.Wmmu ”#2:
. .flnmguhnmmmm.
{!

4.:-

54 a
‘ “ukmmmmmm.
50.3. km
.._. bur
mmwflvfi
.. 14.3.
.. .544. .z .4.-
243 n,.il“..5.n. FL:
.« 545... a...
‘5 n 1:

:1:
ml: .I
9...
.1 .Q
~ !

 

. :.
6 .i.
£2... .r

q...

1..

1.25.
._

 

LA!

afifimu... . . HE

 

 

1 LIBRARY
3 A q Michigan State
0 University

 

 

This is to certify that the
thesis entitled

VARTM PROCESS WITH SOME MODIFICATIONS

presented by

ANUPAM DHYANI

has been accepted towards fulfillment
of the requirements for the

MS. degree in Mechanical EngineerinL

 

 

0&4 <95;

Major Professor’s/Signature

Aim; le-Lzooafiw

Date

MSU is an Affirmative Action/Equal Opportunity Employer

« _.—-—-—o--.---A—-—A—

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K'lProj/Acc8Pres/CIRC/DateDue indd

 

VARTM PROCESS WITH SOME MODIFICATIONS

By

Anupam Dhyani

A THESIS
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Mechanical Engineering

2008

ABSTRACT
VARTM PROCESS WITH SOME MODIFICATIONS
By

Anupam Dhyani

Vacuum Assisted Resin Transfer Molding (VARTM) is a technique which
has many parameters to be controlled and it is not always possible to have the
same quality and similar properties with different batches produced by it. This
makes it a manufacturing technique which is not suitable for mass production of
composites. To be able to control all the parameters, when some of them affect,
directly or indirectly, the others, is a challenge and is the main key to promoting
VARTM. In this study, the properties of the specimens produced by VARTM will
be discussed. The manner in which the properties could be controlled and how
batch by batch repeatability of these samples can be maintained will also be
addressed. A term “BUBBLE TRAP VARTM” is suggested for a modification
made in the setup. Also how the modified technique could be used to
manufacture two surface-finished and transparent (to a degree) specimens will
be discussed. With the transparency, it would be possible to get a better insight
of the voids suspended inside the sample and also will be able to study the inter-
laminar damage which may occur when testing is done on the sample. Overall,
this technique can be used to produce transparent and two surface-finished

specimens and is cost effective.

To my father, Shri Pushp Raj Dhyani,mother Dr. (Mrs) Ashe Dhyani, sister
Apoorva Dhyani and the inspiration in my life, Jaya

iii

ACKNOWLEDGEMENTS

I would like to thank Dr. Dahsin Liu, my research advisor, who has guided me
in writing this thesis, and has taught me many valuable lessons in life. This work
would not have been possible without his faith in me and his invaluable
expenence.

I would also like to thank my colleagues Dr. Elias Shakour, Goijing Li, Shawn
Klann, Kirit Rosario, Jeffrey Fuller and Brandon Gulker. Without whose help,
support, advice and invaluable camaraderie this work would not have
materialized.

Thank you all!

iv

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. viii
LIST OF FIGURES ............................................................................. ix
CHAPTER 1
INTRODUCTION .................................................................................. 1
1.1 Composites background and History ......................................... 1
1.2 Literature survey .................................................................. 4
1.2.1 Flow of resin ................................................................. 4
1.2.2 Compaction / Permeability ............................................... 6
1.2.3 Improvement of Quality ............................................................. 7
1.2.4 General / New Techniques ............................................... 7
1.3 Statement of problem and Objectives ....................................... 8
1.4 Organization of the Thesis ..................................................... 8
CHAPTER 2
VACUUM ASSISTED RESIN TRANSFER MOLDING .................................. 10
2.1 Raw materials ................................................................................... 11
2.1.1 Mold .......................................................................... 12
2.1.2 Mold releasing wax ....................................................... 13
2.1.3 Reinforcing Fabric ...................................................... 14
2.1.4 Resin ......................................................................... 14
2.1.5 Measuring balance ........................................................ 16
2.1.6 Infusion media ......................................................... 17
2.1.7 Release film / peel ply ................................................... 18
2.1.8 Vacuum bag ................................................................ 19
2.1.9 Sealant tape ................................................................ 19
2.1.10 Inlet and outlet piping .................................................... 20
2.1.11 Helix tubing ................................................................. 21
2.1.12 Clamping vice .............................................................. 22
2.1.13 Cutting tools ................................................................ 22
2.2 Equipment .......................................................................... 23
2.2.1 Resin trap ................................................................... 23
2.2.2 Vacuum pump ............................................................. 25
2.2.3 Ultrasonic leak detector ................................................. 26
2.3 Basic setup of VARTM .......................................................... 27
2.3.1 STEP 1: Tool preparation ............................................... 27
2.3.2 STEP 2: Fabric layup... ................................................. 28
2.3.3 STEP 3: Inlet and outlet infusion media placing .................. 30
2.3.4 STEP 4: Peel ply placing ................................................ 31

2.3.5 STEP 5: Infusion media and Helix tubing ........................... 33

2.3.6 STEP 6: Placing inlet and outlet tubing and vacuum bag ........ 35
2.4 Processing ......................................................................... 36
2.4.1 Vacuuming process ...................................................... 36
2.4.2 Infusion process38
2.4.3 De-molding process ................................................... 39
CHAPTER 3
CASE STUDIES .................................................................................. 40
3.1 Resin flow ........................................................................... 40
3.2 Orientation of fibers ............................................................... 42
3.3 Infusion media ..................................................................... 50
3.4 Type of fabric ....................................................................... 56
3.5 Silicone rubber sheet ......................................................... 59
3.6 Compaction ......................................................................... 64
3.7 Specimen dimensions ......................................................... 65
3.8 Final setup ........................................................................ 70
3.9 Changes to the final setup ...................................................... 72
3.10 Summary of case studies ....................................................... 75
CHAPTER 4
TESTING METHODS AND RESULTS ................................................... 76
4.1 Flow front monitoring ............................................................ 76
4.2 Burn off test ........................................................................ 77
4.2.1 Composite density and volume calculation ........................ 79
4.2.2 Fiber density and volume calculation ................................ 80
4.2.3 Matrix volume calculation ............................................... 80
4.2.4 Void volume and percentage calculation ........................... 81
4.3 Edge replication on microfilm .................................................. 81
4.4 Impact testing ...................................................................... 82
4.5 Load-Deflection curve ............................................................ 83
4.6 Summary of testing ............................................................... 85
4.7 Results and discussions ...................................................... 87
4.7.1 Resin flow ................................................................ 87
4.7.1.1 Fabric at various angles .................................... 87
4.7.1.2 Various types of fabric ....................................... 89
4.7.1.3 Various setups ................................................. 91
4.7.1.4 Stacking thickness ............................................ 93
4.7.1.5 Dimensions ..................................................... 95
4.7.1.6 Compaction time .............................................. 95
4.7.2 Load-Deflection curve ................................................... 96
4.7.3 Thickness ................................................................... 99
4.7.4 Summary of results ................................................... 103
CHAPTER 5
LARGE AND COMPLEX STRUCTURES ................................................ 105

vi

5.1 Large structures ............................................................... 105

5.2 Complex and sandwich structures ........................................ 111
5.2.1 Arches ...................................................................... 111
5.2.2 Sandwich structures .................................................... 113
5.2.2.1 Design I ....................................................... 114
5.2.2.2 Design ll ...................................................... 116
5.2.2.3 Design ”I ...................................................... 120
CHAPTER 6
CONCLUSIONS AND RECCOMMENDATIONS ....................................... 123
6.1 Conclusions ..................................................................... 123
6.2 Recommendations for future work .......................................... 127
APPENDICES ................................................................................. 129

Appendix A: Pictures of impacted specimens (backlit) top and rear face....130
Appendix B: Resin flow fronts with varying setups and stacking thickness..132

Appendix C: Properties chart ............................................................ 135
Appendix D: Darcy’s Law ............................................................... 136
REFERENCES .................................................................................. 137

vii

LIST OF TABLES

Table 2.1 Types of resin / hardener combinations .................................. 15
Table 3.1 Flow times at different compaction times ................................ 65
Table 3.2 Flow times using various setups ............................................ 73
Table 4.1 Specimens without infusion media ......................................... 88
Table 4.2 Specimens with infusion media ............................................. 88
Table 4.3 Flow time for various fiber types ............................................. 90
Table 4.4 Properties at various setups .................................................. 91
Table 4.5 Properties at various stacking thicknesses ............................... 94

Table 4.6 Flow time and property variation with varying compaction time. . ...96

Table 4.7 Specimen batches and properties .......................................... 99
Table 4.8 Properties for all the setups ................................................. 101
Table 4.9 Summary Chart ................................................................. 104
Table CI Gougeon Brother’s lnc., PRO-SET Epoxy Technical and Handling Data
....................................................... 135
Table C2 Various available mold release wax ....................................... 135

viii

LIST OF FIGURES

Figure 2.1 Schematic of the VARTM process ................................... 12
Figure 2.2 Glass mold ................................................................. 13
Figure 2.3 Molding wax ............................................................... 13
Figure 2.4 Reinforcing fabric ...................................................... 14
Figure 2.5 Infusion Resin and hardener .......................................... 16
Figure 2.6 Measuring balance ...................................................... 17
Figure 2.7 Infusion media ......................................................... 18
Figure 2.8 Peel ply/ release fabric .................................................. 18
Figure 2.9 Vacuum bag ............................................................... 19
Figure 2.10 Sealant tape ............................................................... 20
Figure 2.11 Inlet and Outlet pipes ................................................. 20

Figure 2.12a) Helix tubing .............................................................. 21

Figure 2.12b) Perpendicular flow ....................................................... 21

Figure 2.13 Clamping vice ............................................................. 22

Figure 2.14 Cutting and scraping tools .......................................... 23
Figure 2.15a) Open and closed view of the resin trap ........................... 24
Figure 2.15b) Schematic of Resin trap ................................................ 24
Figure 2.16 Vacuum pump ............................................................ 25
Figure 2.17 Ultrasonic leak detector ................................................ 26
Figure 2.18 Mold preparation ......................................................... 28
Figure 2.19 Fabric layup ................................................................ 29

ix

Figure 2.20 Infusion media on inlet and outlet ................................... 31

Figure 2.21 Peel ply placing ........................................................... 32
Figure 2223) T-joint and straight joint .................................................. 33
Figure 2.22b) Distribution media on top and helix tubing .......................... 34

Figure 2.23 Inlet and outlet tubing and vacuum bagging ........................ 35
Figure 2.24 Complete VARTM setup ................................................ 37
Figure 2.25 Complete VARTM setup with flow ................................... 38
Figure 3.1 Flow of resin in the fabric ............................................... 41

Figure 3.2 Fiber orientation ......................................................... 42
Figure 3.3 Flow times of different orientations of fibers ......................... 43
Figure 3.4 Flow lines at different time intervals for [016 ........................ 43
Figure 3.5 Flow lines at different time intervals for [1516 ........................ 44
Figure 3.6 Flow lines at different time intervals for [3016 ........................ 44
Figure 3.7 Flow lines at different time intervals for [4516 ............................. 44
Figure 3.8 Flow lines at different time intervals for [90].; ........................ 45
Figure 3.9 Thickness variations at [9015 ............................................. 46
Figure 3.10 Thickness variations at [4515 ............................................. 47
Figure 3.11 Thickness variations at [30]5 ............................................. 47
Figure 3.12 Thickness variations at [15].; ............................................. 48
Figure 3.13 Thickness variations at [0]5 ............................................... 48
Figure 3.14 Setup of fabric at various angles ........................................ 49
Figure 3.15a) Setup with no infusion media ............................................ 51
Figure 3.15b) Setup with infusion media ................................................. 51

Figure 3.150) Setup with half-way infusion media ..................................... 52

Figure 3.15d) Combined picture of all the three setups .............................. 52
Figure 3.16 Flow of resin through the three setups ................................. 54
Figure 3.17 The air bubbles retracting back in the setup with no infusion meg:
Figure 3.18 Types of fabric ................................................................ 56
Figure 3.19a) Bubble trap setup (top view) .............................................. 57
Figure 3.19b) Bubble trap setup (side view) ............................................. 58
Figure 3.19c) Bubble trap setup ............................................................ 58
Figure 3.20 Making the silicone bag .................................................... 60
Figure 3.21 Silicone rubber setup ....................................................... 61
Figure 3.22 Formation of dry spots with the silicone sheet setup ............... 62
Figure 3.23 Cross-sectional view of the setup with / without gap ............... 62
Figure 3.24 Setup with silicone rubber sheet ......................................... 63
Figure 3.25 Setup with different dimensions .......................................... 66
Figure 3.26 Schematic of flow front ...................................................... 67
Figure 3.27 Flow times with varying stacking sequence .......................... 68
Figure 3.28 Degassing of the resin with the vacuum chamber .................. 69
Figure 3.29a) The final setup (top view) ................................................. 70
Figure 3.29b) Final setup (side view) ..................................................... 71
Figure 3.30 Final setup .................................................................... 72
Figure 3.31 Final setup (side view) ..................................................... 73
Figure 4.1 Top and bottom cameras for flow front monitoring ................. 76
Figure 4.2 Burn off testing .............................................................. 78

xi

Figure 4.3
Figure 4.4
Figure 4.5

Figure 4.6

Figure 4.7
Figure 4.8
Figure 4.9
Figure 4.10
Figure 4.11
Figure 4.12
Figure 4.13
Figure 4.14
Figure 4.15
Figure 4.16
Figure 4.17
Figure 4.18
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.53
Figure 5.5b

Figure 5.6

3p and 5p Polishing powders ............................................ 82
Impact loading machine setup .......................................... 83
Load-Deflection curve ..................................................... 84

Backlit impacted specimen showing primary and secondary

damage area ................................................................ 85

Flow time at different stacking thicknesses .......................... 93

Load-deflection curve for various layers .............................. 97
8-layer specimens with different batches .............................. 97
3-layer specimens with different batches .............................. 98
12-layer specimens with different batches ............................. 98
5-layer specimens with different batches .............................. 98
Thickness chart for various layers ..................................... 100
Thickness chart for varying setups ................................... 100
Sample with plastic plate setup ....................................... 102
Sample with plastic plate, thin plastic sheet and peel ply ...... 102
Sample with peel ply and infusion media .......................... 102
Sample with no infusion media or peel ply ........................ 103
Multiple infusion lines used for resin infusion ..................... 106
A typical wind blade ..................................................... 107
Parts of the wind blade ................................................. 107
Manufacturing process ................................................. 108
Manufacturing arched Specimens ................................... 112
Schematic for arched specimen setup ............................. 113
Arched Specimen ........................................................ 113

xii

Figure 5.7 Schematic of the sandwich structure ............................... 115

Figure 5.8 Manufactured sandwich composite ................................. 115

Figure 5.9 Making the foam ......................................................... 117

Figure 5.10 Side view of sandwich structure before infusion ................. 117

Figure 5.11 Infusion process .......................................................... 118

Figure 5.12 Sandwich composite .................................................... 119
Figure 5.13 Side view ................................................................... 119
Figure 5.14 Cross-sectional view of the 1St step ................................. 120
Figure 5.15 Manufactured sample with matching domes ............................ 122
Figure A1 3 Layers ......................................................................... 130
Figure A2 5 Layers ......................................................................... 130
Figure A3 8 Layers ......................................................................... 130
Figure A4 12 Layers ....................................................................... 131
Figure A5 14 Layers ....................................................................... 131
Figure 81 Resin flow with peel ply and infusion media on top ..................... 132
Figure 82 Resin flow with only plastic plate on top ................................... 132
Figure 83 Side view of setup with only plastic plate ................................. 133
Figure 84 Resin flow using perforated sheet, peel ply and plastic plate... 133
Figure BS Resin flow with no infusion media or peel ply ............................ 134
Figure 86 Through thickness flow through a thick stacking sequence .......... 134

xiii

CHAPTER 1

INTRODUCTION
1.1 Composites Background and History

Right from the primitive age, man has always been endeavoring to make
things stronger, from reinforcing his house to strengthening monumental
buildings. This continuous strive to build things stronger has led to the
advancement of the materials he has been using. Prior to knowing the chemistry
of metals and the use of advanced materials, man found his ways of
strengthening all his belongings like his house, his means of transport, his
weapons, his clothes and his utensils. He made his house of mud and reinforced
it with straw so that it doesn’t give in to inclement weather; it was his first idea of
composites and was not named so then though.

With the advent of metals, man got a chance of building stronger
components, things which were stiffer than his “primitive composites”. This
continued for many centuries until he refined the art of making stronger materials.
With the increasing demand and diverse needs, e.g. to conserve energy, man
had to come up with materials which were stronger yet lighter. This is how the
modern composites took birth. In recent times, the art of composite
manufacturing has refined itself to such an extent that right from paving our
roadways with asphalt concrete or Portland cement, we are making the airplanes
lighter as in the Boeing 787 “Dreamliner” which is 50% composite (by weight).
Defense departments all around the world are shifting to composite materials for

their combat vehicles as their use improves maneuverability, gives strength and

high energy absorption. Even though conventional homogeneous metals like
steel, brass and aluminum are still being used for the basic framework of
structures; many advanced structures are being quickly replaced by composite
materials. Not only do they give the advantages of low density (lightweight), high
tensile strength, high stiffness and immense energy absorption, they can also be
tailored for various customized engineering requirements and provide design
flexibility. Composites also add to the overall aesthetics amalgamated with the
engineering requirement. A composite is a combination of two or more materials
(reinforcing elements, fillers and matrix binders) different in form and composition
on a macro-scale. The constituents retain their identities, that is, they do not
dissolve or merge completely into one material although they act in concert.
Normally, the components can be physically identified and exhibit an interface
between one another.

Composites continue to play an important role in industries like aerospace,
automotive, marine, medical, military, recreational and many more. Due to the
rapid spread of the use of composites in many important industries, it becomes
vital to manufacture these composites for various requirements at reduced costs
and high quality. During the last few decades, researchers and industries have
invested lots of money in the development of cost-effective manufacturing
techniques for composites [1]. For composite production, many different types of
techniques are used currently in the industries. For example, RTM (Resin
Transfer Molding), filament winding, autoclaving, SCRIMPTM (Seemann’s

Composite Resin Infusion Molding ProcessTM), and spray gun techniques are

some of the popular techniques available in the composite manufacturing market
currently. While RTM has an advantage of producing composites with high fiber
percentage, SCRIMP has the advantage of producing large composites in a very
short time (e.g. the hull of a boat). Autoclaving possess has the advantage of
making high-quality composites with low void content for maximum structural
efficiency (e.g. the tails and wings of aircrafts). Filament winding (e.g. golf clubs)
can provide engineering properties like high tensile strength. Although RTM,
autoclaving and filament winding are very capable of producing high quality
composites yet they are very expensive procedures as they are labor intensive
and require very sophisticated tooling and high setup costs.

The constant need for a manufacturing technique which produces quality
specimens and is low on cost led the composite manufacturing industry to give
birth to Vacuum Assisted Resin Transfer Molding, commonly known as VARTM.
With the capability of producing composites with a fiber volume fraction close to
autoclaving, VARTM has, over the years, shown the potential of being an
effective technique for manufacturing composites. Researchers all around the
globe, over the past two decades, have shown special interest in this technique
and because of the potential it possesses, have led to the refining of this process
to a great extent. VARTM has developed into a technique, because of its simple
setup and cost effectiveness, which has the potential of replacing classical
techniques like RTM and autoclaving. Although RTM and autoclaving are widely
used for high-quality composite parts, the research on VARTM has led to many

possibilities in which it can be used in different capacities.

Research has been going on for some time and many aspects of VARTM
have been looked into. Since there are many parameters involved in this
technique, studies have been made to address these problems and conclusions
have been drawn out to assist the development of this technique. Some of the
parameters involved are flow of resin in the preform, dimensions of the
composite manufactured, compaction and permeability of the preform.
Researchers have addressed these issues in detail and have come up with some
solutions to the problems encountered.

1.2 Literature survey

The literature on VARTM technique has been surveyed and categorized into
the following areas as the studies are done for various parameters.
1.2.1 Flow of Resin

Studies have been done on the flow of resin through the fiber preform in the
VARTM process. Because voids formed during the mold filling stage of VARTM
become defects in the fabricated part, Johnson and Pitchumani [2,3] studied the
flow and came up with an active flow control strategy to eliminate these defects
by guiding the flow along a desired path. By implementing local induction heating
they found out that the viscosity of the resin and the in homogeneity of the
permeability of the preform layup. By their active flow control they were able to
reduce fill time, improve flow front uniformity and eliminate dry spot formation.

Grimsley et al [4] studied the effects of compaction On the flow of resin. They

also studied the change in the direction of the flow front as the thickness of the

stacked layers of fabric increased. They concluded that as the thickness of the
preform increased, the through thickness flow of resin became faster.

Heider et al [5] studied the flow in thick sections and found out that the flow
times were increased and void formation was more likely to happen, so they
came up with a design methodology where they placed several layers of
distribution media of varying lengths through the thickness of the part. They
improved the infusion times with their setup along with reducing voids; however
the flow field was very complex. They did not address the concerns about
property change due to the inter laminar distribution media.

SCRIMP (Seemann’s Composites Resin Infusion Molding Process) [1], which
is a patented technique, was widely used in the manufacturing of large structures
such as boat hulls. The feasibility of this technique was reviewed by Beckwith et
al.

Yoon et al [6] studied the gravitational effects during the flow of resin in
VARTM for tall structures. They came up with an analytical solution for flow and
pressure prediction by incorporating the gravitational effects; however the
assumptions made were one dimensional, laminar and incompressible resin flow,
and homogeneous and uniform thickness preform. The parametric studies
indicated the important parameters influencing the flow time including mold angle
(00-900), permeability, radius and length of injection tubes and preform cross-
sectional area. They concluded that the flow time could be reduced for downward

infusion (gravity assisted) case where the resin bucket was placed at the same

height as the inlet tube. Also the length of the infusion tube and the area of the
preform could affect the flow times.

Bender et al [7] investigated the flow rate control for VARTM which allowed
vacuum pressure to be applied to the resin bucket to generate a computer
controlled vacuum differential between the injection and vent gate. They also
studied how to control the resin flow to reduce void content in the part. The
system successfully controlled the flow for a wide range of permeability and
viscosity values

Walsh et al [8] suggested a method alternative to conventional VARTM
known as FASTRAC which reduced cost and wastage. They used a double
vacuum bag with an optional distribution media option. The cost was reduced by
eliminating the distribution media thus causing less wastage.

1.2.2 Compaction / Permeabllity

The effect of flow time, permeability and compaction were discussed in
[9-17]. These studies detailed the permeability study and the factors affecting the
permeability. Algorithmic models developed for flow front determination and
permeability effect were also detailed in these studies. Fiber permeability had a
strong influence on resin impregnation behavior. A procedure was developed for
determining the anisotropic in-plane permeability of fiber performs from constant
flow rate molding. Gauvin et al [10] designed and optimized a reliable mold filing
model with accurate description of resin flow through the reinforcement. Dong
[11] proposed a computational procedure for permeability based on flow front

location and flow time data was derived. Hammani [13] discussed the factors

affecting the permeability measurement in VARTM. He characterized the

parameters affecting and observed the effect on infused length and fiber fraction.

1.2.3 Improvement of Quality

The quality of the specimens such as thickness, surface finish, void content
and flow control were compared with other techniques such as RTM and
autoclaving in [1 ,18-24]. Solange [12], Gama et al [13], Juska et al [18] discuss
the various parameters which were involved and the challenges in controlling
these properties. Some new methods were also proposed in these studies such
as silicone bagging and FASTRAC. The thickness variation and surface quality
were discussed in [25,26]. Heider et al [5] discussed the surface quality of the

techniques like SCRIMP, VARTM and silicone bagging.

1 .2.4 General / New techniques

New methods and techniques such as reusable vacuum bags, multiuse
tooling system, water soluble tooling, double chamber VARTM and UV-VARTM
were proposed in the studies [22,23,27-29]. Double chamber VARTM used two
molds which were made of wood or plastic. This process was capable of making
large parts and could be used to make thermoplastics as well. UV-VARTM did
not need extensive training and licensing and reduced the cost of the final
product. Reusable vacuum bag helped reducing the cost of vacuum bags giving

the same finish as the ones which use the one-time used vacuum bag. Water

soluble tooling assisted in making complex shapes as once made, could be

dissolved in water and reshaped for further use.

1.3 Statement of Problem and Objectives

Vacuum Assisted Resin Transfer Molding is a process which involves many
parameters to be controlled. These parameters affect directly or indirectly each
other and thus make them difficult to control simultaneously. They produce large
variations in the properties and quality of the samples produced with VARTM,
due to effect of one or more parameters involved, thus hindering VARTM from
being a commercial composite manufacturing technique. To understand the
behavior of the parameters individually and simultaneously, and to be able to
control most of the parameters involved, is the primary goal of this study. Another
goal is to obtain batch-to-batch repeatability of the samples manufactured by this
technique. Some modifications for achieving the objectives of the study would
also be made to the existing procedure so that the desired result can be

achieved.

1.4 Organization of the Thesis
This thesis is organized into six chapters overall. Chapter 1 gives a brief
introduction to composites, summarizes the literature survey done on composite

manufacturing; it also defines the problem posed and the objectives of the study.

Chapter 2 describes in detail the setup, the raw materials used and the
precautions to be taken while setting up the procedure to manufacture
composites with VARTM.

Chapter 3 details the various case studies done while making the specimens.
It also explains the various modifications made to the current setup and proposes
a suggested setup which helps control some of the parameters effectively. It also
explains how the various setups can be used effectively.

In Chapter 4, the first portion explains the various testing methods involved.
The second portion presents the results and discusses the outcome and gives
reasons for the outcome.

Chapter 5 briefly describes the use of the proposed technique for
manufacturing large and complex structures like sandwich composites and
arches with VARTM.

Chapter 6 summarizes the conclusions drawn and recommendations for
future work in this topic are proposed.

Some tables, pictures and data which are not presented in the chapters are

listed in the appendices.

CHAPTER 2

VACUUM ASSISTED RESIN TRANSFER MOLDING

With so many composite manufacturing techniques available in the industry
it becomes essential to select a process which is easy to use, is cost effective
and produces composite materials which are of high uniformity and quality. This
chapter briefly summarizes the Vacuum Assisted Resin Transfer Molding
(VARTM). It also explains in detail the necessary raw materials and conditions
required to manufacture composites with the VARTM technique.

As the name suggests this technique uses vacuum pressure to pull out resin
from the inlet to the outlet which is connected to a vacuum pump, with vacuum
pressure through an already laid reinforcing fabric material. The reinforcing
material is placed on a single flat tool. On top of this is placed an infusion media
which assists the flow of resin through the reinforcing fabric.

VARTM is an inexpensive technique to produce composites. Unlike Resin
Transfer Molding (RTM), which uses two molds and an injection pressure to
infuse the resin in the fabric, VARTM uses one mold and vacuum instead, this by
itself reduces the cost tremendously. Although fiber volume fraction is slightly
less than that produced by autoclaving, it does not require any sophisticated
equipment such as an autoclave and can be done at room temperature. All these
characteristics make VARTM a potential candidate for replacing some current
techniques to a great extent and by reducing the cost of manufacturing, making

composites readily available in the industry and other applications.

10

There are some elements which are necessary to be considered when this
technique is employed for manufacturing composites. The elements necessary
are reinforcing fabric (such as glass, carbon or Kevlar fiber), binding material (a
mixture of resin and hardener). The tooling or the mold and vacuum bag are the
other elements that are involved in the process. Another important element is the
vacuum pressure which assists the resin to transfer from the inlet to the outlet
wetting the reinforcing fabric.

2.1 RAW MATERIALS

The VARTM system constitutes of laying the preformed fabric onto a
smooth surfaced tool (which acts as the first mold). Some release material like
wax is applied to the mold so that after the cure of the part it is easy to demold
the part from the fixed mold. Then the resin inlet and vacuum outlet piping is
placed at both the ends of the laid fabric. A low permeability infusion media and a
release fabric are placed on top of the fabric and finally the vacuum bag is placed
on top; the vacuum bag is sealed with the help of a sealing tape and vacuum is
drawn using a vacuum pump. After the setup of this process, the resin is infused
(pulled by the vacuuming force of the pump) through the fabric. Thereafter the
inlet and outlet manifolds are clamped so that the resin inlet can be stopped. The
impregnated fabric is let to cure. The curing time is decided by the type of resin
and hardener combination used for the resin impregnation.

The Figure 2.1 below shows the schematic of the VARTM process.

11

 

 
  
   
    

 

Stale“ Infusion media
PImna
Inlet helix
Tubing Outlet Helix tubing
Peel I
Inlet p y
Infusion Outlet Infusion media
Media
Glass
Sealant Tool
tape
. * Outlet piping OUT
Vacuum bag Polished —>
Tool
*— Resin bucket Resin trap —> / '
T Vacuum
Pump
Figure 2.1 Schematic of the VARTM setup
2.1.1 MOLD

This is usually a stiff material which does not change its shape. For making
flat panels it is usually chosen to be a smooth plate so that one side of the
composite has a smooth finish. In this case, glass was used as it had fine finish
and was stiff. It was also used because it was transparent and the resin flow

could be monitored from both sides. Figure 2.2 shows the glass mold.

 

Figure 2.2 Glass mold
2.1.2 MOLD RELEASING WAX
This is usually applied to the tool surface before the placing of the fabric.
After the application of the wax it was set to dry and then wiped off with a clean
cloth or paper towels, giving a fine finish to the tool side and making it so smooth
that the impregnated preform would not stick to it once it cures and could be
taken off easily. A solid release wax with the trade name TR-102 manufactured

by TR industries was used in this study as shown in Figure 2.3

 

Figure 2.3 Molding wax

2.1.3 REINFORCING FABRIC

The backbone of any composite material is the reinforcing fibers. They are
usually stacked in layers when the composite part is produced. The materials
generally are E (electrical Grade) glass, S (structural grade) glass, carbon or
Kevlar fibers. There are many forms of these fibers; some are unidirectional
which are stitched together. Some are woven in coarse manner while some in a
fine weave. Each type of fabric caters to a particular need. Each layer of the
fabric can be stacked at different orientations and of varying thicknesses
depending on the engineering application. Figure 2.4 shows the various fabrics

used in the study.

 

Woven
Fine Glass
Fiber in

Woven Strand

Coarse form
Unidirectional
Stitched

Figure 2.4 Reinforcing fabric

2.1.4 RESIN

With the advancing chemical knowledge, resins are being tailored such that
they can be used in the most effective manner. The resin is used to impregnate
the fabric preform. The characteristics of the resin which directly affect the
VARTM process are viscosity, gel time and the type of resin. Less viscous resin

might infuse very quickly resulting in deformed parts, the deformation is due to

the uneven wetting of the fabric making some portions resin rich and the others
full of voids, while faster curing resins (usually with high viscosity) gel too quickly
and cannot sometimes impregnate the entire part.

Most commonly used resins are vinyl esters and epoxy resins. Styrene
emission is one of the drawbacks of the vinyl esters, which causes shrinkage of
the parts resulting in stresses between the layers. They are not expensive
though. On the other hand, the epoxy resins are used very commonly as there is
fairy less emission. They are provided with hardeners of varying viscosities so
that one can control the type of flow (fast or slow which depends on the resin
viscosity) by changing the hardener. In this study the resin and hardener used
were manufactured by PROSET (resin 117LV and hardener 226).

Table 2.1 below shows the various resin/ hardener used in the study.

Table 2.1 Types of resin / hardener combinations

 

PRO-SET Epoxy Technical and Handling Data

 

 

 

 

 

 

 

 

 

 

 

 

Property 1 17LV/224 1 17LV/226 117LV/229
Density g/ml 1.114 1.09 1.114
Viscosity cP 800 350 310
Mix Ratio Weight 100:33 100:30 100:31
Mlx Ratio Volume 100:34 100:35 100:37
Pot Life 100 9
sample In min

65° F 29 85 1 10
72 ° F 21 44 61
80 ° F 15 25 40
Working Time @ 72°
F 20 min 1 hr 2 hrs
Gel Time @72° F 2.25 hrs 3 - 3.5 hrs 5 hrs

 

 

 

 

15

 

Figure 2.5 Shows the 117 LV and 226 hardener. A detailed properties chart is

available in the Appendix C.

 

Figure 2.5 Infusion resin and hardener

2.1.5 MEASURING BALANCE

A sensitive measuring balance shown in Figure 2.6 was used to measure
the amount of resin and hardener mixture to be used for infusing. This has to be
very sensitive (measures up to .01 grams) as inaccurate weights can cause
uneven curing and change the properties of the resin flow inside the preform,
also cause gel times to change. This was observed when in one of the
experiments the resin and hardener were not mixed accurately, the mixture
started to over-heat and due to this exothermic reaction the mixture in the resin
pot started giving out huge amounts of black fumes with intolerable smell and all
the epoxy spilled out of the pot after melting the pot. Also the gelling color
changed from the regular amber color to a dark brown shade. This inaccuracy

can cause hazardous changes, so care should be taken that the measured

amount is as accurate as the measuring balance can measure, slight changes

may cause irrevocable damage.

 

Figure 2.6 Measuring balance

2.1.6 INFUSION MEDIA

This is used to distribute the resin from the inlet piping to the preform.The
resin moves in the thickness direction thereafter. The infusion media is placed on
top of the preform. The inlet manifold is placed on it as it extends on the inlet side
beyond the preform, on the outlet it only extends a little so that it does not touch
the outlet helix tubing. The gap between the outlet helix tubing and the infusion
media is approximately 31.75 mm (1.25 in). There are many types available in
the market and all are used in different capacities. The infusion media used in
this study was INKA 7001 manufactured by Colbond-USA Inc. Figure 2.7 shows
the various types of infusion media used. It should also be noted that the use of
infusion media qualifies the process as Seeman Composites Resin Infusion
Molding Process (SCRIMP) which is a patented technique. This study is only for

academic discussion and use of the infusion media will only help study the

process better. Also comparing the technique in this study to the already in use

SCRIMP will help better understand the process.

INKA
7001

 

Figure 2.7 Infusion media

2.1.7 RELEASE FILM / PEEL PLY

This is a cloth like material which is porous and allows the resin to flow
through it; however, it does not let the resin stick to it. It also allows the infusion
media not to adhere to the preform. 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 used in this study was Ply by Composites One (Figure 2.8)

 

Figure 2.8 Peel ply/ release fabric

2.1.8 VACUUM BAG

The vacuum bag is easily one of the vital components of VARTM. It forms
the second mold (flexible) and covers all the components such as the fabric, peel
ply, infusion media and inlet and outlet helix tubing. It also maintains the vacuum
pressure gradient and helps contain the resin in the system. It is fixed to the solid
mold (glass base) with the assistance of mastic sealing tape i.e. tacky tape.
There are a few types of vacuum bag, some used are of non stretchy material
and others are flexible. The advantages, drawbacks, functioning and the ways of

set up would be discussed later in the chapter. Figure 2.9 shows the two bags.

 

Non

Stretchy

Green

Vacuum \

bag Stretchy
Vacuum
bag

Figure 2.9 Vacuum bag

2.1.9 SEALANT TAPE
This tape is used for the vacuum bag to stick to the mold; it adheres to both
the vacuum bag and glass mold strongly and thus helps to completely draw out
air and seal the vacuumed section to remain that way. It makes the system
airtight. Apart from its use as the major sealant it is also used in the sealing of the

various joints in the manifold system (inlet and outlet). Also it can be used to

cover any sharp portion inside the bag which could cause the bag to rupture and
ruin the vacuum. The tape comes in various grades depending on the

temperature. Yellow sealant tape was used in this study shown in Figure 2.10.

    

. I _. -..,
'r F?‘ .- ‘ ‘1‘"

 

Figure 2.10 Sealant tape

2.1.10 INLET AND OUTLET PIPING

Inlet piping is used to draw resin into the vacuum chamber and outlet is
attached from the vacuum chamber to the resin trap (discussed later). This pipe
has to be sturdy enough not to collapse with 1atm of vacuum pressure. It should
also be transparent as flow of the resin can be monitored and controlled. In this

study, polyethylene pipes were used as they meet both the requirements.

6.35 mm Inner

 

diameter 9.53 mm Inner
polyethylene dlameter
Inlefloutlet plpe ponetherne
Inletlouuet pipe

Figure 2.11 Inlet and outlet pipes

20

2.1.11 HELIX TUBING

Helix tubing is special spiral tubing which is used for the distribution of the
resin inside the vacuum bag. It is used both on the inlet (to distribute the resin
equally on the inlet) and on the outlet so that the extra resin, flowing away from
the fabric towards the outlet with the help of the outlet manifold, is collected in the
resin trap. The flow of resin is perpendicular to the helical tubing giving maximum

resin distribution. Figure 2.12 shows the helix tubing and the perpendicular flow.

 

Figure 2.12a Helix tubing

Flow of resin perpendicular to helix
tubing

  

“““““““““““““““““““““““““““““
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

\

 
 

Helix tubing

I

Resin
Inlet

Figure 2.12b Perpendicular flow

21

2.1.12 CLAMPING VISE

These vices are used to clamp the piping before the infusion so that the inlet
end is blocked and no air can enter the vacuum chamber, which is the vacuum
cavity within the vacuum bag sealed with the sealant tape. It is also used to
further clamp both the inlet and outlet pipes once the resin has completely
infused. This collapses the piping and help in the control of the flow in the

system. Figure 2.13 shows the vices.

 

Sharp

edged vise

Inlet /Out|et 2:5“
Pipe vice

Figure 2.13 Clamping vise

2.1.13 CUTTING TOOLS

The cutting tools include the scissors; it is used to cut the fibers from large
rolls of fabric. It is difficult to out these fabrics as the scissors do not allow
enough pressure to be applied as the fabrics are tough and cause the fabric to
be cut unevenly. This happens because the fabrics are slippery due to the fiber
being subjected to sizing when being manufactured. Thus it becomes necessary

to use a roller cutter which provides better cutting power and assists in easy

22

cutting of the fabric. The blade of the roller cutter is basically round. The fabric is

placed on a cutting mat and the roller is used to cut the fabric. (Figure 2.14)

 

Heavy

Duty

Scissors Roller
Cutter

Figure 2.14 Cutting and scraping tools

2.2 EQUIPMENT

The equipment used in the entire process is detailed in this section
2.2.1 RESIN TRAP

This is a bucket which is used to prevent the resin from going to the vacuum
pump and thus it collects any extra resin coming out of the resin manifold
towards the vacuum pump. It is sealed so that no vacuum pressure is lost. It also
is mounted with a gage which measures the vacuum pressure and indicates the
loss of pressure should there be any leak in the vacuum bag. It also acts as a
vacuum reservoir which means that even if the vacuum pump is turned off the
trap would not let the maintained vacuum pressure to be lost. Figures 2.15a and

2.15b show the resin trap.

23

 

Figure 2.15a Open and closed view of the resin trap

Pressure
Gage Top lid

Valve \ To
_. \ g) I ——_' pump

Vise to clamp —>
Lid and chamber _

Resin in

 

 

 

 

 

 

R Vacuum
Chamber

 

Resin
Bucket

2.15b Schematic of Resin Trap

24

2.2.2 VACUUM PUMP

The vacuum pump shown in Figure 2.16 is the main driving force which
draws out the entire air from beneath the vacuum bag and maintains a pressure
close 1 ATM or 101 kPa under the vacuum bag, i.e. the vacuum chamber which
sometimes loses a little bit of pressure. The pump makes noise and it sometimes
hinders the process of hearing the leaks which may be present in the vacuum
bag or the interface between the sealant and the vacuum bag. To remove this
defect we use an ultrasonic detector to detect the leaks in the bagging. This is
done by placing the detector’s head near the tacky tape seal and if there are any
variations in the digital meter inbuilt in the detector we know that there is a leak

and it is to be take care off.

 

Figure 2.16 Vacuum pump

25

2.2.3 ULTRASONIC LEAK DETECTOR

The ultrasonic detector used in this experiment was made by Am probe Inc
known by the trade name ULD 300. This is shown in the Figure 2.17 below.
Typically, human can hear frequencies from 20 Hz to 20 kHz, with sound from 20
kHz to 100 kHz being termed ultrasonic. Turbulence created by air or gas, forced
through a small orifice, generates ultrasonic sound. This ultrasonic leak detector
is designed to locate the source of the ultrasonic sound emissions. These
ultrasonic sound emissions are converted by the Leak detector to a range that
can be heard by humans, the sound generated by this unit is 32 times lower in
frequency than the sound that is received. Thus small leaks can be detected
which may be potential cause for the part to be defective with air voids if the leak

is not detected.

 

Figure 2.17 Ultrasonic leak detector

26

2.3 BASIC SETUP OF VARTM
Now that we have discussed the raw materials and equipment in detail, a
detailed study of the set up should be made to understand the process and to

minimize the losses while setting it up.

2.3.1 STEP 1- Tool preparation

This is the starting point of VARTM. The tool or the solid mold has to be
carefully prepared. Mastic sealant tape is placed on the tool to outline the area
which would form the vacuum cavity. The tool is then first cleaned with acetone
which removes any waxy material from the previous tests. Wax is applied with a
smooth cloth provided with the wax to the glass mold and let to dry for 15
minutes; this causes the wax to stiffen and assists the cleaning process.
Thereafter it is cleaned with paper towels to give a smooth and shiny finish; one
of the purposes of applying wax is to fill the micro-pores in the tool, resulting in a
Shiny finish. The wax is removed with the help of paper towels, care must be
taken that no specs of the dried wax are left on the tool when the fabric is placed
on the tool or else this might hamper the smoothness of the glass tool. The

Figure 2.18 below shows the setup of this step.

27

 

 

 

 

Sealant Glass
Tape Tool
Waxed surface
Figure 2.183 Step 1-Mold preparation
Sealant
Tape Waxed surface
as * rs
[ .9 t _ , , L Glass
Tool

 

Figure 2.18b Step 1-Mold preparation
2.3.2 STEP 2- Fabric layup
The next step is to place the fabric on top of the tool. The Figure 2.19

shown below depicts the placing of the fabric on top of the tool.

28

Fabric

 

 

 

Sealant
Tape
Polished Surface after
Wiping off wax
Fabric
Sealant
Tape
Glass
tool

Figure 2.19 Step- 2 Fabric layup (top and side view)

29

2.3.3 STEP 3- Inlet and outlet Infusion media placing

When the fabric is placed, the infusion media, cut in strips of 50.8 mm (2 in)
width and the length equal the preform width, is placed on the inlet side of the
fabric just enough to touch the fabric without overlapping it. This placing of the
infusion media assists the flow of the resin from the inlet tubing to the fabric. Also
since the infusion media is thinner as compared to the fabric, a number of layers
of infusion media should be Stacked to make the thickness equal to the thickness
of the fabric. This ensures the uniform flow of resin to the fabric. The infusion
media, cut into 25.4 mm (1 in) broad strips equal to the size of the preform, are
placed on the outlet side of the fabric just touching the fabric in similar manner as
the inlet side infusion media placing. The thickness of the infusion media on the
outlet does not matter to a great extent as it only assists the resin to flow out of

the system; however it also assists the back flow of the air. (Figures 2.20a and b)

Fabflc
Inlet side
Infusion
Outlet side
Media Infusion
Media (just
touching)
Sealant
Tape
‘— Glass
tool

 

Polished Tool

‘ ‘

INLET OUTLET

Figure 2.20a Infusion media on inlet and outlet (top view)

30

Outlet slde

Inlet side 50.3 mm Fabric Infusion
Infusion Media Media
25.4 mm

    

Sealant Tape \
m _

Glass
Tool

‘ ’

INLET
OUTLET

Figure 2.20b Step 3-Infusion media on inlet and outlet (side View)
bubbles as it does not trap the bubbles due to its high permeability. Figure 2.20
shown depicts the inlet and outlet placing of the infusion media.

2.3.4 Step 4-Peel Ply placing

Peel ply is cut and placed on top of the fabric so that it overlaps the inlet
and the outlet distribution media; it is adjusted such that it exceeds the infusion
media on the outlet side by 25.4 mm (1 in). The peel ply would not allow the
infusion media which is placed on top of it (discussed in next step) to stick to the
fibers once infusion is done. It also acts as an infusion media as it is perforated,
although it cannot be used for that application as it is not porous enough. This
plays a vital role when demolding the specimen as it assists the process by
peeling off easily and hence its name. The Figure 2.21 shown below depicts the

step of placing the peel ply over the fabric in both top and side view.

31

Fabfic

   
 

  

 

Inlet side :- _ 1.. ; Peel
Infusion 1 , ‘ . ' 5:: .5" ,. t: égazi T“ l ‘ - . Ply
i l a . . ' . , ‘ .
Meda , fie _~, _ Oufletflde
‘ ‘ Infusion
medh
Sedant
Tape Glass
Tool
‘ Polished TOOI ‘
INLET OUTLET
Inlet side Peel ply Fabric
Media Infusion
Medh
Sedant
tape
(mass
Tool
INLET OUTLET

Figure 2.21 Step 4—Peel ply placing

32

2.3.5 STEP 5-lnfusion Media and Helix tubing

. In the next step, the infusion media is out to the same size as of the fabric
and placed over the peel ply, care should be taken that the distribution media
does not exceed the fabric on the sides which will give rise to" race-tracking"
(explained later). Helix tubing is placed on top of the infusion media on the inlet
Side at a distance of 38.1mm (1.5 in) from the fabric. On the outlet side it is also
38.1mm (1.5 in) from the fabric and over the peel ply. Helix tubing on both the
inlet and the outlet side is to be connected to the piping. They are connected with
either a T joint or a straight joint. Figure 2.22a shows the two types of joints
which were used in the setup. The advantage of the straight joint is that it is
easier to setup and seal and also the flow of resin from the bucket is more
uniform in the straight joint. In the T- joint, first of all it becomes difficult for it to

seal with the sealant in the periphery and second of all the distribution of resin is

non-uniform.

   

Figure 2.22a T-joint and straight joint
The figure of the classical (straight) inlet and outlet is shown in Figure 2.22b

below. Once these are connected the vacuum bag is placed on the entire setup.

33

H 38.1 mm H 25.4 mm Infusion

 

 

 

 

 

 
 
 
  

media on top
of the fabric

Inlet

HeIIx . é . , Outlet Helix

TubIng :.- f; Tubing
Inlet side
Infusion ‘ Peel plv
Media # ‘ Outlet side

_ Infusion media

Sealant . ' ‘ .. GI
ta ass

pe Tool

‘ POIiShed TOOI
INLET 'OUTLET
Infusion
Inlet Media
Inlet slde Infusion Helix Peel ply
Media TubIng
Fabric Outlet
Sealant tape Tfiizii‘ng

38.1 mm 381 mm
‘ ‘
INLET OUTLET

Figure 2.22b Step 5-Dlstribution media on top and helix tubing

34

2.3.6 STEP 6-Placing inlet and outlet tubing and vacuum bag

This step involves the placing of the inlet and outlet tubing. These are
connected to the helix tubing with the help of straight joints or connectors and are
fixed to the outside periphery of the surrounding sealant tape with the help of
sealant tape so that they do not move. Once this is set up the vacuum bag is
placed on top of the setup and sealed with the help of the sealant tape.

The sealant tape helps to seal the vacuum bag. If the bag is large and does
have some left over portion which is not adhered to the sealant, pleats are made
to avoid any gap between the bag and the sealant tape. After this step the inlet
piping going to the resin reservoir is clamped and the vacuum pump is initiated.

Figure 2.23 below shows the set up of tubing and vacuum bag.

 

 
 
 
 
 
  
 
 
  
 
  
        
 
 
  

—> IN
Sealed Infusion Media
oipina
Inlet helix Outlet Helix
tubing Tubing
Inlet Peel ply
Infusion
Media Outlet Infusion Media
Sealant tape Glass
Tool
—>
I Outlet piping OUT
Vacuum bag Polished
Tool
.— Resin bucket

 

Resin trap —> . '
Vacuum
Pump

Figure 2.23 Inlet and outlet tubing and vacuum bagging

35

Figure 2.23 Inlet and outlet tubing and vacuum bagging continued

Infusion Media Vacuum bag

Inlet Infusion l Peel P'Y .
Media 50.8mm / / FabrIc

1’ Outlet Helix
Tubing

    
  

Sealant tape

Glass
Tool

 

38.1mm
INLET ‘
OUTLET

Inlet and outlet tubing and vacuum bagging (side view)

2.4 PROCESSING
2.4.1 Vacuuming process

Once the entire setup is done as explained earlier the vacuum pump is
started which draws out the air inside the chamber. The vacuum bag takes the
shape of the preform and the materials above and next to it. The pressure is
drawn to 1 ATM of atmospheric pressure and maintained through out the resin
infusion processes. The gage mounted on the resin trap reads the amount of
vacuum pressure in the vacuum chamber. It reads 27psi for maximum vacuum
pressure. If the reading in the gage drops then it is possible that there is a leak.
By monitoring the gage the leaks are detected. Thereafter the leaks are detected
by placing ear close to the sealant tape and a whistling sound is noticed.

Sometimes due to excessive noise of the vacuum pump, it becomes difficult to

isolate the leaking noise and thus the vacuum pump is placed far away from the
setup with large tubing connecting the trap. After sealing the tape the noise is
decreased. Once the audible noises are fade out, but still minute leaks may be
present which can be detected using an ultrasonic detector. Once all the leaks
are taken care off, the pump is shut down to check that there are no leaks. If
there are still some leaks, the gage will indicate a drop in pressure and thus leaks
can be further sealed.

The vacuuming process is a tedious process as, due to human error it is
bound to have some leaks in the vacuum bag-sealant interface. And in this
process the vacuum pressure is the main driving force. Thus careful monitoring
in the vacuuming process is necessary. Figure 2.24 shown below shows the

complete setup ready to be infused with resin.

 

 
 
 
 
 
  
  
 
 
 
      
  
   
 

Sealed Infusion Media
Piping
Inlet helix Outlet Helix
Tubing Tubing
Peel |
Inlet p y
Infusion
Media Outlet Infusion Media
Glass
Sealant tape Tool
Outlet piping

OUT —>
Vacuum bag Polished

Tool

W‘— Resln bucket Resin trap —>

Figure 2.24 Complete VARTM setup

 

,/

T Vacuum
Pump

37

2.4.2 Infusion process

After the complete setup of the vacuuming process, the final step is to
infuse the resin through the preform. The resin and the hardener are measured
using the sensitive balance and mixed with the specified mixture ratio provided
by the manufacturer. The cup is placed under a fume hood, so that the emitting
fumes are drawn out and user is safe. Once the resin is placed under the hood,
the inlet tubing is placed in the cup so that it fully dips in the resin mixture, the
inlet clamping vise is opened and the resin flows through the tubing to the inlet
helical tubing. Once the resin reaches the outlet the inlet tubing is clamped again
so that no air is introduced into the preform. The outlet is clamped so that the
resin coming out of the outlet tubing does not reach the resin trap because once
the resin hardens it is difficult to remove it from the inlet valve of the resin trap.

The pump is left running and the specimen is left to cure overnight. (Figure 2.25)

 

 
 
 
 
 
 
  
 
 
 
  

 

—> IN
Sealed Infusion Media
Plbina
Inlet helix
Tubing '3'». Outlet Helix
g. Tubing
Inlet g Peel ply
mg?" a Outlet Infusion Media
é Glass
Sealant tape Tool
Outlet piping
OUT —>
Vacuum bag Polished
‘\ TOOI
.— Resin bucket _
Resm trap
Vacuum
Figure 2.25 Complete VARTM setup with flow Pump

38

2.4.3 De-molding process

After the specimen is cured, the pump is shut off and the vacuum bag is
removed. The peel-ply which is on top of the preform is then removed; it is easily
removed due to its nature of not sticking to the resin at all. The specimen is
removed easily from the glass mold as the surface is finished coated with wax.

This details the procedure of VARTM. The problems encountered and the
ways to overcome them, also the changes made to the VARTM process are to
be studied in the next chapter with the help of case studies and hands-on

experience with the VARTM technique.

39

CHAPTER 3

CASE STUDIES

Vacuum Assisted Resin Transfer Molding (VARTM) is a process which has
many parameters to be controlled simultaneously and it becomes quite
challenging to maintain the specimen’s repeatability. To analyze this, some of the
parameters involved in the manufacturing of specimens with VARTM were
studied. One parameter at a time was altered, keeping the rest constant, and its
effect was observed and then controlled with respect to the others. Several
changes to the setup were applied too and a final setup was arrived upon. This
chapter discusses in detail the various changes made and their effects in the
manufacturing of glass/epoxy composites with VARTM.
3.1 RESIN FLOW

The flow of resin in the preform plays a vital role in the final part
configuration and properties. In this section we would study the effects of various
fiber type, orientation and thickness of the preform on the flow of the resin. Some
amendments to the setup are also made to study their effect on the flow of resin.

The resin flow in the fabric can be approximately modeled by 1-D Darcy’s Law

Q=K*’W"’.l=g*l (1)
77 x x

 

Where,

= resin flow rate in (cm3/sec)

permeability of the fabric in (cmz)
pressure gradient between the inlet and the outlet in (Pa)

viscosity of the resin in (Centipoises)

area of the fabric (cmz)
position of the flow front in (cm)

Q
K
AP
77
A
x

40

Equation (1) shows that the flow rate Q in the fabric is directly proportional
to the permeability K of the fabric, area A and the pressure Ap gradient between
the inlet and the outlet; the equation also shows that the flow rate is inversely
proportional to the length x of the fabric explaining that as the distance from the
inlet increases the flow will be slower towards the outlet. (Appendix D)

To improve the flow rate a low permeability infusion media (INKA 7001) is
placed on top of the topmost layer of the fabric.(this is a patented process known
as SCRIMP) This improves the flow of resin in the fabric. However due to the
presence of the infusion media the flow becomes as shown in the Figure 3.1

below. The micro-gaps shown in the figure are exaggerated to show it clearly.

Infusion media

Peel ply

 

Microvoids

FLOW
Figure 3.1 Flow of resin in the fabric
The Figure 3.1 above shows how the resin flows through the fibers. Since
the flow is accelerated on the topmost layer and the resin reaches the outlet very
quickly, the fabric is not evenly wetted, giving non-uniformity to the part prepared.

To study the flow, tests were conducted. These tests were conducted with

41

various types of fiber, various orientation and thickness of fibers and will be
detailed in the following sections. Actual picture of flow mentioned in Appendix B
3.2 ORIENTATION OF FIBERS

This test was conducted to observe the flow in the various orientations of
the fibers in the preform. Preform of unidirectional fabric with a density of
0.00041 g/mm2 (12 oz/yardz) were cut in 127 mm X 127 mm pieces and stacked
in 6 layers at orientations of Mg, [1515, [30],;, [4516, [9016 ( fiber oriented in the

direction of the flow is 0°). This setup is shown in the Figure 3.2 below.
{— [015
‘ [1516

~— [3016

‘

.— [45]6 Outlet

 

Figure 3.2 Fiber orientation

42

It was observed that even with the infusion media on top there was variation in
the flow of resins in different orientations. It was observed that the fibers oriented
in [0],; took the maximum time and the one at [90]6 infused the fastest. The Figure

3.3 shows the flow of resin in the various fabric orientations.

Distance vs Time

140
120
100 .
80
60
40 <
20

+ [0 deg]
+ [15 deg]
[30 deg]
.. [45 deg]
+ [90 deg]
—Poly. ([0 deg])
—Poly. ([15 deg])
l — Poly. ([90 deg])

Distance( mm)

 

o 1000 2000 3000 4000 5000 6000 7000, —P°'Y-ll45degl)

Time( sec ) fPon. ([30 deg]) .

Figure 3.3 Flow times of various orientations of fibers

-+ 60
Distance (mm) WRT time (sec) + 120 Ii

180
2.. 240
—i«- 300
4 360
‘ —+— 420
_. 480 i
E __j 540
= 600 k
—~r—”‘ 720
840
960
1200

 

 

 

I 40
I 20
100
80
60 _
40 i
20 E

 

-.luu—i—q Tr—l
l
.
..L.
l
7‘
E
l
I
l
.

 

 

 

”ll-l

 

 

MI 4H
0 II

 

 

 

 

 

 

 

 

 

 

 

Figure 3.4 Flow lines at different time intervals for [015

43

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. —o—+ 60

Distance (mm) WRT time (sec) + 120
180
T 240

I40 ‘ , _‘ '
I20 ‘L»—A _,_j 300
100 1— ,. _._1 360
80 . _ _._ 420
firs—1“.— ._. _ : *_ 480

60 “:3, LET-t _.
40 ‘___ _ ‘-————2_:‘—__—r———‘ .. 540
20 i ”’3 —-——~— ._———- T —:3 600
0 M 720
840
' ' ' ' ' ' ‘ 960
‘ 0 20 40 60 80 100 120 I40 ’ 1200

 

 

Figure 3.5 Flow lines at different time intervals for [1515

 

 

 

 

 

 

 

 

 

 

 

 

 

60

| Distance (mm) WRT time (sec) + 120
180

l - _ . __ _ _ __, 240

140 l . ——‘— —‘ + 300
‘2“ " ’ ° ‘ + 338

100 . ‘7 , ._ . . _

so ,t;..—..:_j_—L;E_n_..- ._ _____.__ a : 480
60 -.-—____ ___,_ +_—:— - 540
20 _““'*———— - _' _2 j 720

0 840

l l v v I I I 960
]— 0 20 40 60 80 mo 120 140 1200

 

 

 

 

 

 

 

 

 

 

 

 

 

 

+ 60

Distance (mm) WRTtlme (sec) _._ 120

180

240

140 _- " ‘— + 300
120 + 360
[00 420
80 T“ 480
60 #7 540
40 ._. 600
20 720
o 840
. . . . . . . 960
0 20 40 60 80 100 120 140 1200

 

 

 

 

 

 

Figure 3.7 Flow lines at different time intervals for [45]5

44

Distance (mm) WRT Time( sec)
+ 60 sec
140 + 120 sec
120 . 180 sec
240 sec 3
+ 300 sec

—A
O
C

—e— 360 sec
—-l— 420 sec
—-——480 sec
——-- 540 sec

600 sec

0 20 4o 60 80 100 120 140 72° 59°

Width (mm) 2:: Z:

Length (mm)

 

Figure 3.8 Flow lines at different time intervals for [9015

Along with the difference in the flow times there were also variations of the
thickness in the parts manufactured. The flow lines were marked at various
intervals and were plotted for different angle orientations as shown in Figures 3.4
to 3.8. The thickness variation in the part was attributed to the uneven flow in the
preform. It could be concluded that the permeability of the fibers changed with
the angle even though the exact nature of the flow with various fiber angles was
not known. However this showed that the variation in the flow changed the part
thickness at different positions in the part.

It was also observed that the flow front was much more uniform in the
lower angled preform lay-up and became less uniform as the angle of orientation
increased. It was also observed that as the angle of orientation increased the
flow in the sides increased as compared to the center portion. Due to this, it was
also observed that as the angles increased, the thickness in the sides also

increased as compared to the center portion, this was attributed to the

45

accumulation of resin in the sides because the resin reaches the outlet faster
through the sides which left the sides resin rich. The Figures 3.9 to 3.13 below

show bag side surface and the variation in thickness with the angled fabric.

Thickness (90°)

127.00

   

50.80
25.40

0 . h
\Length (mm) Wldt (mm)
I I
Flow Direction nlet 5'“

Figure 3.9 Thickness variation at [9015

46

Thickness (45°)

    

50.80
Width (mm)
Length (mm)
Inlet side
Flow Direction

Figure 3.10 Thickness variation at [4515

Thickness (30°)

    
    

127.00

50.80
25.40

Inlet side
Flow Direction

Figure 3.11 Thickness variation at [3015

47

Thickness (15°)

   
 

50.80

0 W'd h
Length (mm) I t (mm)

Inlet side

Flow
Direction

Figure 3.12 Thickness variation at [1515

Thickness (0°)

   
 
  
  

Outlet side 0150 r”

0.100
Thickness (mm [1 .050 s
127.00

50.80
25.40

0

\ Length (mm) I ( )
Flow Direction n at Slde

Figure 3.13 Thickness variation at [015

48

The figures show that the variation in thickness is more in the higher
orientation angle. One other observation made was that all the specimens had
resin accumulated near the inlet and thus the thickness was more at the inlet
side of the specimen. This was due to the loss in the pressure gradient between
the inlet and the outlet and due to that the resin coming from the inlet tubing
accumulated and made a reservoir near the inlet. This caused the inlet side of
the specimen to be thicker as compared to the other sides.

In order to observe the actual flow of resin through the preform it was
necessary to remove the infusion media so that the flow front was not strongly
guided by the infusion media. The same setup was used but this time without

infusion media as shown in Figure 3.14.

[0°]

*-— [15°]

‘—- [30°]

.__ [45°]

.— [90°]

 

Figure 3.14 Setup of fabric at various angles

49

Similar trends were observed as far as the flow was concerned; however, the
thickness variations were lesser. This may be attributed to the fact that the flow
was not accelerated and the variations were minimized, however the trends of
the flow were similar to those with the infusion media, yet the thickness variation
was less. It was also observed that the unidirectional fiber was difficult to infuse,
the reason for which would be discussed later in the chapter. To understand the
effect of infusion media another test was conducted, and discussed in the next

secfion.

3.3 INFUSION MEDIA

As it was observed that the removal of the infusion media form the top of the
topmost layer of the fabric slowed down the process and gave less variation with
the thickness of the part produced. To address the effect of infusion media in the
flow of resin through the preform, varying lengths of infusion media were placed
and the flow was observed with unidirectional 90° fabrics (as that was the fastest
among the lot in the previous test). The setup changed is shown in the following
figures. The infusion media placed to cover the entire preform in Figure 3.15b,
was removed from the preform and placed half way through the preform on the
outlet side as shown in Figure 3.15c. This test was done under the same
conditions as the previous test. From this test a very important observation was

made and is discussed later in this section.

50

Vacuum bag

Inlet -
Infusion / Fabric
Media

 
 

 

 

 

 

 

 

 

 

 

 

 

 

\ Outlet Helix
Sealant tape a: j; ' .. Tubing
if” ' m E ' ’ ' -4... 4 Libe Glass
H Tool
33" "‘m 38.1mm
‘ ‘
INLET OUTLET
Figure 3.15a Setup with No Infusion media
Vacuum bag
I f "I‘M Infusion Media Peel ply
n us on Fabric
Media / / Outlet Helix
Sealant tape Tubing
Glass
Tool
INLET

Figure 3.15b Setup with Infusion media

51

Infusion Media Vacuum bag
Peel ply

   
 
  

Inlet Infusion

Fabric
9“" Outlet Hellx
T
Sealant tape ublng

' ‘ ' ' Glass
fl Tool
38.1 mm 38 1 mm
‘ ’
INLET OUTLET

Figure 3.15c Setup with Half way Infusion media

 

With complete With Half-Way With no Infusion
Infusion media Infusion media media

INLET

Figure 3.15d Combined picture of all the three setups

52

The Figures 3.15a to 3.15d above show the three setups with varying lengths
of infusion media placed on top of the fabric. It was observed that the resin in the
setup with the infusion media was the fastest (319 sec), followed by the one with
half- way infusion media (489 sec) and the one without the infusion media was
the slowest (546 sec). All the specimens were the same dimensions, 127 mm X
127 mm (5 in X 5 in). The main purpose of the infusion media was to accelerate
the process of infusion through the fibers.

Due to the low permeability of the preform, the resin flowed very quickly (213
sec) on the top Side of the fibers with the infusion media as shown in the Figure
3.1. Also because the infusion media was significantly thicker than the peel ply,
the vacuum bag did not conform completely to the media leaving gaps which
assisted in the flow of resin. However, in the case where there was no infusion
media on top the bag very perfectly contoured the laid down fabric and the
permeability of the top layer was reduced as compared to the case with the
infusion media. Due to this the permeability of the fiber was the resisting force in
the setup. No acceleration was provided the flow time on both tool and top side
was (546 sec). In the third case it was also observed that the resin moved in a
Similar trend as the setup with no infusion media but accelerated when it reached
the distribution media placed half way through the fabric. The Figure 3.16 shows

the progress of the flow in all the three cases.

53

 

Figure 3.16 Flow of resin through the three setups

Another very important observation was made in this study. Due to the
smoothness of the vacuum bag in the case where no infusion media was used,
the top layer was smoother as compared to the one with infusion media and peel
ply on top (the roughness in this case was due to the peel ply as it was
perforated in nature and gave corrugations to the top surface profile).

With smooth finish, the composite specimens were visibly clearer than the
ones with the corrugated finish. In the case where only half-way infusion media
was used one half was clear and the other half (with infusion on top) was
unclear. This also confirmed that the presence of peel ply was the main cause of
the clarity of the specimens. Even though the transparency and clarity highly
depended on the degree of matching of the refractive index of both the resin and
the fiber, surface finish played a vital role in the clarity of the specimen. Because
the optical study of the composite was outside the scope of this thesis it would
not be discussed further in great detail.

Another very important observation was made that after the fiber had been
infused completely, the air bubbles could not retract back into the fabric. Also, the
extent of retraction was greatest in the setup with infusion media followed by the

setup with half-way infusion media and then the setup with no infusion media.

54

This extent of bubbles was negated to a great extent with more resistance as in

the case of the setup with half way infusion media.

 

Figure 3.17 The air bubbles retracting back In the setup with no Infusion media

At first this was also attributed to the type of fabric as different fabrics had
different permeability and resistance to this back flow of residing bubbles. The
void percentage was as high as 8% in the setup with infusion media. Another
study was conducted due to this behavior which will be discussed in the next

section.

55

3.4 TYPE OF FABRIC

This test was conducted to see the effect of types of glass fabric on the air
bubble retracing back. The glass fibers used were unidirectional (density 0.00041
g/mm2 (12 oz/yard2)), fine woven 20 fabric (density 0.000406 g/mm2 (12
oz/yardz), coarse woven 2D (density 0.000813 g/mm2 (24 oz/yard2)) and quasi
3D intenrvoven fabric. No distribution media was placed on top (only on the inlet

and outlet) (Figure 3.15a).

    

_ Woven tine Unidirectional Woven Quasi 30
HOW ""35 stitched coarse
Figure 3.18 Types of fabric

It was observed that no matter what the type of fabric was, the bubbles always
retraced back into the fiber after infusion, due to the overlapping distribution
media at the outlet. In other words, it was due to the decrease in gradient
compaction pressure with the increasing distance from the inlet of resin. The
extent of retracted bubbles was most in the setup with the infusion media as the
infusion media assisted the flow of bubbles due to its high permeability. This was
attributed to the fact that since the infusion media placed at the outlet had a
considerable amount of thickness the vacuum bag did not exactly conform to its
shape thus creating channels for it to flow back. To be able to stop these air

bubbles to retract and still maintain the flow through the fiber was a challenge.

56

The infusion media was removed from the top and the outlet side .The most
suitable material within the scope of VARTM was the peel ply which acted as a
fine distribution media, let the vacuum bag take its shape, and did not let the air
bubbles, already trapped in it, to retrace back. It acted as a natural throttle,
trapping the displaced air bubbles and preventing them form retracting
backwards into the fibers. The setup is shown in the Figure 3.19 a), b) and c). A
new name was suggested for this setup “BUBBLE TRAP VARTM”.

This setup was used for all the experiments discussed further in this thesis.
The fact that the peel ply acted as a natural throttle decreased the visible air
bubbles in the specimens made. Testing for the void volume fraction will be

discussed in later chapters.

  

Inlet piping
Sealed I in
pp 9 Fabric
Iniet Helix
Tubing Outlet Hellx
k Tubing
Inlet side
Infusion Peel ply
Media
Glass
Sealant Tool
tape
Outlet I In
Polished Tool 9 p 9
‘
INLET OUTLET

Figure 3.19a Bubble trap setup (top view)

57

Vacuum bag Bubble

 

 

 

 

 

 

     
 

 

 

 

 

I 1 ti f ' / mp
n e n uslon
, Fabric
Medla / / ,/ Outlet Helix
\\ I ’ / . Tubing
Sealant tape - / /
¥i727, E -_ L 1.. A—‘L‘ / ) , A Glass
38.1 mm 38.1 mm
’ ’
INLET OUTLET
Figure 3.1% Bubble trap setup (side view)
Bubble
Trap
Peel ply
\~ r—
m. 0 0 I H I
ut et eix
O / Tubing
m 0
fl 0
O
O
/'
/ fl 0
Fabric /’
Resin FIOW Outlet piping

OUTLET

Figure 3.19c Bubble trap setup

5

OO

Since only the vacuum bag was on top of the fiber in this technique, it gave a
fine finish to the upper surface at a microscopic level but it took the Shape of the
fabric at a macroscopic level, thus the overall surface finish was not as good as
was expected. Due to direct contact of the vacuum bag with the resin, the bag
stuck to the fabric, while demolding. To remove this defect, the bag was coated
with liquid wax (Locktite) with a cloth. It was found that the bag came off easily
after the specimen gelled completely. The need to have a smoother surface on
the bag side was a challenge. Silicone rubber was thought of as a better
alternative as it was flexible and had a smooth surface finish and a tendency of
not sticking to the resin. The tests done with silicone bagging will be discussed in

the following section.

3.5 SILICONE RUBBER SHEET

Silicone rubber was made in the laboratory to get a smooth finish on one
side of the bag so that it could be used as the second finished side in the VARTM
process. The process of making the silicone rubber was very simple and is
discussed briefly. A glass plate was laid on the table and was waxed using ICE
WAX manufactured by the Gentex Company. After the wax dried, it was cleaned
with the help of paper towels to give the glass a smooth finish so that the rubber
placed on it could be easily removed after it dried. The silicone rubber came in
gel form in containers with nozzles and was dispensed using a dispenser gun.
The silicone was dispensed using the gun and was carefully spread with the help

of a brush. When the first layer had dried the next layer was placed on top of that

59

 

Reinforcing
net

     

Applying the
Silicone gel Glass
mold

First Layer

Figure 3.20 Making the silicone bag

layer in the same manner. A net like reinforcing material was placed when the
third layer was still wet. This reinforcing net gave the strength to the rubber. After
the layer of reinforcing material three more layers of the silicone gel were applied
and the setup was left to dry. After all layers had dried, the silicone sheet was
carefully taken off the glass mold and it had a glass like finish on the glass side.

Figure 3.20 shows the entire process of silicone making the silicone sheet.

60

Once the silicone sheet was manufactured, it was used as the second
flexible mold under the vacuum bag. The rubber was cut into pieces exactly the
length of the specimen to be produced and placed on top of the fabric, the

vacuum bag was placed thereafter and sealed. The setup is shown in the Figure

   
  
 

    

 

 

 

 

 

 

 

3.21 below.
Silicone rubber Vacuum bag Bubble
Sheet Trap
Inlet Infusion F bri
Media _/ 3 °
1 Outlet Helix
Sealant tape J .2- a / Tubing
M, , e E... . —
hm ..--.. .r .2. _ Glass
H Tool
38.1 mm 3&1 mm
‘ ’
INLET OUTLET

Figure 3.21 Silicone rubber setup

A few interesting observations were made using the setup with the silicone
rubber sheet. It was observed that the surface towards the silicone sheet side
had glass like finish; however, there was a dry spot which was created. The dry
spot was a region in the composite material where the resin had failed to
impregnate the fabric. The reason for the formation of this dry spot was due to
the race tracking effect.

Race tracking developed because the silicone rubber sheet was thick (6 mm
(0.236 in)). And the vacuum bag on top did not conform enough to it, creating a

low-resistant channel for the resin and thereby a dry spot in the specimen. Figure

61

3.22 below shows the flow pattern of the resin in this setup. Figure 3.23 shows

how the contouring of the vacuum bag fails with the silicone sheet on top.

Inlet \
Pinlna Fabric
Inlet
Helix
Tubing
. Outlet Hellx
Inlet Tubing
Infusion P I I
Medla ee PV
Sealant tape Glass
Tool

 

Void formation Flow front Outlet
Sealed
’
INLET OUTLET

Figure 3.22 Formation of dry spots with the silicone sheet setup

Vacuum bag

 
 
 

Silicone rubber
Sheet

 

 

 

Racetracklng Glass
gap Tool Vacuum bag
J ‘ Silicone
F b - Rubber
3 "° Sheet
I
mgr. _‘.:.."'if..—.;.. t .. .-
/ Glass
No Racetracklng Tool
9813

Figure 3.23 Cross-sectional view of the setup with /without gap

62

A thinner sheet of rubber was thought to be an immediate answer to this
problem; however a thinner sheet would have less stiffness and would conform
to the fabric more readily, failing the purpose of the silicone sheet to obtain

surface finish. Figure 3.24 shows the setup and the manufactured part.

Dry
spots

 

Figure 3.24 Setup with the silicone rubber sheet

Another very important observation was made during this experiment which
was that the part produced was thicker (4.7mm) than the regular set-up
specimen (4.32 mm). The reason of this will be explained in the next section. The
silicone rubber sheet was however the best answer to the two sided finished
specimens and it could, because of its flexible nature could be utilized for 3D
shapes like domes and arches. . Test with and without the silicone rubber on top
of the fabric were done and seen that the thickness of the part with silicon rubber
increased by 17.5% as compared to the one without the silicone rubber. The
surface finish was obviously better in the specimen with the silicone rubber on
top. Since the thickness was difficult to maintain and there was formation of dry

spots with the use of silicone rubber this technique was discarded. This

63

technique, however, was feasible where the geometry of the specimen was
complex. The silicone rubber being flexible could conform to any desired Shape.
The only challenge was to reduce its overall thickness to a degree where it would
give fine finish, producing no race tracking channels and yet not take the exact

shape of the fabric.

3.6 COMPACTION

Compaction plays an important role in the manufacturing of composites with
VARTM. Not only does it affect the resin flow in the specimen, but directly affects
the fiber volume fraction and void volume fraction in the part.

Grimsley, Cano and Loos studied the effects of compaction [4]. They
concluded that as the stack thickness of the fabric increased the compaction
pressure increased and the overall fiber volume fraction increased; this could be
attributed to the increase in the amount of nesting which also meant the fabric
was much more compact and had less air gaps already. This also increased the
flow rate of the resin through the fabric in the through thickness (discussed in
detail in the next section). Also, drawing the vacuum pressure for long hours (5
hours) increased the compaction and thereby increased the fiber volume fraction
in the specimen

Another test was conducted in which the vacuum pressure drawn was left for
varying time duration ranging from 10 minutes to 5 hours. It was observed that
the fiber volume fraction increased by 11.3% from the first (10 min.) to the last

case (5 hours). The effect of leaving the vacuum pump on for longer time also

64

had an effect on the resin flow time and flow front. The Table 3.1 below shows
the various infusion times at different compaction time.

Table 3.1 Flow times at different compaction times

 

 

 

 

 

 

 

 

Batch Dimensions No. of Layers Compaction Infusion time
- 1 6” x 6” 5 tlmeo min 6 min 18 sec

2 6” x 6” 5 15 min 6 min 0 sec
3 6” x 6” 5 1 hr 5 min 52 sec
4 6” x 6” 5 2 hr 5 min 37 sec
5 6” x 6” 5 3.5 hr 5 min 21 sec
6 6” x 6" 5 5 hr 5 min 20 sec

 

 

 

 

 

3.7 SPECIMEN DIMENSIONS

It was also imperative that this technique was feasible for all sizes of parts
manufactured. VARTM is a technique which almost entirely depends on the
vacuum pressure, the permeability and compaction of the fibers, it was important
to determine the amount of area the resin will cover with all these restraints.
Tests with different dimensions of length, width and thickness were done. Fabric
was laid up with dimensions ranging from 127 mm2 (5 in2) specimens to 508 mm
x 203.2 mm (20 in x 8 in) specimens. The positioning of the inlet and outlet tubing
was also changed. The stack lay-up of [0/90] 2D interwoven fabric (density of

0.000813 g/mm2 (24 oz/yard2)) was increased from 3 to 22 layers.

65

 

OUT

   

OUT

 

Short-wide

Figure 3.25 Setup with different dimensions

Various observations were made. The parts with smaller dimension infused
totally, the parts with longer distance between the inlets an outlet did not infuse
completely and due to loss in the pressure gradient towards the outlet the
bubbles trapped in the portion of the fabric not infused, retracted back due to
pressure gradient loss and damaged the parts. The parts that had the fabric laid
up in horizontal direction (shorter length between inlet and outlet, yet larger
widths) infused completely. Figure 3.25 shows the long narrow and short wide

setup.

66

It was also observed that as the stack lay-up of the fiber increased (3 layers
to 8 layers), the flow was accelerated in the layers where there was no infusion
media. Also the through-thickness flow accelerated as the thickness increased.
The flow front also reversed its direction from moving faster in the top (bag side)

to moving faster in the bottom (tool side) as shown in Figure 3.26.

 

 

 

 

.‘/

Ill

FLOW
Figure 3.26 Schematic of flow front

It was also concluded that even though the resin moved large distances with
this technique it was limited because of the loss of pressure gradient and
retarding of the resin flow towards the outlet, shown in Appendix B.

Figure 3.27 shows the various fabrics stacking sequence for 2D woven glass
fabric. It shows how with the increase in the fabric thickness on the flow changed
(becomes faster in thick specimens). It was observed that the fiber volume
fraction increased with the increase in the fabric stacks as the void volume
decreased significantly. It was observed that with the increase in the thickness of
the fabric layup, the thickness variation at different points in the manufactured

parts decreased from 8 % to 4.5 %.

67

180
160 -
140

120 ~

100 . '/’:_.4

‘ II
80‘ ,r

+ 22 layers
‘ + 3 layers
5 layers
3* 8 layers
+ 12 layers
+ 14 layers
-+-O3D-6layers without infusion media ' .
—-QBD-6 layers with infusion media

60*

Distance covered ( mm)

40-

 

20-

0 1 00 200 300 400 500 600 700
Time( sec)

Figure 3.27 Flow times with varying stacking sequence

It was concluded that the flow of resin was accelerated in thick stacks of
fabric. The resin would travel a greater distance and would cover a large area of
the fabric and thus larger parts could be manufactured with the same technique if
the stacks were thick enough. Complex shapes could also be manufactured with
this technique; however that area has still to be studied. Since the vacuum bag
took the shape of the fabric, smooth finish on both sides of the specimen was a

challenge and was to be achieved within the scope of VARTM. Maintaining

68

constant thickness was also another aspect which would make the sample
consistent in volume fraction.

One of the reasons the voids were formed in the fabric was the preexisting
air bubbles in the resin due to mixing. This caused the air bubbles to travel
through the fabric and increase void volume fraction. This formation of air
bubbles was natural and could be only eliminated before the infusion took place.
A vacuum chamber was used and the pre-mixed resin pot was placed in the
chamber. Vacuum was drawn in the chamber. It was observed that the bubbles
rose up to the top of the pot and collected there, reducing the number of air
bubbles in the pot and thus reducing the void formation (6.34 % to 4.79 %) in the
fabric (Figure 3.28). Care was taken that the resin was not “degassed” for large
duration as this would affect the pot life of the resin, which would affect the
infusion time through the fabric. This might also cause premature gelling of the

resin and the manufactured parts may be defective and not give desired results.

 

 

 

 

Bubbles drawn to the top of the pot

Figure 3.28 Degassing of the resin with the vacuum chamber

3.8 FINAL SETUP
To achieve the desired results of constant thickness, low void volume fraction
(high fiber volume fraction), smooth finish and transparency, the setup was

suggested as given in Figures 3.29a and 3.2%.

 

  
 
 
     
  

Sealed Plastlc Plate

Pioinu
Fabric
Inlet helix

Tubing
Inlet
Infusion
Media

Sealant tape

(54".:- "lid—am f}.

Polished
Tool

Vacuum bag

 

‘— Resin bucket
Resin trap —*

T Vacuum
Pump

Figure 3.29a The final setup (top view)

70

Plastic Vacuum bag
Plate '

Inlet Infusion / Fabric
Media ] Outlet Helix

Tubing

  
  

 

 

Sealant tape

 

 

Glass
Tool

 

 

38.1
38.1
mm 'mm
‘ ‘
INLET OUTLET

Figure 3.29b Final setup (side view)

The use of a thin, glass finished, stiff yet flexible plastic sheet assisted in
achieving the desired result. The reason to use a thin sheet was to take into
account the race tracking phenomenon which was the cause of dry spots.

Race tracking is a phenomenon where the infusion resin follows the easiest
path (path of low resistance) either through the side of the setup or the fabric
where there is a gap to pass.

The vacuum bag took the shape of the thin sheet in the sides and prevented
the resin from moving through the sides. The flat plate helped in giving the
specimen a fine smooth finish which in turn helped in improving the transparency
of the specimen. The transparency also depended on the effective matching of
the refractive indices of the resin and the reinforcing fabric. The peel ply at the

outlet helped trapping the bubbles. The thickness was controlled by using a dam

71

(Figure 3.30) of desired thickness and placing dead weight on top of the plate

sitting on the darn.

   

Fabric

 

Glass tool

Figure 3.30 Final setup

3.9 CHANGES TO THE FINAL SETUP

It was desired to have a uniform faster flow of resin, constant thickness,
transparency and two sided glass finished surface in the composite. In the final
setup all except the uniform faster flow was achieved. To address this issue
some amendments were made in the setup. A thin perforated plastic film (to give
the fine finish) known by trade name E2760 release film, manufactured by
Richmond Aircraft, and peel ply (to accelerate the resin flow) were introduced

between the fabric and the plastic plate. This setup is shown in Figure 3.31.

72

Peel ply

  
 

_ Vacuum bag
Re'ease PIaStIC
Film Plate / Fabric
/
Inlet Infusion Media / Ougiet Hellx
:5 n
Sealant tape u 9
Glass
Tool

 

38.1
8 1
mm ‘mm
INLET OUTLET

Figure 3.31 Final setup (side view)

The thin perforated plastic sheet had a property of not sticking to the resin
once the resin dried; also since it was perforated it assisted the flow. Due to the
fine finish of this thin sheet it also gave a fine finish to the sample unlike the case
where peel ply was placed right on top of the fabric. Table 3.2 below compares
the flow times with the various setups.

Table 3.2 Chart showing the flow time using the various setups

sheet, Top
plastic sheet and

no
media and no

 

73

The peel ply assisted the flow of resin through the fabric. The uniform
thickness was taken care of by the plastic plate. This setup gave the desired
faster resin flow, finished surfaces, transparent specimens and uniform thickness
to the specimens.

The fiber volume fraction achieved by this technique was very high (59 %)
and almost as close to that produced by prepreg (63%). Another observation
made was that the gap between the top fiber layer and the thin plate acted as a
distribution media and the flow front accelerated the same way as it did on the
tooling side (Appendix B). This effect was taken to advantage and large panels
were made which infused completely in less time. This eliminated the use of
distribution media and thereby cutting the cost. The simple setup could be used
for complex geometry also. This technique had the advantages of manufacturing
two Sided glass finished parts; it also reduced the cost because the raw material
used was reduced. Another major advantage was the reduction of voids in the
part and thereby achieving a high fiber volume fraction which was close to those
achieved by prepreg. High fiber volume fraction could also be achieved by
making the dam slightly less high as the desired thickness which, after the
application of dead weight, would smother out the extra resin in the sample
thereby giving high fiber volume fraction in the manufactured parts. The
technique was user friendly and could be done at room temperature. The desired
thickness could be achieved by forming dams on the outside of the vacuum bag
and putting dead weight on. Finally it was a clean process and could be taken to

advantage make large parts.

74

 

3.10 SUMMARY OF CASE STUDIES

The various tests studied in this chapter are summarized in this section.
The tests conducted were in this order; first the different fiber orientation was
studied to study the effect of the varying angles in the flow of resin through the
fabric. Flow lines were marked and flow rates were calculated. It was observed
that the fabric oriented at 90° infused the fastest and the one at 0° moved the
slowest

Next the effect of infusion media on the flow was studied by varying the
area of the infusion media. The fiber type was also studied by setting different
types of fiber and monitoring the flow of resin in these fibers.

Effect of fabric compaction was studied by varying the compaction time. The
study for the effects of various fabric dimensions was another part of the
observation.

Various setup changes were made such as use of silicone rubber and plastic
plate to study the various flow patterns. With the help of all the studies conducted
and parameters controlled, a final setup was reached at, which assisted in the
arrival at the final goal, developing batch to batch repeatability with the VARTM

technique

75

 

CHAPTER 4
TESTING METHODS AND RESULTS

This section details the testing methods used in this study. It also
recommends the various precautions and standards to be maintained during the
testing of the specimens
4.1 FLOW FRONT MONITORING

Primarily, the flow rate of resin was observed by using a highly sensitive
stopwatch which could measure up to 1 millisecond. The stopwatch was
activated as the resin flow front hit the fabric after passing through the inlet side
infusion medium. Since the tool on which the samples were fabricated was made
of glass and was transparent, it was possible to observe the flow front of the
resin from both top and bottom sides with the use of cameras. On both the sides
cameras were fixed and the footage was analyzed at various time intervals. Two
cameras were used only for a few tests. In this manner, the variation of the flow
front on the top and bottom could be calculated. The Figure 4.1 below depicts the
setup of the cameras.

Top camera to monitor
'13 top side flow front

Glass mold

    
 

Frame

_ . Bottom camera to monitor
Setup 0' fabrlc . 1 ._- bottom side flow front

Figure 4.1 Top and bottom cameras for flow front monitoring

76

4.2 BURN OFF TEST

This test was used to calculate the void percentage and the amount of fiber
volume fraction in the specimens manufactured by the VARTM process. In this
test, first a small portion of each sample was cut and labeled. The samples were
then placed in the oven to dry off any moisture which might be present. After this,
a specially made balance was used to weigh the samples both in air and water.
In the present study, the balance used was Sartorious BP 221 S. The measuring
scale measured accurately up to 0.1 mg changes. This setup used a base
measurement frame. This frame had a basket on top which was used to measure
the dry and wet mass of the sample. A bridge was used to hold the water cup
such that the bridge and the water cup do not directly contact the measuring
scale. All of this setup was enclosed in a glass cubicle. The scale was first
zeroed and the glass cubicle was closed as the scale was sensitive to the
movement of air. Once the scale Stabilized and there was no fluctuation in the
reading, the sample was placed on top of the basket and the glass cubicle was
closed again. This was the dry mass of the sample and was noted down. After
this, the cup was filled with distilled water to a height of 3-4 cm (usually filled
enough to immerse the sample completely). The basket on top was also
connected to another basket with a thin wire (with negligible weight). This basket
immersed in the water and the sample was carefully placed in it so that there was
no water splash which might cause an error in the measurement. The sample
was again weighed and this was the wet mass of the specimen.

Once the dry and wet mass of the composite specimen was measured the

specimen was taken under a fume hood and a propane torch was used to “burn

77

off" the matrix from the composite. The composite specimen burned with a
continuous flame till the time there was no matrix in the specimen. Once the
entire matrix was burnt out, the continuous flame stopped forming thus we knew
that the entire matrix was burned out. Once the matrix was completely burnt out
the specimen was left to cool down. After it cooled the fiber-only specimens were
again weighed in the same manner as mentioned above. After noting down the
dry and wet mass of the burnt specimens, calculations of volumes, density and
void content could be determined. The calculations are mentioned later in the

section. Figure 4.2 below shows the setup of the test.

Specimen _
Dry » Frame
Pan

Specimen

Water
Bucket

Bfidge

   

(a) Dry mass of un-burnt sample (b) Wet mass of un-burnt sample

 

(c) Burning of matrix (d) No end flame (burnt out matrix)

Figure 4.2 Burn off testing

78

Figure 4.2 Burn off testing continued

 

(e) Wet mass of burnout sample

The following notations are used in this section.

V = Volume of Composite (cm3) pa = Composite Density (g/cmz)

V; = Volume of Fibers (cma) ,0): Fiber Density (g/cma)

Vm = Volume of Matrix (cm3) pm: Matrix Density (g/cmz)

V, = Volume of Voids (cm3) pwater= Water Density (g/cmz)

me = Mass of Composite (9) mf = Mass of Fibers (9)

mm = Mass of Matrix (9) mad/y) = Dry mass of Composite (9)

meme.) = Wet mass of Composite (9) mm,“ = Dry mass of Fibers (g)

mnwe, ) = Wet mass of Fibers (9)
4.2.1 Composite Density and Volume calculation

The Archimedes’ principle of buoyancy was used to calculate the volume
and density of the composite. We needed the density of water and the dry and
wet mass of the composite. The equations below show this relation and thus the

values were determined.

m —m
C(drv) C(wet)
C = —'—————— (4.2.2.1)
pwarer

79

: "1(‘(dr)‘) (4 2 2 2)
—-———VC . . .

(-
422 Fiber Density and Volume Calculation

As detailed above, after the complete burning of the matrix, only the fibers
remained. This burning off of the matrix left some ashes and residue which were
removed by rinsing and drying, otherwise the residue would hinder the accurate
calculation of the fiber mass. The fiber volume and density were also calculated
in the similar manner with the Archimedes’ principle of buoyancy as the
composite volume and density were calculated. The equations below were used
for the calculation.

= mf(dry) _mf(wet)

thlté’r

 

Vf (4.2.2.3)

: mf(dry)

Pf Vf

(4.2.2.4)

4.2.3 Matrix Volume Calculation
By knowing the mass of the fibers and the, composite the mass of the matrix
could be calculated using the formula given in Equation 4.2.2.5 below

mm = ’"c -mf (4.2.2.5)

Once the mass of the matrix was determined, the volume of the matrix could be
determined either by noting the value given by the manufacturer or by curing a

small piece of the matrix and using Equations (4.2.2.1) and (4.2.2.2).

Vm :fl. (4.2.2.6)
Pm

80

4.2.4 Void volume and percentage calculation

After determining the volume of the composite, matrix and fibers, the volume
of the voids could be easily determined by the relation given in equation (4.2.2.7)
below.

VV :VC—Vm —Vf (4.2.2.7)

Thereafter the void percentage in the composite could be calculated using

Equation (4.2.2.8) shown below.

V0id% = ix100
C

(4.2.2.8)

4.3 EDGE REPLICATION ON MICROFILM

This test was used to observe the thickness quality, the void distribution and
the Side profile of the composite specimen. A specimen was cut with a fine
cutting tool so that the edge was straight and could be polished. Once the edge
was cut clean, it was washed to wipe of any residue remaining after the cutting
process. The edge to be replicated was then polished with polishing powders on
a polishing cloth. The polishing powders used in this research were aluminum
oxide powders manufactured by the Buehler shown in the Figure 4.3 below.
Primarily the edge was polished by putting a 5/1 powder on the cloth and some
amount of water was poured to make it like a slurry paste and then polished.
After polishing, the edge was observed to feel the roughness. It was cleaned with
water and then again polished until the edge was smooth. The next step was to

polish it with a 3p polishing powder to give a better finish.

81

 

Thereafter a small amount of acetone was taken in a clinical syringe and
spread in small quantity on the polished edge. Then quickly the edge was placed
on the microfilm and pressure was applied and image of the edge was replicated
on the microfilm. The film was then observed under the microfilm projector. It was

digitally downloaded into the computer.

 

Figure 4.3 3;; and 5p Polishing powders

4.4 IMPACT TESTING

The sole purpose of doing the impact test was to check the consistency in the
properties determined from the test, such as impact energy and stiffness. To test
the impact energy and stiffness of the composites, a DYNATUP 8250 drop
weight testing machine was used. The schematic of the drop-weight impact
testing facility is shown in the Figure 4.4. The setup consisted of a drop weight
with the tup fixed at the end and was guided by two guiding rails. Various
thicknesses of samples were to be measured and thus they were to be clamped

very tightly in a specially made specimen holder. There were many

82

characteristics in the resultant graph but the tests were conducted to measure

only a few properties, which would be mentioned in this section.

Height adjustment clamp

      
  
 

Latch

Crosshead

 

 

Load Cell

Impactor

Guide rail __,

Roller Lever
switch

4.. Toggle switch

Pneumatic
Rebound
Attestor L—

Specimen

 

 

Figure 4.4 Impact loading machine setup
4.5 LOAD-DEFLECTION CURVE
Figure 4.5 below shows the desired load-deflection curve which gave us
information about the two properties which were taken into consideration in this
context, namely the absorbed energy and stiffness. The area under the load

deflection curve was measured to be the energy absorbed by the specimen.

83

However, it could be observed that the last portion of the graph gave a noisy
data. This was not considered while taking the energy absorbed as this was
nothing but the data obtained when the impact tup had perforated the specimen
and it retracted back, the friction due to the tup and the specimen caused the last
part in the graph. This portion was truncated by using a trend line to fit the end
part of the curve so that it reached the zero of the curve shown in Figure 4.5. The

data points after the fit line were eliminated form the analysis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Peak Load

18 - Descending

16-- Ascending — - slope —'

’4" °'°°° f K \\ ‘—
,. S / f \ \ Truncated —
g 8 / 2” Absorbed \ Data _‘
w 5 7/ / Energy —\
g 4. \ ‘ A A A
3 2. Maximum Deflection \‘ng \/ V

0 ' I I I T I I

n 7 4 6 R In l7. l4
Deflection (mm)

 

 

Figure 4.5 Load -deflection curve

There were two sections in ascending portion of the plot. The first Slope
was indicative of the “stiffness” of the specimen. The area under the curve which
was the absorbed impact energy was determined by the integration of the force
values form the zero to the maximum deflection shown in Figure 4.5. The highest

point in the load-deflection curve was the peak load which was the maximum

84

 

load felt by the impactor. The final point where the new trend line touched the
deflection scale was called the “maximum deflection”.

As the samples manufactured in this study were transparent to a degree, the
area of damage could be observed easily with the naked eye or by the
backlighting method. The damage area in each layer could be observed but
distinction could not be made as to which layer had a particular damage area.
The study of damaged area was out of the scope of this study so that has not
been addressed here. Figure 4.6 shown below depicts the impacted specimens

with the damage areas.

 

   
  

Ir”

Primary
Damage
Area

Secondary
Damage
Area

 

Figure 4.6 Backlit impacted specimen showing primary and secondary
damage area

4.6 SUMMARY OF TESTING

Various tests were conducted during the manufacturing of the composites
with the VARTM techniques. The flow of resin was tested with the various setups
by varying the numerous parameters involved, their effect was seen and
optimum conditions were generated. These tests involved using fibers oriented at
various angles, fabric stacked in various thicknesses, the use of infusion media,

and the use of silicone Sheet, plastic sheet and perforated plastic film. The tests
85

were also done by using various types of fabric to monitor the flow in each of the
samples.

The void volume fraction and the fiber volume fractions were calculated with
the help of the “Burn Off” test which is discussed later in the chapter. Another
method incorporated was the “edge replication on a microfilm” of the samples to
see the thickness, surface finish and to monitor the presence of voids; also the
surface finish was recorded to measure the smoothness of the samples
manufactured by VARTM.

Dynamic impact loading was also done to see the consistency in the
properties of the samples produced by the technique. The transparency was
used to determine the damaged areas after the impacted specimens were placed
on a light table and the impact area was observed.

The various testing methods were detailed in this section. The various

observations made by these tests were also presented.

86

 

4.7 RESULTS AND DISCUSSIONS
As the objective in this study was to obtain consistent properties of

Specimens made by VARTM batch after batch, this section details the results
observed of the various tests done and the cause and effects are discussed.
Some important deductions are also made.
4.7.1 RESIN FLOW

The flow of resin through the fabric is one of the governing factors which
affect the properties of the Specimen made by VARTM. As mentioned in previous
sections various setups were used and their flow times were measured. The
objective in this study was to obtain consistent flow of resin and other properties
batch after batch.
4.7.1.1 FABRIC AT VARIOUS ANGLES

The fabric used in this test was unidirectional stitched glass fiber, stacked up
in 5 layers with angles varying from 0° to 90°, 0° being the fibers in the direction
of the resin flow. The graphs (shown in the Figure 3.3) indicate that the flow of
resin in the fibers oriented in the largest angles (90°) is the fastest. This was
attributed to the nesting (fibers cramped closely) [4] of the fibers as the angular
variation within the weave changed. The channels through which the resin
moved were highly diminished as the angle decreased. As a result, the flow of
resin was uneven and void pockets were formed. Also, as the angle of the weave
increased, the flow of resin became more uniform but started to accumulate in
the sides due to less resistance in the flow or resin in the specimen creating resin

rich areas in the sides. Due to the uneven, resin flow the properties such as void

87

fraction and thickness showed large variation. Since the study of the small angle
is out of the scope of this study only results would be presented.

The Tables 4.1 and 4.2 below show the thickness, density and the void
percentage in the sample. The thickness mentioned is the average thickness of
the sample.

Table 4.1 Specimens without infusion media

 

 

 

 

 

 

 

Specimen Orientation Infusion Infusion Thickness(mm) Density Void
no media time (g/cm“) (%)
(sec)
1 (a) [015 No 5760 1.33 1 .83 4.71
1 (b) [1515 No 3120 1.41 1.82 4.91
1 (c) [30].. No 2640 1 .66 1.80 4.98
1 (d) [4516 No 1740 1 .56 1 .79 5.23
1 (e) [9015 No 840 1.87 1.78 5.43

 

 

 

 

 

 

 

Table 4.2 Specimens with infusion media

 

 

 

 

 

 

 

Specimen Orientation Infusion Infusion Thickness(mm) Densi Void
no media time (g/ cm ) (%)
(see)
2(a) [0].; Yes 1380 1 .98 1 .81 4.78
2(b) [1516 Yes 1083 2.01 1.80 4.93
2(0) [3015 Yes 975 2.01 1 .79 5.09
2(d) [4516 Yes 827 2.12 1 .79 5.33
2(e) [9016 Yes 540 2.34 1 .76 5.57

 

 

 

 

 

 

 

By looking at the tables, it was observed that the infusion time was largest in
the weave specimen with the minimum angled orientation both in the case with

and without infusion media. It was also observed that the thickness increased,

88

 

 

 

l

density decreased and void fraction increased with the increasing angled weave
in both the cases.

It was deduced that the average thickness increased with increasing angle
because the flow rate was faster in the large angle weave. Due to the fast flow of
resin through the sample, the accumulation of resin in the sides increased the
average thickness in the sample.

The density decreased with the increase in the angle orientation as the same

amount of resin impregnated the fabric better when the flow of resin was slow.

 

The void percentage also increased with the increase in the angle because of
poor resin impregnation due to fast flow of resin.

These observations were also confirmed when the infusion media was placed.
The average thickness of all the samples with infusion media was more than
those without the infusion media, the void percentage was also higher and the

density lower. All these could be attributed to the infusion time in both the cases.

4.7.1.2 VARIOUS TYPES OF FABRIC

Various types of fabric were tested namely, hand laid [0/9015, unidirectional
stitched [0/9016, fine woven 2D [0/90]6, coarse woven 2D [0/9016 and 03D [0/9016.
The flow was measured in the same manner and the data was collected. All
specimens in this test were oriented in 0/90 manner and were interwoven and no
infusion media was used. All samples were (127 mm)2 ((5 in)2), 6 layers. The
Table 4.3 below shows the flow of resin through the various fabrics and details

the properties of the specimens.

89

Table 4.3 Flow time for various fiber types

 

 

 

 

 

 

 

Specimen Stacking Infusion Thickness(mm) Densi Void (%)
type Sequence time (g/ cm )

(sec)
Hand laid [0/9016 800 1.87 1.81 4.30
Unidirectional [0/9016 847 2.07 1 .85 4.22
stitched
2D(Coarse [0/9015 362 2.36 1 .81 5.32
woven)
2D (Fine [0/9016 497 2.19 1.84 4.90
Woven)
Q30 [0/9015 1800 1.92 1.92 3.47

 

 

 

 

 

 

 

It was observed that the ZD coarse woven had the fastest infusion time and
the Q3D was the slowest to be impregnated. The thickness of the 2D coarse
woven was the maximum among the lot and the density was largest in the 03D
fabric. The void fraction was observed to be largest in the 2D fabric.

The flow was relatively faster in the 2D coarse woven because there were
fiber interstices, creating resin channels which allowed the resin to flow better. In
the 03D the fabric was denser as compared to the rest so the flow was hindered
and slower resin flow was observed. The hand laid and unidirectional stitched did
not vary to a great extent, while fine woven was faster than these.

The thickness variation was the largest in the 2D coarse woven and the least
in the 03D. This was again attributed to the quick flow and thus accumulation of
resin towards the inlet side. All these samples were fabricated without placing the
infusion media on top of the fabric. The void percentage was seen to be largest
in the 2D coarse weave as the flow of resin was uneven and proper impregnation

did not take place.

90

 

 

4.7.1.3 VARIOUS SETUPS

The samples fabricated for this test were made from the same2D coarse
woven and the same stack sequence ([0/9016). The setups used were with
infusion media on top, without infusion media on top, using silicone sheet on top,
using plastic sheet on top, using thin perforated sheet, peel ply and plastic sheet
on top (setups detailed in sections 3.3, 3.5, 3.8, 3.9). The Table 4.4 below Shows
the flow time of resin in all these setups. The table also shows the thickness
variation, void fraction and density of all these specimens.

Table 4.4 Properties at various setups

 

Setup Type Infusion time Thickness Density (g/ Void (%)
(sec) (mm) cm3)

 

With infusion 310 4.45 1.82 4.84
media

 

Without 400 4.32 1.87 4.17
Infusion media

 

With top plastic 382 4.00 1.92 1.71
Plate and darn
only

 

With Silicone 391 4.70 1.73 5.23
top sheet

 

 

With Peel ply, 361 4.00 1.92 1.91
Perforated

sheet, Plastic
plate and dam

 

 

 

 

 

 

It was seen that the flow time of the sample with the infusion media on top
was the fastest (this process is the patented SCRIMP process). The infusion
media accelerated the flow of resin. The infusion time for the setup without the
infusion media was slowest. This was because there was no other factor apart
from the permeability of the fabric to accelerate the flow. The thickness variation

was less than the setup with the infusion media but as compared to the other

91

 

setups it was more. The density was higher than the one with the previous setup
and void fraction was lesser than that with the infusion media on top.

The silicone sheet, also acting as a second mold assisted the resin to move
quicker than the setup without the infusion media. This was because the silicone
sheet was thick (6 mm) as compared to the specimen, and the vacuum pressure
could not compact the fabric layers between the top and the bottom fabric layer
so the micro-channels between each layer, assisted the resin flow. However, i"

since the silicone sheet was flexible and conformed to the shape of the top fabric,

 

the flow of resin was slower on the top layer as compared to those beneath it. As i ‘—
detailed in section 3.5 in Chapter 3, the race-tracking channels in the sides

allowed the resin to flow faster through the sides and thus there were dry Spots

generated towards the outlet side. This phenomenon also led to the

accumulation of resin on the Sides which made the sample thicker through the

sides. The overall thickness was also larger because the vacuum bag could not

compact the Specimen enough because of the silicone sheet.

The setup with the plastic plate on top was faster than that without the infusion
media because the fine gap between the top layer of the fabric and the plastic
plate acted as a path way for the resin to move quickly without any hindrance.
The thickness was very uniform with a standard deviation of 0.13%. This was
because of the fixed dam mentioned in section 3.8 in Chapter 3. The density also
increased and the void volume fraction decreased.

The final setup with the perforated media, peep ply and the top plastic plate
gave the best results as the flow was slower than that with the infusion media on

top. The thickness was maintained with the dam. The density was also

92

 

comparable with the previous setup (with just the plastic plate) the void fraction
was also similar to the previous setup.
4.7.1.4 STACKING THICKNESS

Various stacking layers of fabric were used and the flow time was also
monitored. The number of layers ranged from 3 to 22 layers of 2D coarse woven
fabric, the flow time, volume fraction and the thickness variation were monitored.
All these setups were without the infusion media or the plastic sheet. The Figure
4.7 shows the flow pattern. The Table 4.5 details the thickness, density and void

percentage of the samples.

180

140

80

60

   

Distance covered ( mm)

+ 22 layers —I— 3 layers
40
5 layers ‘~ 8 layers

+ 12 layers +14 layers

0 100 200 300 400 500 600
Time( sec)

Figure 4.7 Flow time at different stacking thicknesses

93

Table 4.5 Properties at various stacking thicknesses

 

 

 

 

 

 

 

 

No of Compaction Stacking Infusion Thickness Densit t! Void

Type layers Time (sec) Sequence time (mm) (g/ cm) (%)
(sec)

2D 3 600 [0/9013 570 1 .79 1.78 4.78
coarse
woven
20 5 600 [0/90]5 318 3.16 1.82 4.22
coarse
woven
2D 8 600 [0/90]. 253 5.68 1 .82 3.98
coarse
woven
20 12 600 [0/90112 240 7.82 1.91 3.94
coarse
woven
2D 14 600 [0/90]... 230 8.29 1 .90 3.65
coarse
woven
2D 22 600 [0/90122 209 10.17 1 .97 3.1
coarse
woven

 

 

 

 

 

 

 

 

It was observed that the thicker the sample the faster was the flow. This
could be attributed to the increase in the through thickness permeability of the
fabric [4]. The resin moved faster in the through thickness direction.

A contradiction with the earlier hypothesis was that if the flow of resin was
faster the impregnation would be non-uniform and the void percentage would
increase. It was seen that in this test this was completely opposite. The thicker
samples showed lesser voids and greater density. This could be attributed to the
fact that since the compaction time was the same for all the samples the amount
of compaction for the thick stacks was not enough to compact it as thus resin
channels were created, thus faster flow. In the previous case where the infusion

media was placed on top the impregnation was non uniform and had more voids.

94

 

4.7.1.5 DIMENSIONS

As mentioned in Chapter 3, various dimensions were used to observe the
flow pattern. Specimens ranging from 127 mm X 127 mm (5 in X 5 in) to 508 mm
X 203 mm (20 in X 8 in) were used and the flow pattern was measured. It was
observed that the specimens with larger distance to cover, covered the same
amount of distance with the same flow rate but in the samples which were larger
in length and placed length wise (inlet- outlet distance longer as shown in Figure
3.25 in Chapter 3) the flow rate diminished as the distance from the inlet
increased. As explained in the Chapter 3, it was because of the loss of

compaction pressure due to Darcy’s Law that the flow retarded.

4.7.1.6 COMPACTION TIME

Another test done was to see the effect of compaction (the amount of time the
vacuum is drawn on the setup). The setup was compacted under vacuum from
10 min to 5 hours. All the samples were of the same dimension, 6 layers of
152.4 x 152.4 mm, the same reinforcing material (coarse woven 2D, [0/90]..), the
same amount of infusion resin( 371.5 g) and no infusion media.The Table 4.6
below shows the effect of compaction time on the flow, thickness, density and

void percentage.

95

 

Table 4.6 Flow time and property variation with varying compaction time

 

 

 

 

 

 

 

 

Specime Compactio Flow time Thickness Densit Void (%)
n No. n time (sec) (mm) (g/ cm )
(sec)
1 0 378 3.22 1.87 5.23
2 900 360 3.20 1 .83 4.77
3 3600 352 3.18 1 .82 4.76
4 7200 337 3.16 1.81 4.56
5 12600 321 3.12 1.81 4.29
6 18000 320 3.11 1.81 4.29

 

 

 

 

 

 

 

It was observed that the flow time improved due to compaction. The overall
void percentage also decreased, the part thickness variation and density was
Similar in all cases. The reduction in void percentage was contradictory to the
earlier hypothesis as the faster flow would diminish the part quality due to the
poor impregnation of the fabric. It was justified that the reduction in the flow time
was about 15 % however the reduction in void percentage was 20%. Also due to
the compaction the flow in the through thickness direction increased which

compensated for the poor impregnation with fast flow time.

4.7.2 LOAD-DEFLECTION CURVE

As discussed earlier in Chapter 3, the impact tests were done on some
specimens. The goal of these tests was to investigate the consistency of the
properties of the specimens manufactured by VARTM. This section discusses
the properties obtained by the load-deflection curve. The Table 4.7 below details

the properties obtained including stiffness, energy absorbed and maximum

96

 

deflection of the samples tested. The Figures 4.8, 4.9, 4.10, 4.11 and 4.12 show
the characteristics of the samples from different batches manufactured by the
VARTM technique developed.

It was seen that the energy absorbed by the specimens increased with the
increase in stacking thickness (overall thickness). This was deduced as more

amount of impact energy would be required to completely penetrate a relatively
b."-

 

 

thick sample.
'3‘.
Load- Deflection i! .

25

20

15 —3 layers
2 — 5 layers
5 10 _ . _ —-— 8 layers
'8 14 layers
0 5
_| . - _ 12 layers

0 .
0 5 10 15 20
Figure 4.8 Load-deflection curves for various layers
Load- Deflection (8 layers)
A —1
5 — 2
I; 3
8 4
..I
0 2 4 6 8 10 12 I4 16

Figure 4.9 8-Iayer specimens with different batches

97

 

Load-Deflection (3 layers)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 .
4 _ 1.4-«4L1
2 3 fN.d.-fim
x 'H,‘ -— 3 layers
; 2 fl unflhfix‘ — 3 layers
8 1 ___r,‘.« 1.. Slayers
A ‘ a,“ 1““ '
0 I... I ' I f ' \ ‘?\
0 2 4 6 8 10 12
Figure: 4.10 3-Iayer specimens with different batches
Load-Deflection (12 layers)
25 1
20 p": fi
15 , = \ 12
;=; .. f] \ ‘1.
8 5 r7 “x :\ "12
_l f" ' .. \‘n
0 - i I I I I I
(I 7 4 6 R 10 I7.
Deflection (mm)
Figure 4.11 12-Iayer specimens with different batches
12 ‘ Load-Deflection (5 layers)
10 ,. ,1: ‘*“-.\
B 2” m “I.
u/ \K‘L r~\
5 VJ! J H‘L‘ >
.x’/ \ "KM. — Senesf
4 / \lr --—— SBI’1882
:5/ \ \ SeriesS
2 J x‘N7x
0 Y‘ r r n \‘im-R
0 2 4 6 8 l0 I2

 

 

 

Figure 4.12 5-layer specimens with different batches

98

 

 

 

 

Table 4.7 Specimen batches and properties

 

Batch Thickness Absorbed Maximum Slope1 Maximum
(mm) Energy (J) Load (Qt) Deflection(mm)

 

3 layer 1.789 : 3 7. 23.68 : 5 % 3.93 : 8 % 0.57 : 4.5 11.62 :4 %
°/o

 

5 layer 3.46 : 2.5% 63.25: 7.5 % 9.53 :10% 1.93 :9 % 11.95 :6 %

 

8 layer 5.727 :470 107.01 :5% 15.87 :2.5% 2.22 :65 13.64 :5%
%

 

12 layer 7.61 : 2 % 200.58 : 7.5 22.55 : 5 % 11.34 : 7 9.75 : 7.25 %
% %

 

 

 

 

 

 

Based on the energy absorption, thickness, maximum load, stiffness and
maximum deflection data, it was observed that the variation in most of the
sample properties was not large and only differed by a maximum of :I: 10 %.

It was deduced that within this range of properties the samples could be

fabricated with VARTM in a consistent manner batch after batch.

4.7.3 THICKNESS

The Figure 4.13 below shows the thickness profile of the various samples

fabricated with all the setups mentioned in the previous setups.

99

 

 

 

 

Thickness vs No. of layers

 

 

 

 

 

 

 

 

 

16 y = 0.4133x + 1.6029
14 ,
g 12 >
E 10 ,
3' 8 —Thickness
g 6 —Linear (Thickness)
2 .
1'5 4
2
0 - ,
0 5 10 15 20 25 30 35
No. of Layers
Figure 4.13 Thickness chart for various layers
Thickness vs. various Setups
4.8
m
4.6 -
E 4.4 , ‘—
5 j . L,
:11 ° -' l .
o 4.2 lSenes1
5 ‘ » . 1
2 l "
J: 4
I- .i
3.8 . =i . '1
3_5 . L3 . mi Li 1 _,_‘
c .. c c o - 06
5.9.9 3.9.55 Log, 2Q; 03.2
3: ch ‘8 2 U) '8 t L3 o: o g (:5 ‘5 5
age see 353 as: 4::
Setup

Figure 4.14 Thickness chart for varying setups

100

 

Table 4.8 Properties for all the setups

 

 

 

 

 

 

 

Setup Thickness(mm) Standard Void (%) Fiber (%)
Deviation

With infusion 4.45 31.7% 4.82 51.78

media

Without 4.32 21.1% 4.13 52.04

Infusion media

With silicone 4.74 62.5% 5.25 49.67
.b_ag

With top plastic 4.00 0.13% 1.72 59.08

Plate and dam

only

With Peel ply, 4.00 0.1% 1.96 59.00

Perforated

sheet, Plastic

plate and dam

 

 

 

 

 

It was seen that the variation in thickness decreased as the number of
stacking layers increased. Also the variation of thickness was the least with the
setup with the plastic sheet on top. The thickness was perfectly controlled by the
dam used in this setup. It was also seen that the density increased and the void
percentage decreased with the setup with the plastic plate on top. The void
percentage reduced to 0.13%and the fiber volume increased to 59.08% overall.
Since it was not possible to add all the figures of the samples, only a few edge
replicas are shown here.

As shown from Figures 4.15 to Figure 4.18, the sample with the plastic plate
had a very controlled thickness, followed by the setup with perforated plastic film,
peel ply and plastic plate. The most undulated was the one with infusion media

on top of the setup.

101

 

WT. ‘ H . -.. ’ T ’—‘
ET“ ““5””; T at. .- -» -- . 4mm
WM “,v”'$‘fr"m_."f 1:" 72;} LL."

Both sides flat with constant thickness

A

Figure 4.15 Sample with plastic plate setup

    

Figure 4.16 Sample with plastic plate, thin plastic sheet and peel ply.

‘ emu-3. gaff-- EMA H‘ 4.45
‘5“ F at!
Wt; .117.

One side flat, the other undulated

 

Figure 4.17 Sample with peel ply and infusion media

102

One side flat, the other undulated

 

Figure 4.18 Sample with no infusion media or peel ply

The arrows in the above figures indicate the thickness of the sample. It was
deduced that with the setup with plastic plate and with plastic plate, perforated
sheet and peel ply, constant thickness samples could be fabricated. The variation
could be controlled to a value of less than 0.13%, which was considered to be a
very acceptable value.
4.7.4 SUMMARY OF RESULTS

The Table 4.9 below summarizes all the tests done and their results. It could
be seen that with final setup (perforated film, peel ply and thin plastic plate) the
flow time improved, the thickness was uniform and the void percentage was
reduced; also the fiber volume fraction improved significantly. The impact
characteristics were similar within the batches. This mentioned technique could
be used to fabricate composite parts repeatedly.

Large sections could also be manufactured. The next chapter explains the
use of the current technique to fabricate large part with controlled parameters like

thickness, void percentage, and fiber volume fraction and infusion times. The

103

material properties could also be tailored with the current technique if the

parameters involved were controlled.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 4.9 Summary Chart
Varied Type No. of Type Average Average
Parameter of samples of Thickness Average Fiber Average
fabric tested setup (mm) Void Volume Infusion
(%) (%) time
(see)
Orientation U 5 1 2.12 5.14 49.32 961
U 5 2 1 .56 5.052 50.06 2820
Fabric Type U 2 2 1.76 5.12 49.78 801
200 2 2 2.07 5.32 49.76 356
20F 2 2 2.19 4.94 51.95 400
L 2 2 1.87 4.21 52.00 869
03D 2 2 1.92 3.65 54.87 1731
Various 2D 3 1 4.36 4.73 48.33 317
setups
3 2 4.32 4.12 50.39 549
3 3 4.00 1.64 59.89 378
3 4 4.00 1.79 58.33 362
3 5 4.76 5.32 46.11 371
Stacking 2D
Thickness
3layer 3 2 1 .79 4.78 48.44 546
5layer 3 2 3.19 4.22 49.33 399
Slayer 2 2 3.95 4.14 50.06 324
Slaver 4 2 5.61 3.98 50.35 297
10layer 2 2 6.25 3.95 51 .00 267
12layer 4 2 7.84 3.94 52.5 247
14layer 2 2 8.29 3.65 52.76 239
22layer 1 2 10.17 3.1 1 54.00 209
30layer 1 2 12.65 3.02 54.33 203
Compaction 6 2 3.15 4.65 50.06 347
Legend:
Setup: 1 - Setup Classical VARTM (SCRIMP) - with infusion media and peel ply
On t0p.

Setup: 2 - Setup with no infusion media or peel ply on top

Setup: 3 - Setup with just plastic plate on top

Setup: 4 - Setup with thin perforated plastic sheet, peel ply and plastic plate.
Setup: 5 - Setup with Silicon bag on top

U - Unidirectional Stitched

2DC - 2- Dimensional coarse woven fabric oriented at 0/90
2DF - 2- Dimensional fine woven fabric oriented at 0/90

L - Hand Laid fabric

03D - Quasi 3-Dimensional Weave at 0/90

104

 

CHAPTER 5

LARGE AND COMPLEX STRUCTURES

To promote the current technique, it is imperative that the recommended
technique could be used for large and/or complex structures for a wide variety of
applications. This chapter briefly describes the use of the technique (with
perforated sheet, peel ply and thin plastic plate on top), mentioned in section 3.9,
for the fabrication of some of such specimens. This chapter is divided into two
sections. The first section discusses how large parts such as yachts and wind
turbine blades are manufactured and how the technique would be implemented.
[Such large structures were not manufactured; however the use of this technique
is explained]. The second section presents the manufacturing of complex shapes
such as arches and sandwich structures.
5.1 LARGE STRUCTURES

Manufacturing large composite structures could be more cost effective than
manufacturing large metallic structures, since it does not require fabricating small
components and joining them together. There is a great advantage in
manufacturing large structures in one infusion. For example the yacht and the
wind turbine blade industries are using VARTM as an essential manufacturing
technique to produce their structures as it does not require secondary processes
and expensive tooling. SCRIMPTM (Seemann’s Composite Resin Infusion
Molding Process) is a commonly used technique. it uses patented infusion media

to manufacture hulls of boats and ships.

105

 

Multiple infusion lines are used to inlet the resin. The infusion lines are
usually placed parallel to each other and resin is inlet sequentially. When the
resin from one infusion line reaches the maximum distance it can travel, that line
is clamped back on and the next adjacent line is unclamped simultaneously. This
is to ensure that the infusion is continuous and there are no obvious voids and
dry spots created during the infusion. The gel time of the resin should be long
enough so that it does not start gelling until the resin from the previous line
reaches the current infusion line. Otherwise, there would be uneven hardening
and properties would be different. Figure 5.1 shows the multiple infusion lines in

the process.

Greater than 3 meters

 

 

 

 

—O

——-—D _

r! H

u n

a! la

'1 I!

H H

u H

H Id

u H

H In

I; I!

z: s: l
I I! I!

In n
G—_ I! I!

37 5:

It I»

5: 5: OUT

II I!

It If

I: H

s: s:

s: :
Outlet '
helix
tubing 1 2 3 4 inlet helix

tubing

 

Figure 5.1 Multiple infusion lines used for resin infusion

Most wind turbine blades are also made of composites and they are often
manufactured with VARTM. Figures 5.2, 5.3 and 5.4 show the overall structure,

details sections and manufacturing process respectively.

‘V

 

Figure 5.2 A typical wind blade

fi’fifi"

 

 

 

 

 

 

 

 

Lower
Part

 

Spars

 

Figure 5.3 Parts of the wind blade

107

 

 

The first step is to place a layer of the reinforcing material (fabric), which is
large enough so that it fits the mold and some of it exceeds the side of the mold.
The mold is then polished with wax and sealant tape is applied to the edges of
the mold. The fabric is then placed on the mold, shaped in the mold of one half of
the wind turbine blade. Next, a layer of peel ply and infusion media are placed on
top of the fabric. Then the inlet and outlet piping is placed. A vacuum bag is
placed on the very top and sealed with the sealant tape on the periphery of the
mold, resulting in a completely sealed setup. Vacuum is drawn thereafter and left
for a long duration such as three hours so that all the excess air is drawn out.
Next the resin is infused and the lay-up is left overnight to infuse. Once the resin
has cured the part is then demolded. Both the top and the bottom parts are
manufactured in this manner. The supporting spars are then placed in between
the two parts and they are bonded together. Figure 5.4 (a) to (f) show the entire

manufacturing process.

 

(a) Fabric

               

 

 

i - Fabric as; “‘
' w ' ' .‘Ma
r. _ .. ‘
o “ “I f... *’ div: .3
‘ v .. ‘ ‘x ‘1.” m "v.7.
-. 4t .9 ‘Q‘ m . . , . .. m

 

Mold
(b) Layup of fabric on the mold

Figure 5.4 Manufacturing process

108

Resin

 

Figure 5.4 Manufacturing process continued

 

Vacuum outlet

 

 

 

Vacuum
pump

(c) Drawing Vacuum

 

 

 

Resin

    
     

Vacuum Outlet

Vacuum
bag

 

Vacuum
Pump

((1) Resin infusion

Figure 5.4 Manufacturing process continued

—’Out

 

 

 

 

A

 

Port Vacuum
PUMP
Outlet

Z : port

\/ / \l \
/ 7; 1'“ REsln 3:319

Flow
inlet Lines
Port

Outlet 7 4. a ,7, , g a" .. R C,

 

 

 

(e) Top view of the resin infusion process

(f) Demolding and final processing

VARTM is used to manufacture wind turbine blades in the above shown

Figure 5.4

110

There are four independent parts in the blade section, the upper and lower
section and two web spars. These parts are manufactured separately and then
joined together with adhesive bonding. The use of infusion media on the outlet
side causes air bubbles in this process towards the outlet ports. This leads to the
removal of the end sections. Therefore, the infusion sections are always made
bigger than the desired size. Besides “bubble trap” technique, mentioned earlier
could be used so that the voids are trapped in the peel ply and do not retract
back.

Since the vacuum bag side is not smooth, it may lead to structural instability
due to the thickness variation throughout the part. A stiff plastic sheet made in
the shape of the structure could be used to match with the mold so that constant
thickness parts could be manufactured. This would improve the properties of the
part produced.

5.2 COMPLEX AND SANDWICH STRUCTURES

in this section arched structures and sandwich composites are discussed.
5.2.1 ARCHES

With the current technique, some complex shapes like arches and domes
could also be manufactured. The idea for this experiment was to demonstrate the
feasibility of the technique for complex structures.

The mold prepared for this specimen was an arched shape made of steel and
polished with wax so that the tool side is smooth. Figure 5.5b shows the setup.
Eight layers of 2D fabric of dimension 203 mm x 203 mm (8 in x 8 in) were

placed on the tool after waxing and the rest of the procedure was done in the

11]

similar manner as explained earlier. The vacuum was drawn and the bag
conformed to the shape, leaving no race-tracking channels. The part was infused
in approximately 4 minutes.

It was observed that the resin accumulated on the edges where the arch
began. This was due to the fact that the fabric did not completely conform to the
shape of the arch, leaving some gap filled by resin. Overall the part thickness

was around 5.65 mm with a standard deviation of 0.85%.

Race tracking
channel

. Arched Mold

Resin rich
edge

 

Figure 5.5a Manufacturing arched specimens

112

Helix
tubing
Vacuum
Infusion Fabric Bag Peel ply

 

 

Sealant tape Glass tool
Figure 5.5b Schematic for arched specimen setup
It was observed that the resin flowed very uniformly thereby impregnating the
entire part and leaving very little visible voids. Figure 5.6 shows the

manufactured part.

 

Figure 5.6 Arched Specimens

5.2.2 SANDWICH STRUCTURES
When materials with high energy absorption-to-weight ratio are mentioned,
the first thing which comes to the foreground is composites. Sandwich

composites enhance this property. As the name suggests, they are made of

113

materials which have lightweight core, such as Styrofoam, which separates fiber
composite faces. The sandwich structure can absorb significant amount of
energy. Some sandwich composites were manufactured with the technique

presented in this study.

5.2.2.1 DESIGNI

The first design for the sandwich composite is shown in Figure 5.7 .The top
layer, bottom layer and the two intermediate layers were made of 2D coarse
woven glass fabric cut into 127 mm x 127 mm (5 in x 5 in) dimensions. Circular
open-celled Styrofoam balls were cut into semicircular with a diameter of 65.5
mm (2.5 in) and height of 9.5 mm (0.37 in) and painted with Styrofoam paint to
make the surface smoother and completely closed. Open-celled means that the
individual cells are permeable and allow materials to pass through it, just like
sponge. The need to make the Styrofoam complete closed, was that the infusion
resin would not seep through the pores of the foam thereby creating a specimen
rich of resin which would increase the overall weight of the sandwich composite.
Also, fiber rich of resin would render the composite, brittleness.

The setup for making the sample is same as shown in the Figure 3.19b
(Chapter 3). The inlet side infusion media was stacked into layers so that it
matched the height of the specimen which was 19.05 mm (0.75 in). This was

necessary because the resin flow in the specimen would be uneven otherwise.

114

+— TOP

(5‘: Layer

1- ”CIA—‘17 Intermediate
”’7‘! /‘ L's} A

 
  
 

. ..‘ -‘— Layers
15:24:}.
" +_ Bottom
Layer

Figure 5.7 Schematic of the sandwich structure

 

Resin Pockets

Figure 5.8 Manufactured sandwich composite
The infusion time for the sample was 159 seconds as there were huge

passage gaps between the domes. The resin flow within the fabric was 194

115

seconds. The gaps between the domes formed resin pockets as shown in the
Figure 5.8.

The infusion took place the same way as explained in section 3.9 .The gaps
promoted the flow of resin and the peel ply at the outlet trapped air bubbles,
leaving the sample with minimum voids. The voids, however, were more as
compared to the rest of the materials used in the previous setups. The cause of
the voids was attributed to the gaps between the domes.
5.2.2.2 DESIGN II

To reduce the amount of gaps between the domes, perfectly matched
domes were made. This would ensure the reduction of the gaps between the two
layers of foam. Polyurethane expanding foam manufactured by US Composites
was used for this. It is available in densities from 0.000064 g/cm3 (2lb per cubic
foot) to 0.00254 g/cm3 (16le per cubic foot). It is available as a two part
expandable liquid. The liquid is a 50/50 mixture and is sold in the following kit
sizes of 1.8 kg, 7.2 kg, 36.2 kg and 45.4 kg (4 lbs, 16 lbs, 80 lbs and 100 lbs).

Egg crate foam (usually made from polyurethane or poly ether) was
used as a base mold for making the final foam. It was coated with silicone
rubber. This was to give the egg crate foam a fine finish, so that the foam made
could be easily demolded. The setup is shown in the Figure 5.9 below. Two
equally sized containers were taken and each part of the components of the
expandable foam was poured into it. Then they were mixed and stirred for
approximately 20 seconds. It was then poured into the silicone rubberized mold

and a glass plate was placed over it, as one-sided flat foam was to be made.

116

 

Some weight was placed on top of the glass plate to make the surface uniform. it
took about 10 to 15 minutes for the foam to fully expand to about 15 times the

original volume. Once made, it was cut to desired shape. Fabrics were placed in

similar manner as the above mentioned setup and infused with VAFlTM

Egg Crate Foam Coated with
31...... W
a -. - .

fl

 

Lower
fabric |ayer Foam Center Top fabric
fabric layer layer

Figure 5.10 Side view of the sandwich structure before infusion

H7

OUT

Outlet

Helix

tubing
3:: Sealant
infusion tape
media

 

Figure 5.11 Infusion process
The dimensions of the fabric were 165 mm x 203 mm (6.5 in x 8 in).This
setup took about 290 seconds to infuse. It was longer than the previous setup.
This was attributed to fewer gaps available for the resin to flow in between the
foam and the fibers. The manufactured parts had visibly fewer resin pockets than
the setup shown in Figure 5.7.
Figure 5.12 shows the manufactured sandwich composite. It was observed
that due to the fabric (coarse woven [0/90]2) in the center portion was not flexible
enough it could not take the exact shape of the domes on both the upper and

lower side even though the domes matches very closely independently. Due to

118

this there was resin pockets developed in the places where the fabric did not
conform to the shape as shown in Figure 5.12.

These resin pockets were potential weak points as the resin deposition would
give brittleness to the composite as a whole. These resin pockets would also
result in additional weight of the composite. Due to the resin-rich pockets, the

fiber percentage of the composite would decrease significantly. Also void content

was higher.

 

Figure 5.12 Sandwich composite

Resin rich
Areas

 

Figure 5.13 Side view

The top, bottom and center fabrics were impregnated completely as observed

which meant that the infusion process was feasible for these structures. The

119

foam was closed celled so the resin did not impregnate the foam and thus the
weight of this structure was comparable to the previous setup where foam domes
were used. Also the thickness was 36.83 mm (1.45 in) with a standard deviation
of 0.93%. It was concluded that the setup could be used for sandwich
composites also.
5.2.2.3 DESIGN III

To minimize the gaps and reduce the resin pockets so that the probability of
void formation was reduced, another method of sandwich structure
manufacturing was used. First, the lower fabric layer (coarse woven [Q/9_0_D and a
polyurethane foam were cut to dimensions of 254 mm x 254 mm (10 in x 10 in).
The foam was placed on top of the fabric layer. Next, two fabric layers were cut
more than the dimensions of the bottom layer and the foam, i.e. 292 mm x 292
mm (11.5 in x 11.5 in). This was done so that these layers of fabric could
conform to the corrugated foam uniformly. Perforated thin sheet and peel ply
were placed on top of it. Then, the vacuum bag was sealed on top, and vacuum

pressure was drawn. This setup was then infused with resin. The schematic of

the 59109 is shown in Figure 5-14 2 layers of intermediate fabric

 

Bottom end foam Bottom fabric
layer

Figure 5.14 Cross sectional view of the 1" step

120

After the infusion was completed and the setup was fully cured, it was
demolded. This setup was placed on a glass plate and freshly prepared foam
mixture was poured on the already infused structure. This was to ensure that the
foam took the shape of the structure below, thereby significantly minimizing the
gaps between the top foam layer and the intermediate fabric. This would reduce
the resin pockets as the top foam layer would perfectly match the intermediate
fabric. A glass plate was then placed to maintain a smooth finish on the top
surface of the foam.

After the foam took the shape, the structure was removed from between the
two plates. Another layer of coarse woven fabric MWas placed on top of the
top foam layer. Perforated thin sheet, peel ply and thin plastic sheet were placed
on top of the fabric and the setup was sealed under a vacuum bag. This setup
was again infused. Since the intermediate layers were already infused and the
foam was closed celled, the resin only traveled through the top fabric layer,
completely impregnating it.

Once the resin had cured completely, it was demolded. It was seen that the
weight of the sample was comparable to the samples produced by the methods
mentioned above. This proved that even though this specimen was larger, it had
comparable weight to the other. It could be concluded that this was due to the
reduction of the resin pockets which were formed earlier. The thickness was also
comparable; it was 36.6 mm (1.44 in) with a standard deviation of 1.03%

Since the top and the bottom fabric layer were clear, the foam beneath them

could be easily seen. However due to the foam not being transparent it was not

12l

 

possible to observe the intermediate layers and the resin pockets. Figure 5.15

shows the manufactured sample.

 

Figure 5.15 Manufactured sample with matching domes

The motivation for making these composites was the impact testing of the
domed structures. It was found out that the domed structures were capable of
absorbing more energy and thus the center core was made in the shape of
domes so that the center layer could absorb more impact energy with the help of

the domed structures.

122

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS
6.1 CONCLUSIONS
From this study, a number of conclusions were reached upon.

a) Resin flow was affected by the angle of orientation of the fabric. The fabric
oriented in the largest angle with respect to infusion direction required the least
infusion time. The difference in flow-path of resin did not vary even when the flow
was accelerated with the use of infusion media placed on top. The fast resin flow
changed the overall thickness in the large angled fabrics. The thickness was the
highest on the sides and the inlet side as there was accumulation of resin due to
quick flow. The void volume fraction increased and density decreased with the

increasing angles.

b) Among hand laid unstitched, stitched unidirectional, 2D fine woven, 2D coarse
woven and 030 woven fabrics, the flow of resin was fastest in the 20 coarse
woven fabric as it provided the least resistance to flow because of large fiber
tows which caused large interstices (the gaps between the tows which act as the
flow channels) in the fabric. The 030 was the slowest to impregnate as the

nesting of the fibers was dense and fine flow channels were diminished.

c) Use of infusion media on the outlet side resulted in the retracting of the

displaced air bubbles, causing infused parts with high void percentage (8%).

Elimination of infusion media from the outlet side and placing peel ply instead

123

trapped the bubbles and did not let them retract back into the sample. This was
named “BUBBLE TRAP ” technique. This reduced the void percentage and

increased the overall fiber volume fraction in the sample.

d) In the setup with 2D coarse woven fabric laid in [ML using no infusion
media and peel ply on the top side, reduced the overall flow of resin
tremendously (29% and higher), but reduced the thickness variation (upto 22%).
The void percentage also decreased from 4.8% to 4.1% and the fiber volume
fraction increased from 48.3% to 52%. The elimination of the infusion media
improved the transparency of the specimen as the finish was smoother than the

setup with peel ply and infusion media on tap.

e) The increase in the stacking thickness of the fabric improved the flow in the
through thickness direction, thereby improving the overall resin flow from 546 sec
to 203 sec. It also decreased the thickness variation from 3% to 1.82% and void

percentage from 4.78% to 3.1% of the sample.

f) Silicone rubber sheet of thickness 6mm was used as a flexible second mold
which was used to give the specimen uniformity and smoothness on the vacuum
bag side, i.e. the top side. Race-tracking, a phenomenon where the infusion resin
follows the easiest path (path of lowest resistance) either through the side of the
setup or the fabric where there is a gap to pass, was observed and it made the

resin to move faster through the sides parallel to the flow and caused the

124

specimen to increase in overall thickness by 8.7%. The flow of resin was faster
than the setup without the infusion media. The void percentage was higher
(5.23%) among all setups used. This silicon bag was a good method to obtain
fine finish on the vacuum bag side; however, it increased the thickness,

increased the void content and decreased the fiber content of the sample.

g) Increasing the compaction time of the setup decreased the flow time of resin in

the fabrics by 15%. It also decreased the void percentage from 5.23% to 4.29%

h) Degassing of the resin improved the overall density, decreased the void

percentage and increased the fiber percentage in the specimen.

i) The use of thin plastic plate on top, instead of the peel ply and infusion media,
improved the surface quality of the specimen and thereby the transparency of the
specimen. It also improved the flow time of the resin (faster than the setup with
no infusion media by 4.5%). The use of control dams helped manufacture
constant thickness parts with standard deviation of 0.13%. This setup decreased
the void content from 4.8% to 1.7% and increased the fiber volume content to

59%.

j) The use of thin perforated plastic sheet, peel ply and plastic plate on top

improved the overall flow close to that with the infusion media (16% slower). The

use of dams controlled the thickness and the variation was 0.13%, similar to that

125

with the plastic plate. The void content decreased to 1.7% and the fiber volume

content increased to 59%.

k) By looking at the overall properties such as absorbed energy, stiffness,
maximum deflection, peak load, void percentage, fiber percentage, resin flow,
and thickness variation, it could be concluded that the specimen produced with
the VARTM setup using perforated sheet, peel ply and thin plastic plate on top,
were consistent and would give a better batch-to-batch repeatability to the
process, making it a more repeatable, cost effective and easy to use process for

composite fabrication.

l) Complex structures such as arches and sandwich composites could be
manufactured with the similar technique. The flow time for the 2D coarse woven
([0/9013), 203 mm x 203 mm, arched specimen was 238 seconds. The thickness
was 5.65 mm with a deviation of 0.85%. For the sandwich composite the flow

time was 194 seconds and the thickness was 19 mm with a deviation of 2.1%.

126

6.2 RECOMMENDATIONS FOR FUTURE WORK

Based on the observations made and lessons learnt, the following
recommendations are made for future work on this topic.
a) All the procedures including fabric lay-up, peel ply placing, vacuum bag
sealing, resin mixing and degassing, in short all the procedures which have
human error involved should be eliminated. To produce repeating samples these

procedures should be automated so that there is no room for human error.

b) The thickness variations during the flow of resin through the fabric should be
monitored using LVDT’s so that run-time monitoring of the thickness can be

observed and resin flow controlled.

c) Since this technique can produce clear samples, and the bubbles traveling
from the inlet to the outlet can be monitored, the void distribution can be
monitored and can further be controlled. The transparency of the samples
produced can be further used for analyzing the damage area in different layers of
laid fabric. Other resin types can be used to improve the transparency of the

specimen

d) Resin flow can be monitored using various types of infusion resin with varying

viscosities. Desired properties can be achieved by varying the resin type.

127

e) Using a thinner silicone sheet (< 2.5 mm) can eliminate the race-tracking effect
and two sided finish can be achieved for the sample with controlled thickness. It
can also be used for complex structures without difficulty as it is flexible and

smooth.

f) A very sensitive camera could be used to monitor the flow front on the sides so

that the flow pattern can be understood more accurately.

9) Multiple infusion lines can be used to produce large structures so that the resin
moves uniformly throughout the structure without creating any loss of vacuum

pressure and develop any void formation.

h) Use of uniform pressure on top, likely air pressure, could be used to eliminate

the control dam, the top plate and the dead weight so that uniform thickness can

be maintained for complex structures

128

APPENDICES

129

APPENDIX A

Pictures of impacted specimens (backlit) top and rear face

NOTE: Nomenclature on samples is to be ignored

 

Figure A3 8 layers

130

 

Figure A5 14 layers

131

APPENDIX B

Resin flow fronts with various setups and stacking thicknesses

 

Figure B1 Resin flow with peel ply and infusion media on top

.‘ e ' -~ ' ‘
.ng’ft
'1 "zvfz-‘Pfqufr .‘”
, 1.. _:, ..

.‘.-

 

Figure 82 Resin flow with only plastic plate on top

132

‘ ‘7 . i ‘ I: _ , .
5 M we. J ‘ .
l * ‘ 5) .." afi‘qu‘: >' f, In

 

Figure B4 Resin flow using perforated sheet, peel ply and plastic plate

133

 

Figure BS Resin flow with no Infusion media or peel ply

 

Figure 36 Through thickness flow through a thick stacking sequence

134

APPENDIX C

Properties charts

Table C1 Gougeon Brother’s lnc., PRO-SET Epoxy Technical and Handling

 

 

 

Data
PRO-SET EpoxyTechnical and Handling Data
117LV/237 1 17LV/
Property 1 17LV/224 1 17LV/226 1 17LV/229 * 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 100230 100231 100230 100231
Mix Ratio Volume 100234 100235 100237 100236 100235
Pot Life 100 9 sample in 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
10 - 12
Gel Time @72° F 2.25 hrs 3 - 3.5 hrs 5 hrs 7 - 8 hrs hrs

 

 

 

 

 

 

 

* Requires at least 125° F Post Cure

Table C2 Various available mold release wax

 

Mold Release

 

 

Wax Non-wax
TR Reg
Temperature Polyvinyl Alcohol (PVA)
TR High
Temperature Silicone Mold Releases

Glaze

 

Honey Carnuba Wax
Meguiar's Mirror

 

Urethane Mold release
Zinc Stearate Mold
Release

Naphtha-based Releases

 

 

135

 

APPENDIX D
Darcy’s Law

_—KA(Ph—Pu)
Q— # L .............................................. (1)

 

Dividing both sides by area and writing in general notations

,u

Here q is the flux (discharge per unit area, with units of length per time, m/s) and
VPis the pressure gradient vector. This value of flux, often referred to as the
Darcy flux, is not the velocity which the water traveling through the pores is

experiencing.

Deriving from 3 D:

yAzur + pg.- — dip = 0 ........................................... (2)
#Téuflpgr—aip ................................................ (3)

Where (p is the porosity. This gives the velocity:
k.-

ui = -— (3:17 - .081) ............................................. (4)
,u

This gives the Darcy’s Law:

kt
q.- = -—(Bip-pgr) ............................................. (5)
p

136

REFERENCES

REFERENCES

1. SW. Beckwith and C. R. Hyland, “Resin Transfer Molding: A Decade of
Technology Advances,” SAMPE Journal, 34(6), 1998, 7-19.

2. R.J. Johnson and R. Pitchumani, “Flow Control Using Localized Induction
Heating in a VARTM Process,” Composites Science and Technology
67, 2007, 669-684.

. R.J. Johnson and R. Pitchumani, “Simulation of Active Flow Control Based on
Localized Preform Heating in a VARTM process, ” Composites: Part A,37,2006
1815-1830.

. B.W. Grimsley, R.J. Cano and AC. Loos, “Characterization of the
Preform Compaction during VARTM Infiltration Process,” Proceedings
of the Twelfth U. 8. -Japan conference on Composite Materials, Pennsylvania
PA. Sep. 22-22, 2006, 442-455.

. D. Heider, P. Simacek, A. Dominauskas, H. Deffor, S. Advani, and J. W.
Gillespie Jr. “Infusion Design Methodology for Thick-Section, Low-Permeability
Preforms using Inter-Laminar Flow Media,” Composite Journal: Part A., 38,
2007, 525-534.

. M.K.Yoon, J. Baidoo, J. W. Gillespie Jr., D. Heider,”Vacuum Assisted Resin
Transfer Molding (VARTM) Process Incorporating Gravitational Effects: A
Closed-Form Solution, " Journal of Composite Materials, 39(24), 2005, 2227-
2242.

. D. Bender, J. Schuster and D. Heider, “Flow rate control during Vacuum
Assisted Resin Transfer Molding Processing,” Composites Science and
Technology. 66, 2006, 2265-2271.

. S. M. Walsh, E. J Rigas, W. A Spurgeon and W. N Roy, “A Non-Contact
Distribution Scheme for Promoting and Controlling Resin Flow for VARTM
Process,” SAMPE Journal, 32, 2000, P 284-293.

. Y.J. Lee, J.H. Wu, Y. Hsu and CH. Chung, “A Prediction method on
ln-Plane Permeability of Mat/Roving Fibers Laminates in Vacuum
Assisted Resin Transfer Molding,” Polymer Science, 2006, 665-670.

10. R. Gauvin, F. Trochu, Y. Lemenn and L. Diallo “Permeability
Measurement and Flow Simulation Through Fiber Reinforcement,”
Polymer Composites, 17(1), 1996, 34-42.

11. C. Dong, “A Fast Permeability Measurement Method Based on Hybrid
Fiber Preforms,” ASME Journal, 127, 2005, 670-676.

138

12. M Nordlund, T.S. Lundstrom, V, Frishfelds and A. Jakovics
“Permeability Network Model for Non-crimp Fabrics,” Composites Part A, 37,
2006, 826-835.

13. A. Hammani, “Key Factors Affecting Permeability Measurement in
the Vacuum Infusion Molding Process,” Polymer Composites, 23(6),2002,
1057—1066.

14. A. W. Chan, D. E. Larive and R. J. Morgan, “Anisotropic Permeability of
Fiber Preforms: Constant flow Rate Measurement,”
Journal of Composite Materials, 27( 10), 1993, 996- 1008.

15. R. S. Luse, J.G. Howard,T. L. Luse and S. G. Advani
“Permeability Characterization Part 1: A Proposed Standard
Reference fabric for Permeability,” Polymer Composites, 16(6), 1995
429-445.

16. A. Dobrowolski, Neal White, “Re-Usable Customized Vacuum Bags,"
SAMPE Journal, 46, 2001, 1192- 1204.

17. Theodore Jacobson, John Crowley, Robert Stratton and Linda Clements
“Low-Cost Reformable Multiuse Tooling System for Composite Materials,”
SAMPE Journal, 46, 2001, 1032- 1042.

18. T. D. Juska, H. B. Dexter and W. H. Seemann, lll
“Pushing the Limits of VARTM, ” SAMPE Journal, 43, 1998, 33-43.

19. E. J Rigas, S. M. Walsh and WA Spurgeon “Development of a Novel
Processing Technique for Vacuum Assisted Resin Transfer Modeling
(VARTM), ” SAMPE Journal, 46(1), 2001, 1086- 1094.

20. S. C. Amouroux “On the role of Membrane to Improve Quality of
VARTM Processed Composites,” SAMPE Journal, 42(1), 2006, 31-40.

21. MA. Livesay, “UV-VARTM Fabrication of Low Cost Composite
Structures, ” SAMPE Journal, 44, 1999, 1338-1344.

22. K.K.Han, B. P. Rice and C. W. Lee, “Double Chamber Vacuum Resin
Transfer Molding, ” SAMPE Journal, 45, 2000, 1547-1556.

23. B. A. Gama, H. Li, W. Li, A. Paesano, D. Heider, and J. W. Gillespie, Jr.

“Improvement of Dimensional Tolerances during VARTM Processing,”
SAMPE Journal, 46, 2001, 55-64.

139

24.

25.

26.

27.

28.

29.

D. Heider, W. Li, K. Engel, andJ. W. Gillespie, Jr.,
“Improvement of surface quality during the VARTM Processing,"
SAMPE Journal, 48, 2003, 282-289.

J. Li, C. Zhang, R. Liang and B. Wang., “VARTM Processes Thickness
Variation Reduction, ” SAMPE Journal, 50, 2005, 2201-2214.

B. W. Grimsley,lP. Hubert, X. Song, R. J. Cano, A.C. Loos and RB. Pipes
“Flow and Compaction during Vacuum Assisted Resin Transfer Molding
Process,” SAMPE Journal, 33, 2001, 1224-1235.

T. Jacobson, J. Crowley, R. Stratton and L. Clements “Low Cost Reformable
Multiuse Tooling System for Composite Materials,” SAMPE Journal, 33, 2001
2123-2136.

R. Vaidyanathan, J Campbell, R. Lopez, J. Halloran, S. Yarlagadda and
J.W. Gillespie, Jr. “Water Soluble Tooling Materials for Filament Winding
and VARTM,” SAMPE Journal,41, 2005, 49-56.

S. M. Lewit, and J. C.Jakubowski “Low Cost VARTM Process for Commercial
and Military Applications,” SAMPE Journal,42,1997,1173-1187.

140