used,
porosity can sometimes be seen with the naked eye, as shown in Fig. 3.5.]. However,
when more advanced fabrics, like carbon, graphite or aramid fabric (i.e. DuPont’s
KEVLAR), the porosity cannot be seen due to the opaqueness of these materials. It is
therefore, imperative that the VARTM manufacturing process be practiced and honed to
avoid having excess porosity in the laminates. The engineer must be comfortable in
taking the steps necessary to avoid air intrusion in the vacuum system and detecting and
eliminating air leaks. The VARTM process was designed to reduce porosity and
composites were designed to replace traditional manufacturing materials by minimizing
weight, maximize strength and maximizing energy absorption. The testing methods
noted above need to be used in any research program to properly produce advanced

composite characteristics in today’s complex structures.

llO

4. COMPOSITE MATERIAL IMPACT TESTING AND RESULTS

This chapter is focused on determining and comparing the energy absorption of flat,
cylindrical arched and domed composite laminates based on low-velocity impact tests.
The results of this research should help in designing prototypes of various armors. The
low-velocity impact operating procedures, laminate testing and data analysis sections
were all developed procedures utilized in the MSU composite laboratory, therefore, the
general guidelines in Sections 4.2 and 4.3 were modified from Schulz [27] regarding
some editing and necessary changes.
4.1 Determining Energy Absorption

All specimens were tested using a low-velocity instrumented drop-weight impact
system from Instron, namely the DYNATUP 8250. The impact results produced the
impact velocity and the load history. The histories of impact load (N), deflection (mm),
velocity (m/s), and absorbed energy (kJ) were obtained subsequently with the use of a
computer program based on Newton’s second law and mathematical integration.
Calculation was also done to determine the impact energy so that a comparison could be
made with the absorbed energy. Each test was run under the same conditions and setup to
eliminate additional variables beyond adjusting the impact energy. The following
sections provide details on the equipment and operating procedures.
4.1.1 Low-velocity drop-weight impact test

A picture for discussion purposes of the low-velocity impact test setup can be seen
below in Figure 4.1.1. There are several important features to note. First, there are safety
doors on all sides of the unit. These safety doors are monitored by the control box, the

crosshead will not move with any of these doors open. The yellow block in the middle is

111

the crosshead, which has a load cell tup and two small flags attached (although not
visible). The crosshead is attached to the release/pick-up actuator. Its purpose is to
release the crosshead by means of pneumatically controlled jaws. When a test is
completed, it is also dropped down to pick up and raise the crosshead for another run.
The tup tip is screwed into a load cell that is located behind the yellow crosshead, it has a
22241N (50001br) capacity and a 12.7mm (0.5”) hardened steel hemispherical tip, called
the tup tip, for impacting the specimen. Assumed to be perfectly rigid, the load cell
measures the load during impact. The two flags run through the infrared velocity detector
right before impact to record the impact velocity at the moment of contact between the
specimen and tup tip. The velocity obtained by dividing the distance between the flags
with the time it takes the flags to run through the detector. The control box allows
adjustment of the height of the crosshead on the guide rails. A pneumatic switch is
pressed to release the crosshead from the rail clamp.

The specimen is clamped at the base of the equipment such that the tup tip impacts
the center of the specimen. If the impact energy is low enough, the crosshead/tup will
rebound several times, further damaging the specimen. To prevent this, a rebounding
system is in place. If a rebound is detected, the rebound actuators will be deployed and
bottom edge of the crosshead will land on the pads of the rebound system before hitting

the specimen again.

112

Safety

  
 

Doors
Crosshead
Guide release/ pick-
Rails up actuator
Crosshead
Tup Tip
Reboun .
cylinder Control
Box
Velocity
detector
Emergency
Specimen Stop
Support

Figure 4.1.1 Instron’s Dynatup Low-Velocity
Impact Tester
To control the amount of impact energy, either deadweight can be added to the
crosshead or the crosshead height can be changed. This allows duplication of the tests.
The crosshead height is determined by measuring the distance from the tup tip to the
impact location on the specimen. The total impact weight (crosshead, tup, flags and

deadweight) is recorded so that accurate impact energies can be calculated.

113

4.1.2 Impact data
When impact takes place, the load cell records the tup load, F (I). To find the

acceleration, Equation (4.1.1) is used, where the tup load is divided by the total impact

mass, m. The data is recorded every 25 us.

40:5"? (4.1.1)

From the acceleration calculation in Equation (4.1.1), the velocity of the tup can be
determined. Equation (4.1.2) is the numerical integration of the acceleration over time.
Since the tup is decelerating during the impact, the integration is multiplied by —1. The

initial velocity, vr, is determined by the infrared detector and is added to this integration.

v(t) = — [a(t)dz + v, (4.1.2)
0
Equation (4.1.3) shows the final calculation to determine the deflection of the

specimen during impact. The velocity is integrated over time from zero to the final time

of the impact.

5(1): My: (4.1.3)

The data acquisition program also calculated the absorbed energy. However, due to
important subtleties in calculating the energy absorbed by the specimen, the calculations
will be covered later in this chapter on data analysis.

4.1.3 Operating Procedure
4.1.3.1 Pro-impact test adjustments
Before running the first test, several adjustments were made to the impact testing

machine. Initially, the specimens were prepared by marking their centers. A specimen

114

was then fixed into the clamping system and centered. The crosshead and tup tip were
lowered by means of pneumatic controls on the control box until the tup tip just hung
over the specimen; the tup tip and specimen were then aligned. While the tup tip was in
this position, the infrared sensor was adjusted up or down such that the second leading
edge of the bottom flag was about 3.2mm (0.125”) beyond the centerline of the plastic
insert in the velocity detector block. This adjustment assured that the velocity at impact
was recorded properly. The crosshead, load cell and tup tip were weighed and mass
noted for future data input.
4.1.3.2 Impact test procedure

Once the pre-impact adjustments were done, testing could begin. The specimen was
centered on the clamping fixture and clamped and the crosshead was adjusted to the first
height. The computer was set to retrieve the data from the load cell and infrared sensor,
once set, the user had 30 seconds to release the crosshead, load cell and tup tip system.
The protection cover was closed and the crosshead release button pressed to release the
crosshead, allowing gravity to accelerate it toward the specimen.
4.1.3.3 Rebounding and perforation

For convenience, the energy change was due to the height change of the crosshead.
The crosshead was adjusted to the maximum height to determine if perforation was
possible at this maximum height. If perforation was reached, the height was decreased
until rebounding occurred for following specimens.
4.1.3.4 Data Acquisition

The voltage signals from the load cell and infrared sensor are then sent to a computer

data acquisition unit. The computer obtains the load and impact velocity. The results are

115

obtained at a rate of 25ps up to 100ms. The computer outputs the load, deflection,
velocity, and absorbed energy for each time step. This data is converted and sent to a
usable file for evaluation by means of an ExcelTM spreadsheet.

Once the load and deflection were obtained, the impact energy could be calculated
by integrating the load-deflection curve. The area under the curve, as calculated by
integration, was the amount of energy absorbed by the specimen. The absorbed energy
could then be compared to that of the impact energy. The performance of a laminate
could be compared as the impact energies were increased. The more energy the
specimen could absorb, the better its performance was. The specimens could then be
compared to each other and the best laminate could be used for further study.

4.2 Data Analysis

The dropped loads from the crosshead system resembled impact forces caused by a
projectile. Based on the voltage and the speed of the tup tip on impact, data was
simultaneously converted into force, acceleration, velocity, displacement, and energy
histories.

Composite armor and structures must absorb an overabundance of loads, from jolting
impact to quasi-static stresses. They must be able to absorb the associated force, be it
impact force from a bullet, shrapnel, shock-wave blast or numerous other forces
experienced by soldiers or law enforcement agents. The structure must also be able to
absorb the energy so the force is not transferred to personnel. The load-deflection
relation and the energy profile are important aspects of this research to better understand

how energy is absorbed by composite structures.

116

4.2.1 Load-Deflection Relation

The fundamental data from the impact experiments performed in this thesis research
was impact force recorded in voltage observed by the load transducer on the Dynatup
8250. As noted before, the voltage was then converted to time, load, energy, velocity and
deflection. From this information, the load and deflection were used as the basis of the
energy study. The load-deflection relation is the most straightforward method to describe
behavior of composites during impact. A load-deflection relation is obtained by plotting
the force of the load against the corresponding displacement throughout the impact. This
relationship provides most of the data for impact analysis. By evaluating the curve at
various points and noting the slope of the curve, the load-deflection history can also give
insight to how a composite damages. The best attribute of the load-deflection curve is
that it visualizes how the composite absorbs the impact energy throughout the whole
impact process.

There are two basic types of load-deflection curves created by the Dynatup software.
The curves are based on whether or not the tup tip penetrates through the specimen or
rebounds off from it. Figure 4.2.1 shows these two types of curves, the closed curve
represents a rebounded specimen and the open curve represents a penetrated specimen.
For the closed or rebounded curve, note how the load increases to a peak load and then
loops back to the beginning, such that the load decreases as the deflection also decreases.
The looping back of the curve is caused by tup rebounding upwards, which makes the
load decrease as the specimen deflects back away from the laminate surface in a negative

direction, creating the looped curve.

117

Load Deflection Curves

 

   

_ Penetrated
, Laminate

  
      
 

Rebounded
" Laminate

 
 
 
 

 

 

 

0 5 10 15 20 25 30
Deflection mm

Figure 4.2.] Types of curves for penetrated laminate and
rebounded laminate

Penetration, on the other hand, takes place as the tip embeds itself into the laminate.
As soon as penetration is reached, there is no rebounding of crosshead and tup, which
results in an open curve. Once the tup tip penetrates through the specimen’s surface, it is
defined to be perforation. When perforation is obtained, the tip tup slows down due to
friction between the tup tip and fibers and matrix in the specimen. Once the specimen has
been perforated, the small load caused by the fiiction is not considered in the energy
absorption calculation.

The load-deflection curves for different impact energies for a particular design are
normally plotted on one chart. Figure 4.2.2 shows such a chart for three-ply, coarse
woven panels, with different impact energies applied. The energies were applied
differently by increasing the initial distance between the tup tip and the top of the
specimens’ domes. The maximum deflections range from 38 to 55 mm. There are four

specimens that rebound the tup tip, and three that succumb to penetration. The maximum

118

load is about 6 kN, and the maximum deflections is just over 55 mm.

Load Deflection curve for Domed Laminate 3-Ply Coarse Woven

 

 

 

Load kN

 

 

 

 

 

 

 

0 10 20 3O 40 50 60
Deflection mm

Figure 4.2.2 Load-Deflection curves for domed laminates

In comparison, the load-deflection relation for flat specimens is shown in Fig. 4.2.3.
The general shape of the curves is much different than that made by the domed laminates.
The peaks, created by the maximum loading, appear much sooner than those in the
domed specimens. This is because the flat specimens do not allow the tip to have any
“extra” travel. The domed composites, on the other hand, are struck by the tup tip and
the dome is inverted, allowing the tup tip to travel farther before reaching maximum load.
The geometrically domed shape is able. to “catch” the top tip, whereas, the flat specimens
abruptly take the load. It can be seen that the maximum load for a flat panel is also less

than that of the domed laminate. The maximum load is only a little over 4 kN.

119

Load Deflection curve for Flat Laminate 3-Ply Coarse Woven

 

 

 

 

 

 

 

4
E 3
1:
“3
o
A 2 ~

I

0 .

Deflection mm
Figure 4.2.3 Load-Deflection curves for flat, penetrated laminates
4.2.2 Impact Stiffness

The stiffness can be divided into two parts for a flat panel. The initial stiffness is
determined by obtaining the load-deflection curve’s slope early in the impact, as seen
below in Figure 4.2.4. There is a bump in the load in both plots. This small bump is not
included in the determination of the stiffness (slope). Notice how the slope of the initial
stiffness changes for the flat panels into the second stiffness at a deflection around 5.0
mm in Figure 4.2.4. The critical point in this stiffness change is believed to be the onset
of delamination. This stiffness change is not as apparent for the arched or domed
composites. Also, depending on the reinforcing fibers and the mode by which the

laminates fail, the second stiffness may not be as apparent even on some flat panels.

120

 

Load vs Deflection Flat Laminates 3 Plies Coarse Woven E-glaes

 

 

 

 

 

4.5-
; Maximum Load
4.01 /
rSecond _ _____ _ _ _L A s
3'51 Stiffness
3.0; “—v —*- — , — e‘ —
52.5713. —-— — J; —— *v—
. ”mp
w 20* 1 Maximum
8 i I / / \ Deflection as per
._1 15 Extension Method

 

 

 

  

Initial Stiff—ness_ '

    

 

o 5 10 15 25
Deflection mm

Figure 4.2.4 Parts of a Flat Composite Load-Deflection Curve

4.2.3 Peak Load

The peak or maximum impact load changes based on specimen curvature and

 

boundary conditions. A flat panel produces a single peak load, the load increases sharply

and then drops sharply with a relatively small deflection. An arched or domed composite,

however, produces different peak load shapes, depending on a penetrated or rebounded

tup tip, as shown below in Fig. 4.2.5. This is due to the dome or arch shape deforming as

it absorbs the load. As the shape deforms, it causes various changes in the load, as seen

on the load-deflection curves, most likely due to failure, delamination and buckling of the

composite. However, oncethe impact inverts the shape and the fibers begin taking the

most strain, a higher peak is noted, which is similar to that seen on a flat specimen’s

curve.

12]

Load Deflection Curve Arched or Domed Laminate

Load after shape

 

inverted from impact

 

4i

Load when shape

deforming and |

 

 

 

 

 

 

 

 

Initial . , ..,. _ __
bucklin under irn act
2 Load 9 P Ii
8 3“: l A 1 1
"i . K N , l,
— F WK “"1 [I Penetrated
“ ' . . .1. M“ A!“ ' Laminate
°~-— . , Rebounded/“F f " .
0 ‘° 2° Laminate ' so

Deflection mm

Figure 4.2.5 Parts of an Arched or Domed Composite Load-Deflection Curve

4.3 Energy Absorption

4.3.] Integration Method

The energy absorbed by the composite during impact is calculated via Equations

(4.3.1) and (4.3.2). It is simply the determination of the area bounded by the load

deflection curves. The load, f(8), defined in Equation (4.3.1) is integrated over the

deflection, 5. The upper limit, 8 ,, is taken as the final deflection for closed curves. For the

Open curves, the limit 5, is determined by the extension method, which is explained in the

next section.

F =f(i5)

6

E. = from

(4.3.1)

(4.3.2)

122

4.3.2 Extension Method and Integration

Determining the area for integration on the open load-deflection curves is critical for
determining accurate energy absorption. Figures 4.3.1 is for a unidirectional E-glass flat
laminate. It shows a plot with an open curve where perforation takes place. A line is
extended to the abscissa at the same slope as the descent of the load during the
penetration process. This line is the “extension” of the load deflection curve to eliminate
the effects of the friction due to the rubbing of the tup with the specimen after
perforation, whereby giving this method its name. The location where the extension
intersects the abscissa is the upper bound, 5 ,, in Equation (4.3.2). The integral of the load
deflection curve, using the new extension method boundary, calculates the area under the
curve, which is the absorbed energy noted from Equation (4.3.2). This is shown

pictorially in Fig. 4.3.1.

Load vs Deflection Flat Laminates 3 Piles Coarse Woven E-giass

 

 

 

 

 

 

 

§
‘\§ / Absorbed Energy

Maximum
' Deflection as per
1.5i , extension method

_ : *\‘
£1 .

0 5 10 15 20

Load kN

 

Deflection mm

Figure 4.3.1 Area calculated under the load-deflection curve is the absorbed energy

123

4.3.3 Energy Profile

The equations for determining the impact energy are given below in Equation (4.3.3)
and Equation (4.3.4). The impact velocity is determined by two factors. The first is the
energy due to kinetic energy, which is the first term of Equation (4.3.3). The variable m is
the mass of the crosshead/tup. The initial velocity, v,-, is determined by Equation (4.3.4),
which is also the impact velocity measured by the infrared sensor/emitter. The second
component of the impact energy is the potential energy generated by the deflection of the
specimen during impact. The additional variables of Equation (4.3.3) are g, the
acceleration of gravity and h ’, the maximum deflection of the specimen. The maximum
deflection is determined by finding the deflection where the extension line intersects the

abscissa for open curves. For closed curves, it is the maximum deflection observed on the

load-deflection curve of the specimen.

I

E. = émv,2 + mgh'= mgh + mgh' (4.3.3)

v, = 2gh (4.3.4)

The energy profile is the key to characterizing the energy absorption of the
composite. The energy profile shown below in Figure 4.3.2 is for a series of flat
composite specimens fabricated with 3 plies of coarse woven fabric. The impact energy
(E) is plotted on the abscissa and the absorbed energy (E) on the ordinate. The scales for
both axes are intentionally the same such that a line can be drawn at a 45° angle, which is
the equal energy line. Any data point that lies on this line means for that given impact
energy the specimen absorbed all of that energy. At the upper end of the energy profile,
the absorbed energy is very close to the impact energy. Once perforation or complete

breakdown of the specimen is reached, the specimen has absorbed the maximum amount

124

of energy. As a result, the data points move away from the equal energy line for
increasing impact energies. In this particular case, the perforation energy or the maximum

absorbed energy is 58 J for the flat specimens.

Fiat Composites
3 Plies Coarse Woven E-Glass
Absorbed vs Impact Energy

 

120

Equal
Energy
Line

 

 

100

Maximum
\ Energy Line

 

 

 

 

 

Absorbed Energy J
8
\
\ .

. \ Absorbed

Energy
20 Results of ~
Flat
Composite

 

 

 

 

0 r . fl . .
0 20 40 60 80 100 120

Impact Energy J
Figure 4.3.2 Energy profile of flat composites

125

4.4 Flat Composites Experimental Results

The flat specimens that underwent low-velocity impact testing were constructed of three
plies of coarse woven E-glass. They were manufactured using VARTM and impregnated
with PRO-SET 117LV Epoxy Resin and 226 Hardener. The samples were cut to roughly
15 cm x 15 cm. The Dynatup 8250 low-velocity impact machine impacted the flat
specimens at six different heights. The results are noted in this section. Table 4.4.1

shows each specimen with its corresponding maximum values as used in this section.

Table 4.4.1 Flat Composite Experimental Results Summary

 

 

 

Height Absorbed Energy Max Deflection Max Load Tup Tip

mm J mm kN

250 30.66 14.12 5.02 Rebound
300 39.38 5.17 5.24 Rebound
350 49.82 17.75 5.14 Penetrate
400 54.22 15.10 5.21 Penetrate
450 55.1 1 17.80 5.75 Penetrate
500 57.97 15.80 5.46 Penetrate

 

 

 

 

 

 

4.4.1 Load Deflection

The load-deflection curves can be seen below in Figure 4.4.1. The specimens are
annotated by the initial height of the tup tip. The maximum load for the flat panels was
about 5.5 kN. The specimens with 250 and 300 mm initial height rebounded and the
remaining four experienced penetration. The maximum deflection for the rebounded
laminates was about 16 mm, whereas the maximum deflection for the penetrated
specimens, as found by the extension method was 18 mm. All of these laminates’ curves
agreed very well with each other, which indicated that the laminates were constructed

similarly.

126

 

Woven Coarse Load Deflection Flat Composite

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7
6 .
—250mm

5 ' —— —— —— - ———300 mm

4 1 350mm
5 I——-400 mm
'0 3 ——450mm
8 . W —500rnm
A 2 I" i '

‘fZiL—r *

0 a— T w "' i ‘05 .“ i .

0 10 20 30 40 50 60 7O 80

 

 

Deflection mm

 

Figure 4.4.1 Load-Deflection curves for flat composites

4.4.2 Energy Absorption

Once the load-deflection curves were created, they were then integrated to determine
the area under the curve, which would result in their corresponding absorbed energies.
The absorbed energies were then plotted against the impact energies applied by the
weight and initial height of the crosshead system. Figure 4.4.2 shows the impact energy
plot for all the flat specimens. It can be seen that the highest absorbed energy by the flat
specimens is about 58 J. This graph is an excellent example of how the rebounding
specimens absorb nearly all of the impact energy, as depicted by the specimens’ plots
being close to the equal energy line. This phenomenon continues to the penetrated
specimens, too. However, once they start deviating away from the line, it can be seen

that the maximum energy absorption capability of the laminate was reached.

127

Theoretically, this line would continue horizontally to the right even if the impact energy

was increased.

Flat Composites
3 Files Coarse Woven E-Glass
Absorbed vs Impact Energy

 

 

 

 

 

 

 

120
100 i~ ~~~ ~ "v 7* __. ~~~~~~
80
H
>.
9.?
g 60 ‘ , f
m
'D
0)
re 40‘"_ T _
O
m
.D
<
20
00 20 40 (so so 100 120

Impact Energy J

Figure 4.4.2 Absorbed energies of flat composites

4.4.2 Failure Results

Figure 4.4.3 depicts the typical damage the flat composites experienced. The
damage around the penetration was localized in all of the flat composites. Because the
tows of a woven fabric are intertwined with neighboring tows, the damage found in these
flat specimens are usually local and do not delaminate far from the penetration. Even the
specimens that rebound, the damage is not much larger than the area of tup tip. In a flat
laminate, the tows are straight, in a horizontal sense, as compared to that of the arched or
domed laminates. Because of this, when the tup tip impacts the fiber bundles, the tows
have nowhere to go except strain until failure. In a domed or arched laminate, there is

extra material due to the geometry, the plies and fibers have more of a chance to interact

128

with each other and take the impact differently. As seen in the next two sections, the

damage will be much different than that of this flat specimen.

    

Figure 4.43. Typical flat composrtedamage by penetrated
tup tip. Damage very localized around penetration.

4.5 Arched Composites Experimental Results

The arched specimens that underwent low-velocity impact testing were constructed
of three plies of coarse woven E-glass, too. They were manufactured using VARTM and
impregnated with PRO-SET 117LV Epoxy Resin and 226 Hardener. The specimens
were cut to roughly 10 cm wide x 15 cm long. The Dynatup 8250 low-velocity impact
machine impacted the arched specimens at seven different heights. The results are noted

in this section and summarized below in Table 4.5 .1.

129

Table 4.5.1 Arched Composite Experimental Results Summa

 

 

Absorbed Max Max
Height Energy Deflection Load Tup Tip

mm J mm kN

450 62.61 49.60 4.05 Rebound
550 75.02 51.50 4.32 Rebound
650 86.62 51.23 5.05 Rebound
700 88.45 45.00 3.48 Penetrate
725 94.39 40.80 3.01 Penetrate
745 104.00 51.98 5.84 Penetrate

 

 

 

 

 

 

 

4.5.1 Load Deflection

The load-deflection curves can be seen below in Figure 4.5.1. The specimens are
annotated by the initial height of the tup tip. The maximum load for the arched panels
was just nearly 6 kN. The specimens with 450, 550 and 650 mm initial height rebounded
and the remaining three experienced penetration. The maximum penetration for a
rebounded arched composite was about 52 mm, whereas, the penetrated arched
composite was about 54 mm when the extension method was used. Notice that there is
an initial peak right around 8.5 mm, this is due to the tup tip experiencing the initial
stiffness of the arched laminate. As the sample’s deflection increases, the load starts to
lessen from about 10 mm to 23 mm. At about 25 mm, the load starts to increase again.
This is due to the arch becoming inverted and starting to absorb and “catch” the tup tip.
The larger peak on the curve is when the arch becomes inverted and allows no more
travel downward. When this happens, the maximum load is experienced by the
specimens. It is interesting to note that all of the specimens, whether they rebound or
penetrated by the tup tip, follow the same general curve. This is a good indication of the
quality and consistency of the VARTM manufacturing process. If the samples were not

consistent, their load-deflection curves may not agree so well with each other.

130

4.5.2 Energy Absorption

 

Woven Coarse E-Glass Arched Composite

 

 

 

 

 

 

 

 

 

 

 

 

 

8
i
V 7

6

p —450 mm

5 “M“ '— *z’ ' ' e; - ----- -——550 mm
1 650 mm
1 E 4i _ — — — h— __ ..._._.700 mm
. '1: Ni
1 8 3 r“. ——725 mm
‘ ._1 .71 _745 mm.
1 2 “ fl I [Je'f‘i‘ : _— l..- .__._- _, ,1

A a ' i 1"“ I
1 'i‘m‘ at “a A’sr“‘w¢h" r ' »
0 10 20 3O 4O 50 60 70

 

Deflection mm

 

Figure 4.5.1 Load-Deflection curves for arched composites

Once the load-deflection curves were created, the curves were then integrated to
determine the area under the curve, which resulting in their corresponding absorbed
energies. As before, the absorbed energies were then plotted against the impact energies
applied by the weight and initial distance of the crosshead system. Figure 4.5.2 shows
the impact energy plot for all the arched specimens. It can be seen that the highest
absorbed energy by the arched specimens is about 104 J. The impact energy of this
sample was only 104.4 J, which shows that almost all, 99.6%, of the impacted energy was
absorbed by the arched specimen. This graph shows how almost all of the specimens
absorb nearly all of the impact energy, as depicted by the specimens being close to the

equal energy line.

131

Arched Composites
3 Plies Coarse Woven E.Glass
Absorbed vs Impact Energy

 

120

1004..—————— 7 4—7 77.. —~ .——~—.¥7 W J L. —— A_.— .—

 

\

 

Absorbed Energy J
a a
i
i
i
i
i
i

i
i
i
i
i
i
i

 

 

 

0 20 40 60 80 100 120
Impact Energy J

Figure 4.5.2 Absorbed energies of arched composites

4.5.3 Failure Results

Figure 4.5.3 shows an arched composite that was impacted and rebounded the tup
tip. The damage for this specimen was also typical for arches of similar construction.
The dark or shaded areas are all areas of delamination. Actually, there are more areas
that are delaminated than areas that maintained intact. A result that was typical for these
arched specimens was to have more delamination on one side or the other of the point of
impact. This is mainly due to the how the laminate buckles and deforms when the impact
of the tup tip is exerted on the specimen. The laminates in a curved composite, such as
this arched specimen, react much more differently than those in flat panels. The
delamination is correlated to the amount of energy the specimen is able to absorb. This

particular sample rebounded the impact of the tup tip.

132

Point of Impact

    
    

Delamination
Figure 4.5.3 Damaged arched composite with majority of
plies in arch delaminated, shown as the dark or shadowed
areas.

Figure 4.5.4 shows a laminate’s side View. This laminate experienced penetration

and had delamination to the point of creating a space about 1mm wide its entire width.

Again, this delamination favored one side, common among arched laminates.

Figure 4.5.4 Perforated arcecomposrte wrth major ’
delamination, approximately 1mm wide.

133

4.6 Domed Composites Experimental Results

The domed specimens that underwent low-velocity impact testing were constructed
of three plies of coarse woven E-glass, as well. They were manufactured using VARTM
and impregnated with PRO-SET 117LV Epoxy Resin and 226 Hardener. The specimens
were cut to roughly 15 cm x 15 cm. The Dynatup 8250 low-velocity impact machine
impacted the arched specimens at six different heights. The results are noted in this

section are summarized below in Table 4.6.1.

Table 4.6.1 Domed Composite Exgrimental Results Summary

 

 

Absorbed Max Max
Height Energy Deflection Load Tup Tip

mm .1 mm RN

710 82.29 45.34 3.25 Rebound
740 84.29 38.26 5.41 Penetrate
750 98.74 44.88 6.06 Rebound
750 101.33 56.78 2.79 Rebound
790 100.98 46.19 5.65 Penetrate
850 113.22 48.69 5.10 Penetrate

 

 

 

 

 

 

 

4.6.1 Load Deflection

The load-deflection curves for the domed laminates can be seen below in Figure
4.6.1. The specimens are indicated in the legend by the initial height of the tup tip. The
maximum load for the domed panels was just over 6 kN. The specimens with 740, 790
and 850 mm initial height rebounded and the remaining three experienced penetration.
The maximum penetration for a rebounded domed composite was about 55 mm, whereas,
the penetrated domed composite was about 48 mm when the extension method was used.
Notice that there is a bump around 22 mm, this is due to the tup tip experiencing the
initial stiffness of the domed laminate. As the sample’s deflection increased, the load

slightly decreased for most of the samples between 25 mm to 38 mm, with only one

134

exception. At about 40 mm, the samples’ loads started to increase again. This was due to
the arch becoming inverted and starting to absorb and “catch” the tup tip, where it
absorbed the most load. When this happened, the maximum load was experienced by

the specimens.

Woven Coarse E-Glass Domed Composite

 

 

—7501

—850i
790
--7‘IO
—740
—750

Load kN

 

 

 

 

 

 

 

 

 

0 10 20 30 40 50 60 70
Deflection mm

 

Figure 4.6.1 Load-Deflection curves for domed composites

4.6.2 Energy Absorption

Once the load-deflection curves were created, the curves were then integrated to
determine the area under the curve, which resulted in their corresponding absorbed
energies. As before, the absorbed energies were then plotted against the impact energies
applied by the weight and initial distance of the crosshead system. Figure 4.6.2 shows
the impact energy plot for all the domed specimens. It can be seen that the highest
absorbed energy by the domed specimens is about 113 J. The impact energy of this
sample was 114.5 J, which shows that almost all, 98.9%, of the impacted energy was

absorbed by the domed specimen. This graph shows how nearly all of the specimens

135

absorbed nearly all of the impact energy, as depicted by the specimens, being close to the
equal energy line. As with the results from the arched specimens, it can be seen that
domed specimens may have had more potential for additional energy absorption, too.
This was because the domed specimens’ curve did not create a smooth curve of absorbed

energy that deviated away from the equal energy line like the flat specimens noted above

 

 

 

 

 

 

 

in Fig. 4.4.2.
Domed Composites
3 Plies Coarse Woven E-Giass
Absorbed vs Impact Energy
120
100 ~~-~ ——-— —~— -A - ~ ——— TL-

he

>3 80
DD
l-n
8

LH 60 ' _

1::
d)
'8

o 40
m
.0
<1

20

0 . -
0 20 40 6O 80 100 120
Impact Energy I
Figure 4.6.2 Absorbed energies of Domed composites
4.6.3 Failure Results

Figure 4.6.3 shows a domed composite that was impacted and perforated. The
damage for this specimen was also typical for domes of similar construction. The dark or
shaded areas are delaminated laminate and are annotated inside a dashed white line. In
this specimen, it can be seen that the top part of the dome delaminated, and the base

mostly stayed intact. In addition, all domes, whether experiencing penetration or

136

rebound, experienced the dark area in the shape of a cross that followed the 0° and 90°
fiber tows. None of the specimens experienced this dark cross at any other angle. This
darker area was caused by severe buckling of the fibers, which contained ripped and torn
fibers, and consequently made them darker than the normal delamination when shown on
a light table. For this penetrated sample, there are more areas that were delaminated than

areas that remained laminated.

Areas of severe
buckling/torn
fibers

\
Delamination

 

Figure 4.6.3 Penetrated Domed lamnate showing areas of
delamination and severe fiber tear.

Figure 4.6.4, below, shows a domed specimen that rebounded the tup tip. This
specimen has much more delamination around the entire dome area than the penetrated
specimen. There is only one area of the laminate that remained intact, as noted in Fig.
4.6.4. However, similar to the penetrated sample above in Fig. 4.6.3, it has the dark cross

oriented inline with the fiber bundles caused by fiber tear when the dome buckled. The

137

 

samples that rebounded all sustained more damage in the dome area than those that were

penetrated by the tup tip. As noted above, the domes tended to catch the tup tip and

absorb the energy.

    
  
      
 
   
 
 
     

‘1 so?

av

 

. Only . Deiar‘pinationfi
. . , - he
laminated

. ,4?

..area left on ' ~.:
dome '- *

 

11??
-? ' Delaminatio
‘ 1 . line .

Figure 4.6.4 Rebounded Domed laminate showing small
area of intact lamination.

4.7 Impact Test Conclusions

Figure 4.7.1 shows the energy chart for all of the tested laminates. It can be seen that
when a laminate is configured in different shapes, such as a flat, arched and domed shape,
the amount of energy the laminates are able to absorb increases. This is believed to be
due to the different boundary conditions associated with each laminate. The flat
laminate’s fiber aligrunent is in one plane. When it is impacted, the fibers immediately
start to take a strain. The fibers can only strain and stretch so much and then there is
failure. When the flat laminate fails, it has a much smaller damaged area and it is much

localized to only a few unit cells away from the point of impact.

138

Domed, Arched and Flat Composites
3 Plies Coarse Woven E-Glass
Absorbed vs Impact Energy

120

 

i
i
i
i
i
i
i

eDomed .

8

 

I Arched [

s
i
i
i
i
i
i
i

Absorbed Energy J

 

i
i
i
|
i
i

 

 

 

o 2‘. 40 .‘o to 160 120
Impact Energy J
Figure 4.7.1 Absorbed energies of 3 ply coarse woven composites

The arches have a fiber configuration that incorporates “extra” fibers into the
structure. The length of the arc of the arch is longer than the chord that the arch spans.
The extra fibers allow the shape deflected when impacted and tend to “soften the blow,”
so to speak. This gives the arched laminates more energy absorption capability.

Boundary conditions have an effect on energy absorption. In this study, the flat
specimens were clamped between two steel platens. The Arched specimens were bolted
because they were narrower than the rest of the specimens. Since the shape only allowed
for attachment on two sides, the ends were bolted to minimize slippage. The domes
allowed to be clamped because the shape was only in the middle and flat on all four sides.
Clamping on four sides of the flat and domed laminates helped reduce slipping when
being impacted. The arched shape is only bounded on the two ends and the other two

sides are open. Since fiber alignment is a key in a composite’s strength, having the fibers

139

marry down into a common plane, as in a dome, the composite can work as one unit,
absorbing energy from the entire perimeter of the dome’s base. The arch, on the other
hand, all the energy is dumped into only the ends.

The dome is a shape that is bounded by the circumference of the floor of the dome,
or the circular section made by the base of the dome. The fibers are now bounded on all
sides of the impact shape. When the dome is impacted, the fibers cannot move one way
or the other, they must deform within the volume of the dome. In doing so, the results
are apparent in that the laminate buckles on the fiber orientation lines. The tows that
radiate to the base of the dome at the 0° and 90°, become the sacrificial areas of the dome
deformation and break and buckle. However, as seen above, the rest of the dome is not
safe, either. Unless the dome was perforated, nearly all of the dome’s surface area was
delaminated almost to the base of the dome. This indicated that the entire structure of the
dome took part in absorbing energy from the tup tip. This deformation is different than
that found on the arched laminates. The arched laminates buckled primarily on one side
of the tup tip, between the impact point and the base of the arch, whereas the other side
did not normally delaminate. The arched composites were penetrated at almost 92 J of
impact energy and the domed composites were penetrated at 105 J. The domed shape
took over a 12% increase of impact energy to be penetrated, and absorbed 10% more
energy.

It can be concluded that the shape of a laminate has much to do with its ability to
absorb energy from impact. This is very useful when designing personal or vehicular
armor because geometric shapes, specifically domes, can be used within the armor to

help protect its wearer from impact of bullets, shrapnel or blast waves.

140

 

5. CONCLUSIONS, RECOMMENDATIONS AND FUTURE STUDY
5.1 Conclusions

The main purpose of this research was to investigate the capabilities of Vacuum
Assisted Resin Transfer Molding (VARTM) for the manufacttuing of geometric
composite structures which could be tested for energy absorption and determine if the
increased benefits of geometric shaped would increase effectiveness of personal and
vehicular armors. VARTM was deeply studied to understand the manufacturing
process’s capabilities and provide a document for other to successfully manufacture
composite parts with this method. VARTM only requires a humble list of manufacturing
tools as compared to that of other composite manufacturing processes; however, there are
some unique materials that need to be applied carefully to avoid air entrapment or
porosity within the vacuum system. There are six basic steps used in the VARTM
process, each step requires attention to detail to minimize laminate porosity. The details
noted in this document should help anyone embarking into a composite manufacturing
task large or small.

Electrical-grade fiberglass preforms were infused with the VARTM method and
evaluated for their permeability. It was found that woven fabrics, both coarse and fine
plain weave, have a higher permeability than the unidirectional electrical-grade fiberglass
used in this study. The higher permeability was linked to the natural resin channels that a
woven cloth possesses as the fiber tow bundles weave their way around each other and
create spaces from their change in direction. It was also found that the unidirectional
fabric that was used did not have a defined tow bundle like that found in the woven

fabrics. Because of this, when laminate preforms’ fiber orientations were solely one

141

 

direction, the undefined tow bundles would nestle together so well and compact, that they
would create dry areas next to the plate glass mold. The highest permeability of the
woven laminate preforms was more than 25% higher than the highest unidirectional
laminate preform. In addition, the size of the unit cell in a laminate had an effect on the
initial infusion rate; however, overall permeability was nearly identical for coarse or
woven laminate preforms.

The porosity of infused laminates was then evaluated to determine the effectiveness
of the VARTM manufacturing process. Since most advanced composites are created
with opaque laminates, such as carbon, graphite or aramid fibers, internal porosity cannot
be visually examined. Therefore, it is important to apply proven techniques, such as
Archimedes’s principle of buoyancy to calculate the void content and porosity in a
laminate. This study found that porosity could be found and used in a manufacturing
setting to determine the laminates porosity to confirm the quality of infusion VARTM
provides.

Lastly, domed, arched and flat panels were tested with a low-velocity impact
machine. Load-deflection curves were created for the different specimens, which
determined the most load a specimen withstood and what its maximum deflection was.
This curve also provided valuable information when it was integrated; it provided the
amount of energy the specimen absorbed. The spherical domed specimens absorbed the
most energy out of all the shapes, whereas, the cylindrical arched domes followed next,
and then the flat geometry trailing in last. It is believed that the domed composites are
able to absorb the most energy out of these three geometries because of the boundary

condition and the shape of the impacted surface. The laminates with a shape molded into

142

its surface were able to deform with the impact and absorb the resultant energy. Flat
specimens of the same construction, on the other hand, are not able to deform like their
shaped counterparts, and therefore unable to withstand as much load nor as much impact
energy.
5.2 Recommendations
5.2.1 VARTM

As noted in the many pages of Chapter 2, there are many components and variables
that can be incorporated into a VARTM manufacturing process. Because of this, each
proposed laminate must be evaluated to ensure that efficiency of VARTM is maximized.
The following are some recommendations that should be incorporated for future VARTM
manufacturing sessions.
Unidirectional fabric — It is recommended that unidirectional fabric with more defined
tows be tested for permeability. Theoretically, unidirectional fabrics should have a
higher permeability due to long, uninterrupted resin channels. The permeability results
married with the impact results should allow for selection of best material for future
study.
Barometer — It is recommended that a barometer be obtained and utilized during the
VARTM process. Since permeability is related to atmospheric pressure, more accurate
results could be obtained by knowing the local atmospheric pressure.
Digital vacuum gauge — For the same reasons a barometer should be incorporated, a
digital vacuum gauge with sensitivity of a tenth of a kPa should be employed to monitor

the vacuum pressure more closely. Since the pressure difference is found in the

143

denominator of the permeability equation, Equation (3.1.2), it would allow more accurate
calculations.
Infusion mold — It is recommended that a female domed mold be used to infuse
composite parts. Sometimes a thicker laminate would cause bridging around the base of
the dome because the atmospheric pressure was not enough to draw it tightly to the
mold’s surface. Other manufacturing methods (RTM) are available that would help this,
but the simplicity of VARTM would be lost.
5.2.2 Porosity Calculations
Burn out test — The burning out of matrix from the composite with a propane torch could
have been more thorough with a crucible designed for this purpose. Not only are
crucibles safer for the operator, they also burn the composite more thoroughly and under
ideal combustion conditions. In addition, they can be arranged so that the fibers in the
laminate are not damaged during the burning process.
5.3 Future Study

This research has demonstrated that domed composites absorb more energy than
composites of similar construction, but with different geometries. This gives high
potential of domes with various sizes, shapes, and configurations. Personal and vehicular
body armor incorporating variations of domes in the laminates are currently being
studied. The natural tract that this study should go is to increase the tests from the current
low-velocity impact testing to high velocity test such as those prescribed by National
Institute of Justice. Passing of these tests would allow armors to be placed into

production for the protection of military and law enforcement agencies.

144

APPENDIX

145

 

 

Technical Data

117LV Resin/224 Hardener

Laminating Epoxy

This combination is intended specifically for resin infusion and closed mold processes. It
is not appropriate for open molding.

The 117LV/224 Epoxy system is formulated for laminating synthetic mmposite structures
using closed mold processes. The 117LV/224 mixture will provide a working time of ap-
proximately 20 minutes at 72° F. A typical laminate will be gelled in approximately 2.25
hours at 72° F.

MIXING

Combine the 117LV Resin with 224 Hardener following the ratios by weight or volume
shown in the table. Stir the mixture thoroughly and transfer to the feed containers con-
nected to the resin distribution system.

CURING

The 117LV/224 mixture maintains excellent working properties until gel takes place. The
mixture will temper and continue to cure over the next several days at room temperature,
and after two weeks will reach an acceptable degree of cure for most applications. Ele-
voted temperature post cure will increase the degree of cure and improve the mechanical
and thermal properties.

Vike recommend building and testing sample laminates using proposed materials and
manufacturing processes to confirm working and curing characteristics under antidpated
use conditions.

HANDLING CHARACTERISTICS (Notfiir srxrr'fimhbn plotters)

Property Mixed Resin/Hardener

Density .............................. 93 lb/gal

Viscosity (at? 72°F (ASTM D—2393—80) ...... 800cps

Mix Ratio (117LV Resin:224 Hardener) Target Acceptable Range
by weight ............................. 100133 100.342 to 100.29.]
by volume ............................ 100234 100.363 to 10030.8
Pot Life (ASTM 02427-71) 100g

@691: ................................ 29 min

@7213 ................ 21min

(£68051: ................................ 15 min

 

146

 

 

£02956me ms “635.8 on c. E: ”mos—m> 339:.
scoEwuuo...EE .52: .305 .5; badge Eu: 3?: mFoEUuam to...
.3523. mam—w up: .0 .350 2: E Eton—E 2:32

.AUmDV uoUoEEEmU wficgom 12:39.50 a mEm: US$52.40 :

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

med wmd Ed and mmd www-m— Em< ASE—Ln: p380: comma: pod
m2 5 $2 $4 e: 3 Se 80 .3 as 255::
mi m3 a: 22 m2 1&qu 3 we 8 320
m3 “3 9: EH m2 mid 2a? a; Cozy eaasmsfl 8565 as:
8+3? 8%»? Bias mama? mo+m8m Red 2.8.... $3 2332 .233:
.23: 83; gag 33: 52.2 cad Ema. is amcmsm Essa:
meanest. 3&me 9333 £58.... mafia? «8-0 Esme. as: mazes: $23
3 mm. we ad m6 $30.0 Ehm< g; conmmcflm @223.
58.: §.: 5.2 ES: 29.... $.90 E? can: inseam amnae
89.: $55 83: Ex: $2: mood Ea? fig 2a; 5.5238
um mm mm mm low Baud E? 5 82$ amass:
memocm— .E moan—ac: wau—omfi E moan—ac: mxuuSNx
+5 ma x t. +2 2 x E + E m x E + 2 2 x E ease 68m uofimz amok beacon EUR» .5

 

 

flsuwfim PSU

 

 

 

 

“28383 wmfifimom >45:

mmHHmmnMOmm A<U$>$L l \CAOAE UZE<Z=Z<A

 

 

147

 

 

Technical Data

117LV Resin/226 Hardener

Laminating Epoxy

This combination is intended specifically for resin infusion and closed mold processes.
It is not appropriate for open molding.

The 117LV/226 Epoxy system is formulated for laminating synthetic composite structures
using closed mold processes. The 117LVRZ6 mixture will provide a working time of ap-
proximately 60 minutes at 72° F. A typical laminate will be gelled in approximately 3—3.5
hours at 72° F.

MIXING

Combine PRO—SET 117LV Infusion Resin with PRO-SET 226 Hardener following the ra-
tios by weight or volume shown in the table. Stir the mixture thoroughly and transfer to
feed containers connected to the resin distribution system.

CURING

PRO-SET 117LV/226 mixtures maintain excellent working preperties until gel takes place
and the mixture cures to a B-stage, during which it may be brittle. Brittleness is more evi-
dent at lower cure temperatures. The epoxy mixture will temper over the next several
days at room temperature, and after two weeks will reach an acceptable degree of cure for
some applications. Elevated temperature post—cure will increase the degree of cure and
improve mechanical properties and high temperature performance.

We recommend building a sample pane] using proposed materials and procedures to con-
firm working and curing characteristics under your shop conditions.

HANDLING CHARACTERISTICS (Not for snufication warms)

Property Mixed Resin/Hardener

Density .............................. 9.1 lb/gal

Viscosity (2} 72°F (ASTM D-2393-80). . . . . . 350 cps

Mix Ratio (117LV Resin:226 Hardener) Target Acceptable Range
by weight ............................. 100:30 1002319 to 10027.1
by volume ............................ 100:35 10037.0 to 1001315
Pot Life (ASTM D-2427-71) 100g 500g

(23653: ................................ 85 min 55 min
Q2172°F ................................ 44 min 34 min
(238031: ................................ 25 min. 23. min

 

148

 

 

 

.ccmE 852:2 Sam :55: .5 >. {is :6... 2?: 22: Us? 72.
.cfluwxmuomm an 23:38 on 2 .5: $323129;
.§_¢.:Ea mud—w up: .0 $35 or: am 32.59: 2:!» .AUmDv potato—uu mficgvm .3 fiesta r. £52.. to: E: 290 . .

 

Comb ”:5 .EZEKEoH Scam .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we «.8 mad and 8c and Ema. 53.: .863: .2 E. .95
he a: a: a: as :E «as; fiestas 252:
ME 9: e: a: 5 :9... 2395 88. 5222.35
E 9: a: a: «.2 39¢ .53 E soured 5:
8.53m mo+m$a 8+Emm noises 358m and 2.5... as: 3.382 E35
§§ Roda 3.2 3.2 032 cad 251.. fie $53 3:55
8&8... £1me aim—am aims m 8+ww~ m 35 Em... is 3352 asap
3 on «a 3 390 E? St ccfimcoa was;
awe: l 85.8 83 «new amen 2e? 9an 35% £28.
mafia $6.3 «3.3 use“ 8a.: mood 3:3 as: Es; Ea 2 68
a a 3 E a Sad E? E 29$ mecca:

:3 x sea 22 x so: 5 x ea; 52 x as: $8: x
+ Emfi x PM + 2mm x .5. + .32 x PM + Ema x r:— ..mE~.—. 58% «.9232 amok krone:— 1.43mi..—

uBfio—«UW uuaU

 

 

$53933 cmqfimom >45:

mmimm—LOME 12:3 m>Im

“rum-Dunn—

 
 

 

 

149

 

 

Technical Data

117LV Resin/229 Hardener

Laminating Epoxy

This combination is intended specifically for resin infusion processes.
The 117LV/229 Epoxy system is formulated for laminating synthetic composite structures
using closed mold processes. The 117LV/229 mixture will provide a working time of ap

proximately 120 minutes at 72° F. A typical laminate will be gelled in approximately 5
hours at 72° F.

MDGNG

Combine PRO-SET 117LV Infusion Resin with PRO-SET 229 Hardener following the ra—
tios by weight or volume shown in the table. Stir the mixture thoroughly and transfer to
feed containers connected to the resin distribution system.

CURING

PRO-SET 117LV/229 mixtures maintain excellent working properties until gel takes place
and the mixture cures to a B-stage, during which it may be brittle. Brittleness is more evi-
dent at lower cure temperatures. The epoxy mixture will temper over the next several
days at room temperature, and after two weeks will reach an acceptable degree of cure fc
some applications. Elevated temperature post-cure will increase the degree of cure and
improve mechanical properties and high temperature performance.

We recommend building a sample panel using proposed materials and procedures to cor
firm working and curing characteristics under your shop conditions.

HANDLING CHARACTERISTICS (Not for spufimiim punmes)

Property Mixed Resin/Hardener

Density .............................. 9.3 lb/gal

Viscosity {it 72°F (ASTM D-2393-80) ...... 310 cps

Mix Ratio (117LV Resin:229 Hardener) Target Acceptable Range
by weight ............................. 100:3] 10033.2 to 10028.2
by volume ............................ 100:37 '1 00:39.4 to 10033.5
Pot Life (A STM 0-2427-71) 100g 300g

(o'72°F ............................... 110 min 68 min
(31803F ................................ 61 min 44 min
(if883F ................................ 40 min 34 min

 

 

150

 

 

.2: uEua.E..5.£ .58.. .8533. >xona .3: 393 296699.. .1;
dacfiEkfia ma 2.5.53 on o. a: $93.? .33»...

£35.53 $3» or: .5 .920 95 n. Eton“: .3.—a.) .Aumab Bro—5.81.4 ucicnum 15:23.25 a $53 _rE_EcB...D ..

Erma "CK. BEEKEmP 5.5M .

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

and | We as E Nfid 2.53.

m: E m3 «.2 E .
«B m: «b E I :5 238 5:12:23 ago
a: .2: SH ”2 3.3 2.5.x E :38ch .8:
mo+m§s 8:23 mimosa 8&8“. 8&3... 8nd 3:9. as: $382 Essa
ans 3%: among «2.2 Sud 2.5. as: eswcgm :2:qu
8&8”. 2+ MKS. 8+8? 8+m 3m 89o 5.9.. as: 33 82 mass
2 «all hm LIN WW3 25a 33 conawcoa 27.59
$3: I BE: Es ES tamed Ema. . a gem was:
use: ”MS 3.: Sn: I mood 52 as 32» some. 68
x. mm a l 5 mm 88d 2&2 o 829 $265:

:3 x new" 23 x 9...: in x so: 3.: x .5: 3.83 fl x .
+ :3. x E + 2.3 x E + 22 x .E + :5 x L». has; 58¢ gaggafl brie; .31.:

£3.23 35

 

 

Roamfiuwm mNQEmmm >1K:

mm Edy—Ow: 1_<U~ m>IL

“rum-Dunn

 

 

 

 

151

 

 

Technical Data

117LV Resin/237 Hardener

Laminating Epoxy

This combination is intended. specifically for resin infusion and closed mold processes. It i
not appropriate for open molding

The 117LV/237 Epoxy system is formulated for laminating synthetic composite structures
using closed mold processes. The 117LV/237 mixture will provide a working time of ap-
proximately 260 minutes at 72° F. A typical laminate will be gelled in approximately 7-8
hrs. at 72° F.

MIXING

Combine PRO-SET 117LV Infusion Resin with PRO-SET 237 Hardener following the ra-
tios by weight or volume shown in the table. Stir the mixture thoroughly and transfer to
feed containers connected to the resin distribution system.

CURING

PRO-SET 117LV/237 maintains excellent low viscosity characteristics providing a good
flow rate until gel takes place and the mixture cures to a brittle ”13" stage. Elevated tem-
perature post cure of 15°F to 180°F is required for mixture to reach final cure.

We recommend building sample panels using proposed materials and procedures to un-
derstand working and curing characteristics under your shop conditions.

HANDLING CHARACTERISTICS (Not fiir snu'fiadim purposes)

Property Mixed Resin/Hardener

Density .............................. 9.0 lb/gal

Viscosity (61, 72°F (ASINI D-2393-80) ...... 360 cps

Mix Ratio (117LV Resin:237 Hardener) Target Acceptable Range
by weight ...... _ ....................... 100:30 ICD:33.9 to 10027.1
by volume ............................ 100236 1002402 to 1002322
Pot Life (ASTM D-2427-71) 100g 500g

(«72°F ............................... 281 min 139 min

(a 80°F ............................... 190 min 124 min

(I!- 88°F ............................... 10:3 min 7'5 min

 

152

 

 

 

.CcoEoEEEE LEE ESE»; >205 =3: $5.5 0.5.0.09. 73.
03.55009 3 093$:ch an 2 .2: £03m.» 7.395.

dang—E. 3E...— 3: .0 ~85 or: n. watch? «20> spawn: uBQEEEU mcECch 3.02805 a uni: 03:53.00 . .

Sam“ ":5 EEEemEP~ Essa .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

  

 

 

 

 

 

 

  

 

 

  

 

 

 

 

 

 

 

 

8o 02 was lelwfl $.00 25‘ 39.20 363: as .5 use
80 R; s2 s2 1 we saw; 953? e £m§05
ems mg 9.3 an" :9.» 0.532050. sexism; $.30
§ $0 «2 a: ”$0 20.? E 538000 :8:
Bio? 91080 0:33 mo+m§m 8::va 8.0 is? as: $0282 .2320
30: 03: $5 $3: and .8; 2&0 20.9 $9 smfism .83:
gigs 808+ mo+mo~m 8+is $.90 254 _ 238: seems
F 3. ms 2 $.90 50.? £0 canawcoa 2:80
ES: 8:: 33: 3.2 $.90 2&2 .. z :85 225
$03 03.3 £2 $.12 $90 3:? as: 2m; assessed
mm mm mm mm 30.0 5? A0 229 emcee:

:3 x ”roe 23 x :2 :3 x .98" 3.: x mom: 2:0
+ .53 x E + £2 x .5. + 22 x b. + £2 x LL. .950. :58. 3.32 .89 03%: 7.5.0:

fisvwaum VSU

 

 

 

 

$000.83 anfimmM >45:

mEHMmmOyE A<U5>TE

 

 

153

 

 

Technical Data

1 17LV Resin/239 Hardener

Laminating Epoxy

This combination is intended specifically for resin infusion and closed mold processes. It i
not appropriate for open molding

The 117LV/239 Epoxy system is formulated for laminating synthetic composite structures
using closed mold processes. The 117LV,'239 mixture will provide a working time of ap-
proximately 420 minutes at 72° F. A typical laminate will be gelled in appron’mately 10— 12
hours at 72° F.

MDGNC

Combine PRO-SET 117LV Infusion Resin with PRO-SET 239 Hardener following the ra-
tios by weight or volume shown in the table. Stir the mixture thoroughly and transfer to
feed containers connected to the resin distribution system.

CURING

PRO-SET 117LV/239 maintains excellent low viscosity characteristics providing a good
flow rate until gel takes place and the mixture cures to a brittle ”B" stage. Elevated tem-
perature post cure of 125°F to 180°F is required for mixture to reach final cure.

We recommend building sample panels using proposed materials and procedures to un-
derstand working and curing characteristics under your shop conditions.

HANDLING CHARACTERISTICS (Not for simficahkm mam)

Property Mixed Resin/Hardener

Density .............................. 9.2 lb/gal

Viscosity (if 7291’ (ASTM D~2393-80) ...... 290 cps

Mix Ratio (117LV Resin:239 Hardener) Target Acceptable Range
by weight ............................. 100:31 10034.8 to 1002279
by volume ............................ 1(1):.35 100142.?» to 100:34.0
Pot Life (ASTM D-2427-71) 100g 500g

@T72°F ............................... 465 min 225 min
@801: ............................... 360 min 163 min
(9188?? ............................... 219 min 128 min

 

154

 

 

jays—3.0.53 “XE «505.3,. Aria“. .3: Eu? 325.9...th 7:.

.Efiaucfiunm we vgbmccu .3 2 no: $935.15..th

dung—5 awn.» 05 we .3.—.0 in u. @300qu 925) came SEEEUEU m: 205m .E.E.:...c_0 r. x01: Firetrafi . .

Cam.“ Feb 8 22090.40. Eccx .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

  
 

 

 

 

 

 

 

 

 

 

$6 nmd end 36 03-0 Em< 237$ HERO... gum z: 303
80 03 :3 8“ 1538; 953? H “.2825
"3 m3 amfi mg :05 Esugmmfimb 00:03th mwdnu
0.3 mm: 53 La 3.0.0 292 05 0020230 EmI
8+is 01%; mo+mmma 3&3 a 2&0 E? a 3.352 3550
3.0: mm

M32 mg: 85 “8.0 Ed 20.? $3 amass .8:qu
8+0”? 8+ was mo+m8a 853* 0080 a? a. 5:82 2550.
3 ma 3 3 £90 :02 $3 conawcca 2:59
amna moms 83 «8.3 30.0 .52 cgacfi
was: 3&3 some ~32 mood E? 2%: Eu; ccvmemscu
a 1F 3 8 sad 20$ 00 eogmw we: as:

aw x 32: £2 x use: :3 x ".03; E: x now: «:6
+ :5 x .E + 52 x ,5. + £3 x b. + :5 x as. .95... :33. was»: «as bras... .31.:

Bsuwnum ~30

 

 

 

 

$80050 @3580 >03

mmmeEOML A<UE>IL

 

 

155

REFERENCES

['1]

[3]

[4]

[5]

[7]

[8]

[9]

[10]

[11]

Daniel, I. M., Ishai, 0., Engineering Mechanics of Composite Materials, 2nd ed.,
Oxford Univ. Press, New York, 2006, Chaps. 1,2.

DeLuca, E., Prifii, J ., Betheney, W., Chou, S.C., “Ballistic impact damage of $2-
glass-reinforced plastic structural armor,” Composite Science and Technology,
Vol. 58, 1998, pp. 1453-1461.

de Rosset, W.S., “Patterned armor performance evaluation,” International Journal
of Impact Engineering, Vol. 31, 2005, pp. 1223-1234.

Baker, A. , Dutton, S., Kelly, D., ed., “Composite materials for aircraft structures,
2nd Edition,” Reston Virginia, American Institute of Aeronautics and
Astronautics, Inc., copyright 2004.

Ambur, D.R., Stames, H. Jr., “Effect of curvature on the impact damage
characteristics and residual strength of composite plates,” 39th
AIAA/ASME/ASCHE/AHS/ASC Structures, Structural Dynamics, and Materials
Conference, AIAA Paper No. 98-1881, 1998.

Chun, L., Lam, K.Y., “Dynamic analysis of clamped laminated curved panels,”
Composite Structures, Vol. 30, 1995, pp. 389-398.

Herszberg, I., Weller, T., “Impact damage resistance of buckled carbon/epoxy
panels,” Composite Structures, Vol. 73, 2006, pp. 130-137.

Kistler, L.S., “Experimental Investigation of the impact response of cylindrically
curved laminated composite panels,” American Institue of Aeronautics and
Astronautics, Inc., Vol. 4, 1994, pp. 2292-2297. 141

Kistler, L.S., “Low Velocity Impact on Curved Laminated Composite Panels,”
Ph.D. Dissertation, Aerospace Engineering Dept., The University of Michigan,
Ann Arbor, MI 1996.

Kistler, L.S., Waas, Anthony M., “On the response of curved laminated panels
subjected to transverse impact loads,” International Journal of Solids and
Structures, Vol 36, 1999, pp. 1311-1327.

Gupta NK, Easwara-Prasad GL, Quasi-static and dynamic axial compression of
glass/polyester composite hemi-spherical shells, International Journal of Impact
Engineering, 22 (1999), 757-744

156

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

Gupta NK, Velmurugan R, Experiments and Analysis of Collapse Behaviour of
Composite Domes under Axial Compression, Journal of Composite Materials, 36,
(2002), 08, 899-914

Cui Z, Moltschaniwskyj G, Bhattacharyya D, Bucking and large deformation
behaviour of composite domes compressed between rigid platens, Composite
Structures, 66 (2004) 591-599

Easwara Prasad GL, Gupta NK, An experimental study of deformation modes of
domes and large-angled frusta at different rates of compression, International
Journal of Impact Engineering, 32 (2005) 400-415

Rao KP, Ganapathy S, Failure analysis of laminated composite
cylindrical/spherical shell panels subjected to low-velocity impact, Computers &
Structures, 68, (1998), 627-641j

Her Shiuh-Chuan, Liang Yu-Cheng, The finite element analysis of composite
laminates and shell structures subjected to low velocity impact, Composite
Structures, 66 (2004) 277-285

Chandrashekharak, Schroeder T, Nonlinear impact analysis of laminated
cylindrical and doubly curved shells, Journal of Composites and Materials, 1995,
29, 2160-2179

Ranger Boats, 96 Ranger Road, Flippin, AK 72634,
http:wawgreatlakesbass.com/fishing/buildboatl .htrn

JHM Technologies, Inc., 1088 Grant Street, Fenton, MI 48430,
wwwrtmcompositescom

All American Racers, Inc., 2334 South Broadway Santa Ana, CA 92707,
http://www.a]lamericanracers.com/mfg/composite.html

NetComposites, Tapton Park Innovation Centre, Brimington Road, Chesterfield,
S41 OTZ, UK, http://www.netcomposites.com/education.asp?sequence=57

Jean-Yves Poirier, Solo Infusion, Professional BoatBuilder Magazine, P. O. Box
78, 41 WoodenBoat Ln., Brooklin, ME 04616-0078, 103, October/November
2006, www.proboat-digital.com

http://www.boatdesign.net/articles/foam-core/

Gutowski, TG, Advanced Composite Manufacturing, John Wiley & Sons, Inc.,
New York, 1997

157

[25]

[26]

[27]

[28]

[29]

Editor, Sizes. Darcy. http://www.sizes.com/units/darcv.htm. Last revised 4 Feb,
2002. Accessed 01 Apr, 2007.

Editor, Wikipedia. Darcy. http://en.Wikipediaorg/wiki/Darcy. Last revised 18
Mar, 2007. Accessed 01 Apr, 2007.

Schulz, P.E., “Effects of Arch Camber and Boundary Condition on Impact-Based

Energy Absorption,” Master’s Thesis, Mechanical Engineering Dept., Michigan
State Univ., East Lansing, MI, 2006.

Instron Headquarters, 825 University Ave, Norwood, MA, 02062-2643,
www.instron.us

Gougeon Brothers, Inc., P. O. Box 656, Bay City, MI 48707,
www.mosetepoxycom

158

    
  

 

1|ljlllllijjjlljljfiljlll1111161131

     

9