' riff!“

.-
a

n5?

‘ VIA
1””!

 

 

.' $2.2
. 1‘
‘4 1:5:

*1

 

mm

(L)

‘1) \L4

in ,

43;;

CM
(L)

S‘\

This is to certify that the
thesis entitled

Development and Characterization of Biocomposite Cellular
Beams and Plates for Load-Bearing Components

presented by

Mario J. Quagliata

has been accepted towards fulfillment
of the requirements for the

MS. degree in Civil Engineering

 

 

 

 

//' Mjor

MSU is an Affirmative Action/Equal Opportunity Institution

 

-—, * ‘VO .-

9-1" ~04-

. v

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/Dateoue.p65-p.15

DEVELOPMENT AND CHARACTERIZATION OF BIOCOMPOSITE
CELLULAR BEAMS AND PLATES FOR LOAD-BEARING COMPONENTS

By

Mario J. Quagliata

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Civil and Environmental Engineering

2003

ABSTRACT

DEVELOPMENT AND CHARACTERIZATION OF BIOCOMPOSITE
CELLULAR BEAMS AND PLATES FOR LOAD-BEARING COMPONENTS

By

Mario J. Quagliata

Increased environmental awareness and interest in long-term sustainability of
material resources has motivated considerable advancements in composite materials
made from natural fibers and resins. However, despite the developments on the
technology of biocomposites materials, their lower stiffness and strength properties has
limited their applications to non-load-bearing components.

The stiffness of a structural component depends both on the material and
structural stiffness. Thus, the material stiffness shortcomings of biocomposites can be
overcome through efficient structural configurations. Popular designs for advanced
composites based on this idea are cellular and sandwich structures which respectively
maximize material and structural performance.

The objective of this research was to investigate the feasibility of using
biocomposite materials for primary load-bearing components in civil structures by
utilizing high-performance structural forms. The research work presented in this thesis
focuses on the investigation of this concept through the development, manufacturing, and
characterization, both experimental and analytical, of laboratory-scale cellular beams and
plates. These studies were used to analytically investigate the feasibility of using
biocomposite materials for full-size structural components for civil structures, and assess

the future research needs for the development of load-bearing biocomposite structures.

Copyright by
Mario J. Quagliata

2003

To my parents; Michael and Susan
for their love, support, and guidance throughout my life

iv

ACKNOWLEDGEMENTS

I am most grateful to my advisor, Dr. Rigoberto Burguefio, for his insights,
encouragement, time, and patience, which provided me with a guiding light to follow
throughout this research. This work was enabled and sustained by his vision and ideas.

I wish a very special thanks to Dr. Amar Mohanty and Dr. Manjusri Misra for
providing their expertise in biocompostie materials, and for their guidance,
encouragement, and enthusiasm throughout this research.

I am grateful to Geeta Mehta for her guidance in the laboratory, and for all her
help in manufacturing the cellular beams and plates. Her friendship and encouragement
helped me through the difficult times in this research.

I wish to thank the member of my master’s committee, Dr. Ronald Harichandran,
Dr. Manjusri Misra, Dr. Amar Mohanty, and Dr. Amit Varma, for their time and
insightful comments.

I wish thanks to James “JC” Brenton for his help in fabricating the cellular beam
and plate molds, and Siavosh Ravanbakhsh for his help in preparing the test setup and
testing the cellular beams and plates at the Civil Infrastructure Laboratory.

I would like to acknowledge Robert Jurek, Michael Rich, and Kelby Thayer from
the Composite Materials and Structures Center for their help in the manufacturing and
material testing performed in the laboratory for this project.

This study was funded by a 2002 Research Excellence Funds award through
Michigan State University's Composite Materials and Structures Center, which is

gratefully acknowledged.

TABLE OF CONTENTS

 

 

 

 

 

List of Tables ix
List of Figures xii
Chapter 1 Introduction 1
1.1 Overview ............................................................................................................. l
1.2 Fiber Composites for Load Bearing Civil Structures ......................................... 2
1.3 Bio-Based Composite Materials ......................................................................... 3
1.4 Load-Bearing Biocomposites .............................................................................. 7
1.5 Objective and Scope ........................................................................................... 9
Chapter 2 Material Development and Characterization 11
2.1 Overview ........................................................................................................... 1 1
2.2 Material Test Approach .................................................................................... 12
2.2.1 Tensile Testing .............................................................................................. 13
2.2.2 Impact Strength Testing ................................................................................ 14
2.2.3 Coefficient of Thermal Expansion Testing ................................................... 15
2.2.4 Moisture Absorption Testing ........................................................................ 17

2.3 Biocomposite Material Systems ....................................................................... 18
2.3.1 Materials and Constituents ............................................................................ 19
2.3.2 Manufacturing ............................................................................................... 21
2.3.3 Test Matrix .................................................................................................... 23
2.3.4 Tensile Test Results ...................................................................................... 24
2.3.5 Impact Test Results ....................................................................................... 31
2.3.6 Thermal Expansion Test Results .................................................................. 34
2.3.7 Moisture Absorption Test Results ................................................................ 36

2.4 Hybrid Material Systems .................................................................................. 39
2.4.1 Materials and Constituents ............................................................................ 40
2.4.2 Manufacturing ............................................................................................... 41
2.4.3 Test Matrix .................................................................................................... 44
2.4.4 Tensile Test Results ...................................................................................... 45
2.4.5 Impact Test Results ....................................................................................... 55
2.4.6 Thermal Expansion Test Results .................................................................. 56
2.4.7 Moisture Absorption Test Results ................................................................ 57
Chapter 3 Cellular Beams and Plates 60
3.1 Overview ........................................................................................................... 60
3.2 Structural Forms ................................................................................................ 60
3.2.1 Cellular Sandwich Structures ........................................................................ 60
3.2.2 Hierarchical Cellular Structures .................................................................... 62
3.2.3 Hybrid Cellular Structures ............................................................................ 63

3.3 Experimental Studies ........................................................................................ 63
3.3.1 Test Matrix .................................................................................................... 65

vi

3.3.2 Materials and Constituents ............................................................................ 69

 

 

3.3.3 Manufacturing ............................................................................................... 70
3.3.4 Test Setup ...................................................................................................... 74
3.4 Cellular Beams .................................................................................................. 76
3.4.1 Observed Behavior and Test Results: Cellular Beams ................................. 76
3.5 Cellular Plates ................................................................................................... 78
3.5.1 Observed Behavior and Test Results: Cellular Beams ................................. 79
3.6 Comparison with Wood Beams and Plates ....................................................... 86
3.6.1 Observed Behavior and Test Results: Wood Beams .................................... 88
3.6.2 Observed Behavior and Test Results: Wood Plates ...................................... 90
Chapter 4 Analytical Studies 94
4.1 Overview ........................................................................................................... 94
4.2 Properties of Randomly Oriented Short Fiber Composites .............................. 95
4.2.1 Continuous Fiber Composites ....................................................................... 96
4.2.2 Aligned Short Fiber Composites ................................................................. 106
4.2.3 Randomly Oriented Short Fiber Composites .............................................. 111
4.2.4 Comparison with Measured Material Properties ........................................ 118
4.3 Cellular Sandwich Structures .......................................................................... 125
4.3.1 Analysis of Cellular Sandwich Structures .................................................. 125
4.3.2 Comparison with Measured Values ............................................................ 128
4.4 Flexural Performance ...................................................................................... 130
4.4.1 Material Indices .......................................................................................... 131
4.4.2 Shape Factors .............................................................................................. 135
4.4.3 Performance Indices with Shape Factors .................................................... 138
4.4.4 Micro-structural Shape Factors ................................................................... 140
4.4.5 Measured Flexural Performance ................................................................. 143
4.5 Hierarchical Cellular Structures ...................................................................... 145
4.5.1 Analysis ....................................................................................................... 147
4.5.2 Limitations .................................................................................................. 152
4.6 Prestressed Hybrid Sandwich Structures ........................................................ 153
4.6.1 Thermal Effects in Hybrid Cellular Plates .................................................. 154
4.6.2 Temperature Prestress Effects ..................................................................... 168
4.6.3 Prestressing Effects in Full-Scale Panels .................................................... 171
4.7 Performance and Feasibility Studies ............................................................... 172
Chapter 5 Results, Feasibility and Recommendations 180
5.1 Overview ......................................................................................................... 180
5.2 Experimental Studies Summary ...................................................................... 180
5.2.1 Material Systems ......................................................................................... 180
5.2.2 Structural Forms .......................................................................................... 181
5.3 Analytical Studies Summary ........................................................................... 182
5.3.1 Material Systems ......................................................................................... 182
5.3.2 Structural Forms .......................................................................................... 183
5.4 Feasibility Assessment .................................................................................... 184
5.4.1 Automated Manufacturing .......................................................................... 185
5.4.2 VARTM Cellular plates .............................................................................. 186

vii

5.4.3 Testing of VARTM Cellular Plates ............................................................ 197

 

5.5 Recommendations for Future Work ................................................................ 203
5.5.1 Material Systems ......................................................................................... 203
5.5.2 Structural Forms .......................................................................................... 204

Chapter 6 Conclusions 206
References 209

 

viii

LIST OF TABLES

Table 1.1. Typical properties of glass fiber and natural fibers (after Brouwer 2000) ........ 4
Table 2.1. Unsaturated polyester (UPE) resin system ...................................................... 20

Table 2.2. Average properties of natural fiber and unsaturated polyester composites ..... 20

Table 2.3. Physical characteristics of the reinforcing fibers ............................................. 20
Table 2.4. Biocomposite material system naming convention ......................................... 24
Table 2.5. Physical characteristics of the composite material systems ............................. 24
Table 2.6. Tensile test results summary for biocomposite material systems .................... 25

Table 2.7. Modulus of elasticity (MOE) test results for biocomposite material
systems ...................................................................................................................... 26

Table 2.8. Ultimate tensile strength (UTS) test results for biocomposite material
systems ...................................................................................................................... 27

Table 2.9. Tensile elongation at break test results for biocomposite material systems 29
Table 2.10. Specific tensile stiffness of biocomposite material systems .......................... 30
Table 2.11. Specific tensile strength of biocomposite materials systems ......................... 31
Table 2.12. Impact strength (Izod Test) test results for biocomposite material systems. 32

Table 2.13. Coefficient of thermal expansion (CTE) test results for biocomposite
material systems ........................................................................................................ 34

Table 2.14. Moisture absorption test results for biocomposite material systems ............. 37

Table 2.15. Average properties of the hybrid material systems with biocomposite and

E-glass material systems ........................................................................................... 41
Table 2.16. Hybrid material system plate naming convention ......................................... 45
Table 2.17. Tensile test results summary for hybrid material systems ............................. 46
Table 2.18. Modulus of elasticity (MOE) test results for hybrid material systems .......... 47

Table 2.19. Ultimate tensile strength (UTS) test results for hybrid material systems ...... 49

Table 2.20. Tensile elongation at break test results for hybrid material systems ............. 51

Table 2.21. Specific tensile stiffness of hybrid material systems ..................................... 53

Table 2.22. Specific tensile strength of hybrid materials systems .................................... 54
Table 2.23. Impact strength (notched Izod) test results for hybrid material systems ....... 56
Table 2.24. Coefficient of thermal expansion (CTE) test results for hybrid material
systems ...................................................................................................................... 57
Table 2.25. Moisture absorption test results for hybrid material systems ........................ 58

Table 3.1. Test matrix of cellular beams and plates with unsaturated polyester resin ..... 67
Table 3.2. Flexural test results summary of cellular biocomposite beams ....................... 78

Table 3.3. Flexural test results summary of cellular, and hybrid plates with green
hemp fibers ................................................................................................................ 82

Table 3.4. Flexural test results summary of cellular, and hybrid plates with green
hemp fibers ................................................................................................................ 83

Table 3.5. Flexural test results summary of cellular, and hybrid plates with flax fibers.. 86
Table 3.6. Flexural test results summary of cellular, and wood beams ............................ 89
Table 3.7. Flexural test results summary of cellular, and wood plates ............................. 92

Table 4.1. Orthotropic moduli of a unidirectional fiber-reinforced composite by
mechanics of materials ............................................................................................ 104

Table 4.2. Measured modulus of elasticity of biocomposite material systems
compared with results from the Halpin-Pagano quasi-isotropic approximation 121

Table 4.3. Measured tensile strength of biocomposite material systems compared
with results from geometric averaging ................................................................... 124

Table 4.4. Measured versus predicted flexural stiffness of the biocomposite cellular

beams ...................................................................................................................... 129
Table 4.5. Measured versus predicted flexural strength of the biocomposite cellular

beams ...................................................................................................................... 129
Table 4.6. Measured versus predicted flexural stiffness of the cellular plates ............... 130

Table 4.7. Measured versus predicted flexural strength of the biocomposite cellular
plates ....................................................................................................................... 130

Table 4.8. Material properties and material indices for minimum weight ...................... 134

Table 4.9. Section shape factor (sections of equal area) ................................................. 138

Table 4.10. Measured flexural performance of the biocomposite cellular beams .......... 145
Table 4.11. Temperature deflection calculations of the biocomposite cellular plate
with bottom carbon layer (CPl-C) using discrete equations .................................. 163
Table 4.12. Temperature deflection calculations of the biocomposite cellular plate
with bottom carbon layer (CPI-C) using the elasticity solution ............................. 166
Table 4.13. Temperature deflection calculations of the biocomposite cellular plate
with top and bottom carbon layers (CP2-C) using the discrete equations .............. 167
Table 4.14. Pultruded fiberglass panel data .................................................................... 175
Table 4.15. Precast prestressed concrete panel data ....................................................... 175
Table 4.16. Oriented strand board/expanded polystyrene core — insulated structural
panel data ................................................................................................................ 175
Table 4.17. Dimensions of bio-panels for structural applications (SI units) .................. 178
Table 4.18. Dimensions of bio-panels for structural applications (US units) ................. 179
Table 5.1. VARTM cellular plate test matrix ................................................................. 198
Table 5.2. Flexural test results summary of VARTM cellular plates ............................. 202

Table 5.3. Flexural test results of the VARTM plates with all cellular plates tested in
Chapter 3 ................................................................................................................. 202

xi

LIST OF FIGURES

Figure 1.1. North American major end markets of natural fiber composites, year

2000 ............................................................................................................................. 6
Figure 1.2. Comparison of natural fiber (NF) and glass fiber (GF) - unsaturated

polyester (UP) composites .......................................................................................... 7
Figure 1.3. Sandwich structures and cellular structures in nature ...................................... 8
Figure 2.1. Schematic of the material experimental program ........................................... 11
Figure 2.2. Sample layout for material property tests ....................................................... 13
Figure 2.3. ASTM D790 Type I tensile specimen ............................................................ 14
Figure 2.4. Impact strength testing, ASTM D256 ............................................................ 15
Figure 2.5. Impact strength testing apparatus ................................................................... 15
Figure 2.6. Thermal apparatus used for coefficient of thermal expansion tests ............... 16
Figure 2.7. Environmental chamber used for moisture absorption testing ....................... 17
Figure 2.8. Close-up view of natural fibers used for manufacturing ................................ 19
Figure 2.9. Manufacturing process of compression molded biocomposite plates ............ 22
Figure 2.10. Compression molded biocomposite materials .............................................. 23

Figure 2.11. Modulus of elasticity (MOE) test results for biocomposite material
systems ...................................................................................................................... 26

Figure 2.12. Ultimate tensile strength (UTS) test results for biocomposite material
systems ...................................................................................................................... 28

Figure 2.13. Tensile elongation at break test results for biocomposite material

systems ...................................................................................................................... 29
Figure 2.14. Specific tensile stiffness of biocomposite material systems ......................... 30
Figure 2.15. Specific tensile strength of biocomposite material systems ......................... 31

Figure 2.16. Impact strength (Izod Test) test results for biocomposite material
systems ...................................................................................................................... 33

Figure 2.17. Impact strength (Izod Test) test results for biocomposite materials only 33

xii

Figure 2.18. Coefficient of thermal expansion (CTE) test results for biocomposite
material systems ........................................................................................................ 35

Figure 2.19. Comparison of Coefficient of thermal expansion (CTE) with
conventional materials .............................................................................................. 36

Figure 2.20. Moisture absorption test results for biocomposite material systems ............ 38

Figure 2.21. Percent weight absorbed versus square root of time for biocomposite

material systems ........................................................................................................ 38
Figure 2.22. Hybrid material system ................................................................................. 39
Figure 2.23. Fiber mat fabrics used as face sheets for hybrid material systems ............... 41
Figure 2.24. Manufacturing process of compression molded hybrid plates ..................... 43
Figure 2.25. Hybrid compression molded plates .............................................................. 44

Figure 2.26. Compression molded plates of fiber mat materials and polyester resin ....... 44
Figure 2.27. Modulus of elasticity (MOE) test results for hybrid material systems ......... 47
Figure 2.28. Modulus of elasticity (MOE) test results for face sheet material systems 48
Figure 2.29. Ultimate tensile strength (UTS) test results for hybrid material systems ..... 50

Figure 2.30. Ultimate tensile strength (UTS) test results for face sheet material
systems ...................................................................................................................... 50

Figure 2.31. Tensile elongation at break test results for hybrid material systems ............ 52
Figure 2.32. Tensile elongation at break test results for face sheet material systems ...... 52
Figure 2.33. Specific tensile stiffness of hybrid material systems .................................... 54
Figure 2.34. Specific tensile strength of hybrid materials systems ................................... 55
Figure 2.35. Impact strength (notched Izod) test results for hybrid material systems ...... 56

Figure 2.36. Coefficient of thermal expansion (CTE) test results for hybrid material
systems ...................................................................................................................... 57

Figure 2.37. Moisture absorption test results for hybrid material systems ....................... 59

Figure 2.38. Percent weight absorbed versus square root of time for hybrid material
systems ...................................................................................................................... 59

Figure 3.1. Periodic cellular sandwich bema and plate .................................................... 61

xiii

Figure 3.2. Hierarchical cellular sandwich beam and plate .............................................. 62

Figure 3.3. Hybrid cellular beam and plate ....................................................................... 63
Figure 3.4. Schematic of the sample matrix for the experimental studies ........................ 66
Figure 3.5. Pictorial summary of the cellular beam test units .......................................... 67
Figure 3.6. Pictorial summary of cellular plates with green hemp/UPE .......................... 68
Figure 3.7. Pictorial summary of the cellular plates with raw hemp/UPE ....................... 68
Figure 3.8. Pictorial summary of the cellular plates with chopped flax/UPE ................... 69
Figure 3.9. Specially designed molds used for manufacturing the cellular beams and
plates ......................................................................................................................... 70
Figure 3.10. Preparation of the mold and material system ............................................... 72
Figure 3.11. Manufacturing process of biocomposite cellular beams and plates ............. 74
Figure 3.12. Schematic of flexural test setups .................................................................. 75
Figure 3.13. Pictures of flexural tests on cellular beams and plates ................................. 76

Figure 3.14. Force versus displacement response at mid-span of cellular
biocomposite beams .................................................................................................. 77

Figure 3.15. Specific stiffness of cellular biocomposite beams ........................................ 78

Figure 3.16. Force versus displacement response at mid-span of cellular and hybrid
plates with green hemp fibers, and unsaturated polyester ........................................ 81

Figure 3.17. Force versus displacement response at mid-span of cellular and hybrid
plates with green hemp fibers, and unsaturated polyester, zoomed in view ............. 81

Figure 3.18. Force versus displacement response at mid-span of cellular, and hybrid
plates with green hemp, and raw hemp fibers, and unsaturated polyester ................ 83

Figure 3.19. Specific stiffness of cellular and hybrid plates with green hemp fibers,

and unsaturated polyester .......................................................................................... 84
Figure 3.20. Force versus displacement response at mid-span of cellular and hybrid

plates with flax fibers, and unsaturated polyester ..................................................... 85
Figure 3.21. Douglas Fur beam and plate test units .......................................................... 86
Figure 3.22. Wood plate test units .................................................................................... 87

xiv

Figure 3.23. Force versus displacement response comparison of cellular, and wood

beams ........................................................................................................................ 89
Figure 3.24. Specific stiffness of cellular and wood beams ............................................. 90
Figure 3.25. Force versus displacement response comparison of cellular, and wood

plates ......................................................................................................................... 92
Figure 3.26. Specific stiffness of cellular and wood plates .............................................. 93
Figure 4.1. Orthotropic lamina, and representative volume element ................................ 97
Figure 4.2. Representative volume element for loading in the l-direction ...................... 98
Figure 4.3. Predicted versus measured E1 ......................................................................... 99
Figure 4.4. Representative volume element for loading in the 2-direction .................... 100
Figure 4.5. Variation of composite moduli with fiber volume fraction .......................... 101
Figure 4.6. Representative volume element loaded in the l-direction ........................... 102
Figure 4.7. Representative volume element loaded in shear .......................................... 103

Figure 4.8. Variation of the longitudinal modulus of an aligned short fiber composite
with fiber aspect ratio, using the Halpin-Tsai equation .......................................... 107

Figure 4.9. Schematic representation of matrix shear deformations in a short fiber
composite ................................................................................................................ 108

Figure 4.10. Single fiber composite element showing the fiber-matrix interfacial
stress and the fiber internal stress for stress transfer by slip ................................... 108

Figure 4.11. Variation of the longitudinal strength of an aligned short fiber
composite with fiber aspect ratio, using the model by Piggott ............................... 110

Figure 4.12. Effect of fiber aspect ratio on the stress-strain response of short fiber
composites ............................................................................................................... 111

Figure 4.13. Comparison of measured elastic modulus of a green hemp/polyester
biocomposite with analytical predictions ................................................................ 119

Figure 4.14. Comparison of measured elastic modulus of a green hemp/polyester
biocomposite with analytical predictions, zoomed in view .................................... 120

Figure 4.15. Results of the analytical models compared with the measured elastic
modulus of a green hemp/polyester biocomposite .................................................. 120

XV

Figure 4.16. Measured modulus of elasticity of biocomposite material systems
compared with results from the Halpin-Pagano quasi-isotropic approximation 122

Figure 4.17. Comparison of measured strength of a green hemp/polyester

biocomposite with analytical predictions ................................................................ 123
Figure 4.18. Measured tensile strength of biocomposite material systems compared

with results from the geometric averaging approach .............................................. 124
Figure 4.19. A beam of square section, loading in bending about the x-x axis .............. 132
Figure 4.20. Effect of relative density on the relative performance ............................... 144
Figure 4.21. Overview of the analysis process for hierarchical hybrid cellular

biocomposite beams and plates ............................................................................... 146
Figure 4.22. Unit cell model for the mechanical behavior of cellular solids .................. 148
Figure 4.23. Section geometry of a general laminate ..................................................... 151

Figure 4.24. Accuracy of the proposed analysis procedure for varying number of
cellular layers .......................................................................................................... 153

Figure 4.25. Temperature deformations in unbonded and bonded materials with a: >
aq ............................................................................................................................. 156

Figure 4.26. Example cross-section with discrete element ............................................. 158

Figure 4.27. Calculated stress and strain distributions due to thermal effects in the
biocomposite cellular plate with bottom carbon layer (CPl-C) ............................. 165

Figure 4.28. Calculated stress and strain distributions due to thermal effects in the
biocomposite cellular plate with top and bottom carbon layers (CP2-C) ............... 168

Figure 4.29 Force-strain response of the biocomposite hybrid cellular plate with a
bottom carbon face sheet, with correction for temperature prestress effects .......... 170

Figure 4.30. Force-strain response of the biocomposite hybrid cellular plate with top
and bottom carbon face sheets, with correction for temperature prestress effects . 170

Figure 4.31. Strain and stress calculations and distributions due to temperature
effects in a full-scale biocomposite cellular plate ................................................... 172

Figure 4.32. Performance biocomposite cellular beams with beams of conventional
construction materials ............................................................................................. 173

Figure 4.33. Performance of full scale biocomposite cellular panels versus
conventional flooring systems ................................................................................ 174

xvi

Figure 4.34. Schematic of the dimensions selected for the bio-panel structural

components ............................................................................................................. 177
Figure 5.1. Schematic of the VARTM manufacturing process ...................................... 187
Figure 5.2. Hybrid pre-form fabric: raw hemp core with jute face sheets ...................... 188

Figure 5.3. Hybrid pre-form fabric: raw hemp core with glass face sheets and top and

bottom carbon ......................................................................................................... 188
Figure 5.4. Pre-compression of the hybrid fabric pre-forms .......................................... 189
Figure 5.5. Preparation vacuum bagging materials ........................................................ 191
Figure 5.6. Placement of pre-form bio-fabric in the cellular plate mold ........................ 192
Figure 5.7. Preparation of dry pre-form fabric and vacuum bagging materials ............. 193
Figure 5.8. Sample with vacuum bag, resin ports, and vacuum ports ........................... 195

Figure 5.9. Resin flow front movement (total time to complete impregnation = 12
minutes) ................................................................................................................... 196

Figure 5.10. Cut through VARTM cellular plate showing fiber wet-out through the
entire depth .............................................................................................................. 197

Figure 5.11. VARTM cellular plate with a non-wrapped jute fabric/raw hemp core
pre-form .................................................................................................................. 198

Figure 5.12. VARTM cellular plate with a wrapped jute fabric/raw hemp core pre-
form ......................................................................................................................... 198

Figure 5.13. VARTM cellular plate with a warpped glass strand mat/raw hemp core
pre-form and top and carbon layers (47% fibers by weight) .................................. 199

Figure 5.14. Force versus displacement response at mid-span of VARTM cellular
plates, compared with a green hemp plate made by hand layup ............................. 201

Figure 5.15. Zoomed-in view of the force displacement response of VARTM cellular
plates ....................................................................................................................... 201

Figure 5.16. Comparison of the force versus displacement response at mid-span of
the VARTM plate with top and bottom carbon versus the hand layup cellular
plates with carbon face sheets ................................................................................. 202

xvii

CHAPTER 1 INTRODUCTION

1.1 Overview

Fiber reinforced polymer (FRP) composites have surpasses their initial target
applications in the aerospace industry to become a viable material alternative in the
sporting goods, automotive, and construction industries. High performance FRP
composites made with synthetic fibers such as carbon, glass, or aramid embedded in
polymeric matrices provide the advantages of high stiffness and strength to weight ratio
and increased chemical inertness compared to conventional construction materials, i.e.,
wood, concrete, and steel. In spite of these advantages, the wide-spread use of synthetic
FRP composites has been limited, among several factors, due to their higher initial
material costs, their use in non-efficient structural forms, and their environmental impact.

Increased environmental awareness and the interest in long-term sustainability of
construction materials have thus challenged the development of environmentally
friendlier alternatives to synthetic oil-based FRP composites (Mohanty et a1. 2000).
Natural-fiber—reinforced polymer composites, or biocomposites, have emerged in the past
decade as an environmentally friendly and cost-effective option to synthetic FRP
composites. Despite the interest and environmental appeal of biocomposites, their use
has been limited to non-primary, or non-load-bearing applications due to their lower
strength and stiffness compared with synthetic FRP composites (Biswas et al. 2001).
Recent developments, however, have shown that the properties of “engineered”
biocomposites are comparable with those of entry-level structural synthetic fiber

composites, namely glass-FRP (Mohanty et a1. 2001, Rout et al. 2001).

While biocomposite materials with specific properties equivalent to entry-level
structural materials are feasible, this performance level is still not enough to compete
with our existing construction materials. However, the performance of a component
depends on both its material and structural properties. The lower material stiffness of
biocomposite materials can thus be overcome by using efficient structural configurations
that place the material in specific locations for highest structural performance for
minimum material use and thus minimum weight (Gibson and Ashby 1988). This
concept has been recognized for some time and has recently gained new attention due to
the way that nature’s own materials and structures follow this principle.

The above concepts have been integrated in this thesis, where engineered
biocomposites and their arrangement in efficient cellular and sandwich configurations are
shown to lead to environmentally friendly and structural efficient structural components.
In the following section the background on the current uses of FRP composites in load
bearing civil structures is presented to give context to the use of biocomposites in the
civil sector. This is followed by background on biocomposite materials, and high-
performance cellular and sandwich structural forms that can be used to efficiently arrange
biocompostie materials. The chapter is concluded with the objective and scope of the

research work presented in this thesis.

1.2 Fiber Composites for Load Bearing Civil Structures

High performance FRP composites, with synthetic fibers such as carbon, glass, or
aramid, are increasingly being considered for use in civil infrastructure applications
ranging from rehabilitation of buildings and bridges to the construction of new structural

systems (Karbhari and Seible 2000). The advantages of FRP composites that have

attracted their use in civil structures includes: high specific properties, enhanced fatigue
life, corrosion resistance, and tailorable properties (Barbero and GangaRao 1991).
Specific applications of FRP composites in civil infrastructure include: retrofit of
columns in seismic regions, strengthening of degraded beam, slab and deck structures,
replacement and new construction of bridge decks, and new structural systems (Karbhari
and Seible 2000). FRP composites have also been used for reinforcement and
prestressing cables in reinforced and prestressed concrete, and have been manufactured in
standard pultruded shapes (Barbero and GangaRao 1992, and 1991).

Despite the advantages of high performance FRP composites, their wide-spread
implementation in civil infrastructure has been limited, primarily due to: (i) higher initial
material cost, (ii) lack of familiarin in most areas of civil engineering, (iii) lack of
comprehensive standards and design guidelines at the present, and (iv) need for
integrated materials-process-design structure in product development (Karbhari and
Seible 2000). While biocomposites materials may be a low-cost, environmentally
friendly alternative to synthetic FRP composites, biocomposites will also encounter the
same issues limiting the implementation of high performance FRP composites in civil

infrastructure.

1.3 Bio-Based Composite Materials

Natural fibers embedded in a natural or synthetic polymeric matrix, known as
biocomposites, have gained recent interest because of their low material and
manufacturing costs, light weight, high specific modulus (elastic modulus over density),
and environmentally friendly appeal (Mohanty et al. 2000). Depending on their origin,

natural fibers may be grouped into: bast (stem), leaf, and seed types (Brouwer 2000).

The best know examples are: (i) bast: hemp, flax, jute, kenaf, and ramie; (ii) leaf: sisal,
henequen, and pineapple leaf fiber; (iii) seed: cotton, coir. Bast fibers have the highest
mechanical properties (especially: hemp, flax, and ramie) and thus, are the ones typically
considered for high-quality applications. Typical mechanical properties of these fibers
together with E-glass fibers are given in Table 1.1. The density of natural fibers is about
half that of E-glass (Table 1.1), which makes their specific strength quite comparable,
while the elastic modulus and specific modulus are comparable or even superior to E-

glass fibers.

Table 1.1. Typical properties of glass fiber and natural fibers (after Brouwer 2000)

 

 

 

 

 

 

 

 

 

 

 

Fibers
E-Glass Flax Hemp Jute Ramie Sisal
2.55 1.40 1.48 1.46 1.50 1.33
2400 800 - 1500 550 - 900 400 - 800 500 600 - 700
73 60 - 80 70 10 - 30 44 38
29 26 - 46 47 7 - 21 29 29
3 1.2- 1.6 1.6 1.8 2 2-3
oisture
u 7 a 12 12-17 11
bsorption (%)
Ema/"9 (s) 1.3 0.5 - 1.5 0.6 - 1.8 0.35 1.5 - 2.5 0.6 - 0.7
Raw
[MT/"9 (3;) 1.7 - 3.8 2 - 4 2 - 4 0.9 - 2 1.5 - 2.5 --
Fabnc/mat

 

 

 

 

 

 

 

 

The applications for which biocomposites have been studied include interior and
exterior paneling of automobiles, interior paneling in rail vehicles, furniture, and
replacement of wood products in housing applications. Daimler Chrysler (2001) has used
biocomposites of green flax and hemp fiber mats with polyester for under body panels of

the EvoBus (a touring coach). Daimler Chrysler found replacing the under body panel,

previously a glass fiber composite, with biocomposites required 83 percent less energy to
manufacture, and the resulting part was 40 percent cheaper. In addition, the same tools
and machine used to manufacture the part with glass-fiber composites were used for
manufacturing with biocomposites. Biocomposite materials have also been used for
interior paneling applications in automobiles, including door paneling and rear shelf
paneling (Biswas et a1. 2002). The interior paneling of rail vehicles in Germany has also
been manufactured using biocomposites, mainly for weight savings over glass-fiber
composites (Nickel and Riedel 2003). In this application the standards for fire protection
are very high in Germany. Thus, the biocomposites were treated with flame retardants to
attain the high levels of fire-protection required by the standards. Phenix Biocomposites
(2003) of Mankato, Minnesota have developed biocomposite materials with decorative
surfaces for furniture, table tops, wall panels, and other home and office finished surface
applications. These applications by Phenix Biocomposties show that biocomposites can
be manufactured and used in products where attractive surface finishing is required. Jute
and coir based biocomposites have been developed in India as substitutes for plywood
and medium density fiber boards for low-cost housing needs (Biswas et a1. 2001). Other
products under development include the use of sisal-based biocomposites as panels and
roofing sheets, which could also be used as false ceilings and partition boards. The
current market uses of bio-based composites in North America are shown in Figure 1.1

(Mohanty et a1. 2000).

Others (7%)
Automotive (8%)

  
  
  

lndustrial/
Consumer (10%)

Building
Products (75%)

Figure 1.1. North American major end markets of natural fiber composites, year 2000

Consideration of biocomposites for load-bearing, i.e. structural, applications has
been neglected due to their low stiffness and strength in comparison with conventional
construction materials. However, recent research on biocomposite materials has shown
that “engineered,” or treated, natural fibers can lead to biocomposites with properties that
can compete with glass fiber composites (Figure 1.2). This has motivated further
research initiatives that consider biocomposites as a technical and environmentally
conscious alternative to E—glass reinforced fiber composites (the most common synthetic
fiber composites) without sacrificing performance. Natural fibers/unsaturated polyester
composites show lower density, equal flexural modulus, comparable flexural strength,
but relatively poor impact strength as compared to a glass fibers composite as show in

Figure 1.2 (Mishra et a1. 2002, Mohanty et al. 2000).

Others (7%)
Automotive (8%)

  
   
 

Industrial]
Consumer (10%)

Building
Products (75%)

Figure 1.1. North American major end markets of natural fiber composites, year 2000

Consideration of biocomposites for load-bearing, i.e. structural, applications has
been neglected due to their low stiffness and strength in comparison with conventional
construction materials. However, recent research on biocomposite materials has shown
that “engineered,” or treated, natural fibers can lead to biocomposites with properties that
can compete with glass fiber composites (Figure 1.2). This has motivated further
research initiatives that consider biocomposites as a technical and environmentally
conscious alternative to E-glass reinforced fiber composites (the most common synthetic
fiber composites) without sacrificing performance. Natural fibers/unsaturated polyester
composites show lower density, equal flexural modulus, comparable flexural strength,
but relatively poor impact strength as compared to a glass fibers composite as show in

Figure 1.2 (Mishra et a1. 2002, Mohanty et a1. 2000).

 

[ IUP30wt°/oGF DUP35wt%M=

 

100%:

   
    

80%1

     
  
  

 

38 kJ/mm‘

60°/o‘

  

l

  

40%-

20°/o‘

20 kJ/mm2

 

 

 

 

 

 

0%:
Flexural Flexural inpact Density
Strength Modulus Strength

Figure 1.2. Comparison of natural fiber (NF) and glass fiber (GF) — unsaturated polyester (UP)
composites

1.4 Load-Bearing Biocomposites

While the properties of biocomposites can compete with those of E-glass
composites, their strength, and particularly stiffness, is lower than that of conventional
structural materials. This has limited their application primarily to non-load—bearing, or
non-structural, components. One way to overcome their low mechanical properties is
through hybridization with stiffer and stronger (i.e., synthetic) fibers (Mishra et al. 2003,
Biswas et al. 2001). A more efficient way of improving the performance of biocomposite
components is to use them in high-performance structural forms. Scrutiny of nature’s
materials shows that high structural efficiency can be achieved by strategic material
organization (Gibson and Ashby 1988, Gunderson and Thorp 1993). Optimized
hierarchical cellular arrangement of the material leads to enhanced and efficient use of

the material system and thus structural components at reduced weight and cost.

The stiffness shortcomings of biocomposites can then be overcome by efficient
structural configurations. By looking to nature for efficient structural forms, the most
common are cellular sandwich structures, which consist of a dense skin integrally
connected to a low-density cellular core, as shown Figure 1.3a. Cellular materials are
made of an interconnected network of solid struts or plate, which have the general form
of either two-dimensional honeycombs (Figure 1.3b) or three-dimensional foams (Figure
1.3c). By integrally connecting the low—density cellular core with the dense skin face
sheets delamination between the core and face sheets can be avoided, which can lead to

premature failure of sandwich structures (Jones 1999).

 

 

 

 

(c) Cancellous bone 3-D cellular structure (Gibson and Ashby 1988)

Figure 1.3. Sandwich structures and cellular structures in nature

1.5 Objective and Scope

The objective of this research is to investigate the feasibility of using
biocomposite materials for primary load-bearing components in civil structures. High-
performance structural forms, such as cellular and sandwich configurations, will be
investigated with the aim of improving the structural performance of biocomposites for
application in load-bearing civil structures. The thesis focuses on the investigation of this
concept through the development, manufacturing, and characterization, both
experimental and analytical, of laboratory-scale cellular beams and plates. The potential
of cellular biocomposites for full-size load-bearing structural components is studied by
comparing extrapolated results to the performance of commercial components made with
conventional construction materials, and by selecting dimensions of cellular
biocomposite structures for civil applications based on deflection limits given in
structural design codes. The research work involved the following tasks:

0 Experimental material characterization of biocomposite material systems

0 Identification and characterization of high-performance structural forms using
laboratory-scale cellular beams and plates

0 Identification and validation of analytical tools for biocomposite materials and
cellular structures

0 Analytically studies on the feasibility of using biocomposite materials for load-

bearing structural components

The research presented in this thesis integrates all the aspects of structural design,
from selection of the material and structural form, to manufacturing of the structural
component. Thus, a holistic view of the design process for biocomposite structural
components is shown, which emphasizes the impact each link in the design process has
on the performance of the resulting structural component. Understanding this interaction
between the various aspects of design can help engineers design efficient structures,
which meet the desired performance for a specific application.

Based on the above research tasks, the chapters of the thesis are organized in the
following order:

0 Chapter 2: Material development and characterization

Chapter 3: Cellular beams and plates

Chapter 4: Analytical studies

Chapter 5: Results, feasibility, and recommendations

Chapter 6: Conclusions

10

CHAPT ER 2 MATERIAL DEVELOPMENT AND CHARACTERIZATION

2.1 Overview

A material experimental program was undertaken with two main objectives: (i) to
establish a minimum database for biocomposite materials to asses their current status and
evaluate their feasibility for structural applications, and (ii) to provide the necessary
material characterization for the material systems used in the structural experimental
studies (Chapter 3). Hybrid material systems using hemp fibers in combination with
chopped E-glass, and woven jute fabrics were also investigated. In addition, material
systems using chopped E-glass, unidirectional carbon, and woven jute mats were
investigated to provide the necessary material characterization for the integral face sheet
materials used in the structural experimental studies (Chapter 3). The performance of the
biocomposite material systems was compared with that of E-glass composite materials.

A schematic of the material experimental program is shown in Figure 2.1.

 

Material
Experimental Program

 

 

 

 

 

Material
Test Approach

I
l l

Biocomposite Hybrid
Materal Systems Material Systems

 

 

 

 

 

 

 

 

 

 

 

Face Sheet
Materials

 

 

 

Figure 2.1. Schematic of the material experimental program

11

In the following sections, the test approach and testing procedures followed by the
test results from the biocomposite and hybrid material systems are provided. The
manufacturing procedures for both the biocomposite and hybrid material systems is also

presented and summarized in photographic sequences.

2.2 Material Test Approach

The mechanical properties of the material systems were studied through American
Society of Testing Materials (ASTM) material testing procedures for fiber-reinforced
plastics. The following mechanical properties were evaluated in the material
experimental program:

0 Tensile modulus, strength, and elongation at break

0 Fracture toughness

0 Coefficient of thermal expansion

0 Moisture absorption
Details of the testing procedures used for each material property are provided in the
following sections. All of the material characterization testing was conducted at the
facilities of the Composite Materials and Structures Center (CMSC) at Michigan State
University.

Specimens for performing the mechanical property tests were prepared by
manufacturing a 200 mm x 200 mm compression molded plate for each material system
type. Manufacturing details of the compression molded plates are provided in sections
2.3.2 and 2.4.2. The number and size of the test specimens required for each test were

prepared using the layout shown in Figure 2.2. The samples were cut from the plate

12

using a band saw, and were then machined to the proper shape and dimensions using

 

 

 

 

 

 

 

 

 

 

standard templates.
Impact (10)
/ 0.50-9.50"
l
./ l
r i
ll
Tone-lie (arfi‘
\ // oso'xzso'
/— Thermal
:/ Expansion (4)
+4 0.35'x0.50'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.2. Sample layout for material property tests

2.2.1 Tensile Testing

The tensile properties of the material systems were determined using ASTM
D790 specifications for sample shape, size, conditioning, and testing procedure. ASTM
Type I dogbone specimens were used: a 13.0 mm x 57.0 mm narrow section with 20 mm-
wide grip ends and a total length of 165 mm (Figure 2.3). A Universal Testing Systems
testing frame with an external laser extensometer was used to conduct the tests. The
applied load and displacement on the narrow section of the sample were recorded,
converted to a stress-strain response, and used to determine the tensile modulus, strength,
and elongation at break. The samples were only tested in one direction due to the random

orientation of the natural fiber reinforcement.

13

 

 

165 mm

\r—fl min—>1

<1: Force Force t:1>

\—

 

 

 

 

 

 

 

20mm~ 13mm—

3:115 mm

Figure 2.3. ASTM D790 Type I tensile specimen

2.2.2 Impact Strength Testing

The Impact strength of the material systems was measured using ASTM D256
specifications for sample shape, size, conditioning, and testing procedure. ASTM Izod
type test specimens were used: 63 mm x 12.7mm samples with a 45° notch and bottom
radius of 0.25 mm at the sample mid-length (Figure 2.4a). The samples were tested using
ASTM Test Method A (Izod Type), where the sample is cantilevered in the setup with the
notch facing the striking edge (Figure 2.4b).

The samples were tested by swinging the pendulum, with striking edge, and
striking the notched cantilevered sample. The energy lost (required to break the sample)
as the pendulum swings through the sample is measured from the distance of the
pendulum follow through. The energy lost per unit of specimen thickness at the notch
(J/m) is reported as the impact strength. The testing apparatus used for this experiment
program (Figure 2.5) automatically measured the energy lost and computed the impact
strength. Wind and friction losses in the system were measured and used to correct the

impact strength value.

14

<1: Pendulum
— Striking Edge

45- W—\
J— /\ /\ i Notch—\\< 22'0 mm
10.16 mm 1 R-Msn-m 127 mm
l <31.s mm>l i /5i:_:d/ 7133”"\
<——63 mm———— é/y/ \Jfi:\§\\\

 

 

 

 

 

 

(a) Izod Type test specimen (b) Izod test (Test Method A)

Figure 2.4. Impact strength testing, ASTM D256

 

Figure 2.5. Impact strength testing apparatus

2.2.3 Coefficient of ThennaI Expansion Testing

The coefficient of thermal of the material systems was measured using the

Thermal Mechanical Analysis (TMA) apparatus at the CMSC (Figure 2.6). The two

15

main components of the TMA apparatus are the probe and the furnace. The probe holds
the sample and measures the change in length during the test. The furnace applies heat to
the sample at a specified rate of temperature increase. The temperature and displacement
are measured throughout the test to generate a strain-temperature curve. The dimensions
of the test specimens were 6.3 mm in height by 12.7 mm in width. The samples were
dried in a vacuum oven at 80°C and 100 kPa pressure for at least one hour before testing
to eliminate any dimensional changes due to moisture effects.

The samples were tested by heating the specimen from room temperature to
140°C at a rate of 4°C/minute. The strain and temperature of the sample were measured
throughout the test by the TMA apparatus. The linear slope of the strain-temperature

curve was reported as the coefficient of thermal expansion.

 

Figure 2.6. Thermal apparatus used for coefficient of thermal expansion tests

2.2.4 Moisture Absorption Testing

The moisture absorption of the material systems was measured using an
environmental chamber (Figure 2.7), which maintains a constant temperature and
humidity. The dimensions of the test specimens were 12.7 mm by 63 mm. The side
edges of the samples, which were exposed when the samples were cut from the
compression molded plate, were sealed with unsaturated polyester resin to be consistent
with the material on the surface of the specimen.

The moisture absorption testing was performed by storing the samples in the
environmental chamber at 30°C and 90% humidity and measuring the increase in weight
of the samples. On the first days of testing, measurements were taken at: 1 hour, 2 hours,
4 hours, 8 hours, 18 hours, 24 hours, and 48 hours. After the initial readings,
measurements were taken once a week until the readings reached a steady state (no
longer absorbing moisture). The percentage of the sample weight absorbed was plotted
against the square root of time to determine when the moisture absorption had reached a
steady state. The percent weight absorbed once the readings had reached steady state was

reported as the degree of moisture absorption.

 

Figure 2.7. Environmental chamber used for moisture absorption testing

17

2.3 Biocomposite Material Systems

The material test program was initiated by identifying potential biocomposite
material systems and performing material testing to evaluate their probable use for load-
bearing structural components. Hemp fiber was determined to be the stiffest, cheapest,
and most readily available fiber type for use in load-bearing structural applications.
Thus, hemp was used as the base natural fiber reinforcement in the biocomposite
materials. Flax fibers were also used for specific applications. Unsaturated polyester
resin was also determined to be the most readily available and cheapest resin type for use
in load-bearing structural applications. Thus, unsaturated polyester was used as the
matrix for the biocomposite and hybrid material systems. The performance of the
biocomposite material systems was compared with that of E—glass fiber reinforced
composite materials.

The fiber content in the material studies was selected to match the fiber content in
the cellular beams and plates manufactured for the structural experimental program
(Chapter 3). In most cases, the fiber weight fraction (ratio of fiber weight to total weight)
was between 10-20%, which corresponds to roughly an 8-16% fiber volume fraction.
The E-glass material system was manufactured with approximately 30% fibers by weight,
which corresponds to roughly a 15% fiber volume fraction. At this fiber weight fraction
the E—glass and biocomposite material systems have approximately the same fiber
volume fraction. Because of the high percentage of resin used in the material systems,
resin was squeezed out of the sample during the pressing of the compression molded
plates. The squeeze out of resin during the pressing of the sample made it difficult to

control the fiber content in the manufactured plates for material characterization. Thus,

18

the fiber content could not exactly match the contents in the cellular beams and plates

manufactured for the structural experimental program.

2.3.1 Materials and Constituents

The biocomposite plates for material characterization were designed to match the
material systems used for the cellular beams and plates for the structural experimental
program (Chapter 3). All biocomposite material systems were composed of either hemp
fibers or chopped flax fibers and unsaturated polyester resin with a fiber weight fraction
of 10-20%, which corresponds to roughly an 8-16% fiber volume fraction. The hemp
fibers (Figure 2.8a) were chopped with an average length of 6mm, and purchased from
BioInnova of Germany. The flax fibers (Figure 2.8b) were chopped with an average
length of 2 mm, contained 50% core material, and were purchased from Flax Craft of
New Jersey. The E-glass fibers were chopped with an average length of 7 mm, sized for
polyester resin, and purchased from Owens Coming of Ohio. The resin system used was
ortho unsaturated polyester (UPE) with methyl ethyl ketone peroxide (MEKP) catalyst
(1% by weight of resin), and cobalt naphthenate (CN) promoter (0.03% by weight of

resin). The resin system composition is show in tabular form in Table 2.1.

    

(a) Hemp fibers (b) Chopped flax, 50% fibers, 50% core

Figure 2.8. Close-up view of natural fibers used for manufacturing

l9

Table 2.1. Unsaturated polyester (UPE) resin system

 

 

Percentage by
Resin System Component Weight of Resin
Ortho Unsaturated Polyester Resin (UPE) --
Methyl Ethyl Ketone Peroxide Catalysis (M EKP) 1%
Cobalt Naphthenate Promoter (CN) 0.03%

 

A summary of the average mechanical properties of the material systems used in
the material and structural (cellular beams and plates) characterization studies is
presented in Table 2.2. The physical characteristics of the fiber types used for each
material system, including fiber weight fraction, fiber volume fraction, fiber length, and

aspect ratio, are provided in Table 2.3.

Table 2.2. Average properties of natural fiber and unsaturated polyester composites

 

 

Fiber Volume Elastic Tensile

Fraction Density Modulus Strength
Fiber type (%) (g/cm“) (GPa) (MPa)
Green Hemp 10 1.18 3.81 14.47
Green Hemp 13 1.19 4.40 14.91
Green Hemp 20 1.22 5.23 16.79
Unprocessed Green Hemp 25 1.14 8.65 10.08
Raw Hemp 15 1.21 6.15 19.49
Flax 35 1.10 3.37 13.31
E-Glass 15 1 .36 7.80 34.70

 

Table 2.3. Physical characteristies of the reinforcing fibers

 

 

Fiber Wt. Fiber Vol. Fiber Fiber

Fiber Type (%) (SAL Length (mm) Ud
Green Hemp 13 10 2- 10 60
Green Hemp 17 13 2 ~10 60
Green Hemp 25 20 2 - 10 60
Unprocessed Green Hemp 25 25 5 - 20 125
Raw Hemp 20 15 8 - 12 100
Flax 33 35 1 - 2 15
E-Glass 29 15 6 - 8 467

 

20

 

2.3.2 Manufacturing

Characterization of the biocomposite material systems was done through ASTM
testing of coupon samples taken from compression molded plates. The typical
manufacturing process of the compression molded plates is summarized in the
photograph sequence shown in Figure 2.9. The manufacturing process began by first
preparing a 200 mm x 200 mm picture frame mold (Figure 2.9a), followed by top and
bottom Teflon coated aluminum sheets, and top and bottom steel plates (Figure 2%).
The fibers (Figure 2.9c) were dried in a vacuum oven at 80°C and 100 kPa of pressure for
at least 10 hours before manufacturing. The polyester resin system (Figure 2.9d) shown
in Table 2.1 was measured and mixed. The resin was then poured on the fibers (Figure
2.9c) and mixed by hand (Figure 2.9f) until the material was consistent. The impregnated
fibers were then placed in the picture frame mold (Figure 2.9g). The natural fibers have a
tendency to clump and tangle together when mixed. Thus, to ensure a uniform sample
care must be taken to evenly distribute the material in the mold.

Once the impregnated fibers were placed in the mold, the top Teflon and steel
plates were placed on top of the material. The entire mold was wrapped with aluminum
foil to catch any resin that squeezed out of the sample during pressing. The sample was
then placed in the press for curing under compression (Figure 2.9h). The sample was
cured under 550 kPa of pressure for a total of 4 hours. During the 4 hour curing period
the following temperatures were used: 100°C for 2 hours, followed by 150°C for 2 hours.
Typical biocomposite compression molded plates and samples for material testing are

shown in Figure 2.10.

21

 

 

(c) Tray with hemp fibers
\. ' , “‘ ’ ‘ L '

 

(g) Impregnated biofibers in mold (h) Compression curing of sample

Figure 2.9. Manufacturing process of compression molded biocomposite plates

22

 

Figure 2.10. Compression molded biocomposite materials

2.3.3 Test Matrix

Material property tests were performed on the following reinforcing fiber types:
green hemp, unprocessed green hemp, raw hemp, flax, and E-glass. Unsaturated
polyester was used as the matrix for all the composite material systems. Each material
type was given a sample identification label (ID) depending on the material and fiber
weight fraction used. The first letter of the ID, P, stands for compression molded plate.
The next two digits of the ID stand for the material type, as shown in Table 2.4. The final
two digits stand for the fiber weight fraction used in the material. The physical

characteristics of the fiber types used for each material system are provided in Table 2.5.

23

Table 2.4. Biocomposite material system naming convention

 

 

ID Label Fiber Type Fiber Physical Description
01 Green Hemp short fibers, low core content
02 Unprocessed Green Hemp long fibers, high core content
03 Raw Hemp long fibers, low core content
04 Flax short fibers, 50% fibers, 50% core
05 E-Glass short chopped fibers, sized for polyester

 

Table 2.5. Physical characteristics of the composite material systems

 

Fiber Wt. Fiber Vol. Density

 

SAMPLE ID (%) (%) (g/cms)
Green Hemp/UPE - 13 wt.% P-01-13% 13 10 1.18
Green Hemp/UPE - 17 wt.% P-01-17% 17 13 1.19
Green Hemp/UPE - 25 wt.% P-01-25% 25 20 1.22
Unprocessed Green Hemp/UPE P-02-25% 25 25 1.14
Raw Hemp/UPE P-03-20% 20 15 1.21
Flax/U PE P-04-33% 33 35 1 .10
E-Glass/UPE P-05-29°/o 29 15 1 .36

 

2.3.4 Tensile Test Results

The tensile stress-strain response of each biocomposite material system was
measured using ASTM D638 and used to determine the modulus of elasticity, ultimate
tensile strength, and the elongation at break. The results for each of these tensile
properties are presented in the following sections. The specific modulus and strength of

the biocomposite material systems are presented at the end of the section.

2.3.4.] Tensile Characteristic Values

A summary of the characteristic tensile properties including: modulus of elasticity
(MOE), ultimate tensile strength, and elongation at break are shown in Table 2.6 for the
biocomposite material systems. The results from each of these tensile properties are

presented in detail and discussed in the following sections

24

Table 2.6. Tensile test results summary for biocomposite material system

 

MOE Tensile Str. Tensile

 

SAMPLE ID (G Pa) (M Pa) Elm (%)
Green Hemp/UPE - 13wt.% P-01-13% 3.81 14.47 0.32
Green Hemp/UPE - 17wt.% P-01-17% 4.40 14.91 0.30
Green Hemp/UPE - 25wt.% P-01-25% 5.23 16.79 0.33
Unprocessed Green Hemp/UPE P-02-25% 8.65 10.08 0.16
Raw Hemp/UPE P-03-20% 6.15 19.49 0.33
Flax/UPE P-04-33% 3.37 13.31 0.33
E-Glass/UPE P-05-29% 7.80 34.70 0.50

 

2.3.4.2 Modulus of Elasticity

The results for the modulus of elasticity are given in Table 2.7 and shown in
Figure 2.11. The test results show that, on average, the modulus of the E-glass material
system was 1.5 times that of the biocomposite material systems (Table 2.7). However,
the unprocessed green hemp material system, with a modulus of 8.65 GPa, showed a
higher value than the E-glass material system, with a modulus of 7.80 GPa (Table 2.7).
The raw hemp material system, with a modulus of 6.80 GPa, showed a modulus
comparable to the E-glass material system. It should be noted that the E-glass fibers were
sized (chemically treated for improved adhesion) for polyester resin (see Table 2.4, page
24), which may lead to improved properties for the E-glass material system.

The results from the green hemp material systems show increased modulus of
elasticity with increase in fiber weight fraction. For example, as the fiber weight fraction
of the green hemp material systems increased from 13% to 25%, the modulus increased
from 3.81 GPa to 5.23 GPa, an increase of 37% (Table 2.7). This trend is expected, as

the fibers are stiffer and stronger than the matrix material.

25

 

Table 2.7. Modulus of elasticity (MOE) test results for biocomposite material systems

 

MOE Std. Dev. Coeff. of

 

SAMPLE ID (GPa) (GPa) Variation
Green Hemp/UPE - 13wt.% P-01-13% 3.81 0.43 11%
Green Hemp/UPE - 17wt.% P-01-17% 4.40 0.41 9%
Green Hemp/UPE - 25wt.% P-01-25% 5.23 0.82 16%
Unprocessed Green Hemp/UPE P-02-25% 8.65 2.80 32%
Raw Hemp/UPE P-03-20% 6.15 1.12 18%
FIax/UPE P-04-33% 3.37 0.24 7%
E-Glass/UPE Pris—29% 7.80 1 .2_9 1 7%

 

 

 

.5 .e
N §

.a
O

MODULUS OF ELASTICITY (GPa)
O O

 

 

Green
Hemp/UPE -
13wt 96
Green
Hemp/UPE -
17M %
Green
Hemp/UPE -
25wt.%
Unprocessed
Green
Hemp/UPE

Raw
Hemp/UPE
Flax/UPE
E-GIaes/UPE

Figure 2.11. Modulus of elasticity (MOE) test results for biocomposite material systems

2.3.4.3 Ultimate Tensile Strength

The results for the ultimate tensile strength are provided in Table 2.8 and shown
in Figure 2.12. Test results show that the strength of the E—glass material system was, on
average, 2.3 times that of the biocomposite material systems (Table 2.8). As stated in the

previous section, the E-glass fibers were sized for polyester resin (Table 2.4, page 24),

26

which may lead to improved composite properties. Despite this fact, the strength of E—
glass fibers are in general much stronger than natural fibers (see Table 1.1, page 4).

The results from the green hemp material systems show increased tensile strength
with increase in fiber weight fraction. For example, as the fiber weight fraction of the
green hemp material systems increased from 13% to 25%, the strength increased from
14.47 MPa to 16.79 MPa, an increase of 16% (Table 2.8). This trend is expected, as the
fibers are stronger than the matrix material.

Results from the tensile strength tests have a high variation for most material
systems, with coefficients of variation (standard deviation divided by mean value)
ranging from 25% to 50% (Table 2.8). The variation in the results may be due to
manufacturing defects on the surface of the compression molded plates, which may lead
to premature tensile failure and cause scatter in the results. Variation may also be due to
impurities, such as the wood like core material present in the natural fibers. These

impurities may act as crack initiators or lead to poor fiber-matrix adhesion.

Table 2.8. Ultimate tensile strength (UTS) test results for biocomposite material systems

 

Tensile Str. Std. Dev. Coeff. of

 

SAMPLE ID (M Pa) (M Pa) Variation
Green Hemp/UPE - 13wt.% P-01-13% 14.47 5.74 40%
Green Hemp/UPE - 17wt.% P-01-17% 14.91 5.67 38%
Green Hemp/UPE - 25wt.% P-01-25% 16.79 4.19 25%
Unprocessed Green Hemp/UPE lP-02-25% 10.08 5.89 58%
Raw Hemp/UPE P-03-20% 19.49 3.10 16%
FIax/UPE P-04-33% 1 3.31 2.63 20%
E-GIass/U PE P-05-29% 34.70 1 7.49 50%

 

27

 

l

 

11111
I

LIL ll

8

initniirtii
l

TENSILE STRENGTH (MPa)
3 8

 

 

 

O
L

Green
Hemp/UPS -
13wt.%
Green
Hemp/UPE -
17wt.%

Green
Hemp/UPE -
25M 96

it
no
3

Figure 2.12. Ultimate tensile strength (UTS) test results for biocomposite material system

Hemp/UPE
Raw
Hemp/UPE
Flax/UPS
E-GlasslUPE

2.3.4.4 Elongation at Break

Test results for the tensile elongation at break are provided in Table 2.9 and
shown in Figure 2.13. The results from the elongation at break have a high variation for
most material systems, with coefficients of variation ranging from 20-75% (Table 2.9).
As with the ultimate tensile strength results, that variation in results may be due to
manufacturing defects on the surface of the compression molded plates, which may lead
to premature tensile failure and cause scatter in the results.

The test results show that the elongation at break for the E-glass material system
was, on average, 1.5 times that of the biocomposite material systems (Table 2.9). The
tensile elongation at break of the biocomposite material systems was very similar, with an

average elongation at break of 0.32% (Table 2.9). However, the ultimate strain of the

28

 

unprocessed green hemp material system was only half that of the other biocomposite

material systems (Table 2.9).

Table 2.9. Tensile elongation at break test results for biocomposite material systems

 

 

 

 

Tensile Std. Dev. Coeff. of
SAMPLE ID Elong(%) (%) Variation
Green Hemp/UPE - 13wt.% P-01-13% 0.32 0.16 50%
Green Hemp/UPE - 17wt.% P-01-17% 0.30 0.15 49%
Green Hemp/UPE - 25wt.% P—01-25% 0.33 0.07 22%
Unprocessed Green Hemp/UPE P-02-25% 0.16 0.12 75%
Raw Hemp/UPE P-03-20% 0.33 0.07 20%
Flax/UPE P-04-33% 0.33 0.16 47%
E-G lass/U PE P-05-29% 0.50 0.21 41 %
0.80
0.70

§

TENSILE ELONG (%)
O O
8 s

O
L.
O

 

 

 

 

0.00
r“- it it: i E 5% if “
0%; (SEE egg; g; g E g

 

 

Figure 2.13. Tensile elongation at break test results for biocomposite material systems

2.3.4.5 Specific Tensile Properties

The specific tensile stiffness is provided in Table 2.10 and show in Figure 2.14,

while the specific tensile strength is proved in Table 2.11 and shown in Figure 2.15.

29

Both the specific stiffness and strength show similar trends to the modulus of elasticity
(Figure 2.11) and ultimate tensile strength (Figure 2.12). Because of the low fiber weight
(and volume) fraction used for the E-glass material system, the density of the material is
relatively low. Thus, the specific properties of the biocomposite material systems do not
compare well with the tested E—glass material systems. If higher volume fractions were
used for both the E-glass and biocomposite samples, improvement in specific properties

of the biocomposite materials would have been more noticeable.

Table 2.10. Specific tensile stiffness of biocomposite material systems

 

 

Density MOE E/p
SAMPLE ID (gicm3) (G Pa) [GPa/(gismafl

Green Hemp/UPE - 13wt.% P-01-13% 1.18 3.81 3.24
Green Hemp/UPE - 17wt.% P-01-17% 1.19 4.40 3.69
Green Hemp/UPE - 25wt.% P-01-25% 1.22 5.23 4.28
Unprocessed Green Hemp/UPE P-02-25% 1.14 8.65 7.59
Raw Hemp/UPE P-03-20% 1.21 6.15 5.07
Flax/UPE P-04-33% 1.10 3.37 3.06
E-GIass/UPE P-05-2_9% 1 .36 7.80 5.74

 

 

N

h

60

M

d

    

SPECIFIC srrrruess [GPa/(grem’n

 

,3

O 3
III III III '8 .0 li’ E 'i’

:13! ell-39 cit-32 ":0- g

1%: ii: iii §§§ 52% E 5
GEfi 0E5: 0&3 55E 0 u. 2
o o o a. e z 0
a: I: 1: g :I: “I,

Figure 2.14. Specific tensile stiffness of biocomposite material system

30

Table 2.11. Specific tensile strength of biocomposite materials systems

 

 

Density Tensile Str. o/p
SAMPLE ID (gicma) (MPa) [MPa/(gicmsfl
Green Hemp/UPE - 13wt.% P-01-13% 1.18 14.47 12.29
Green Hemp/UPE -17wt.% P-01-17% 1.19 14.91 12.51
Green Hemp/UPE - 25wt.% P-01-25% 1.22 16.79 13.73
Unprocessed Green Hemp/UPE P-02-25% 1 .14 1 0.08 8.85
Raw Hemp/UPE P-03-20% 1 .21 19.49 16.07
Flax/UPE P-04-33% 1.10 13.31 12.06
E-Glass/UPE P-0L@% 1 .36 34.70 25.51

 

 

 

SPECIFIC STRENGTH [MPa/(glcm°)]

 

s= :3

Green
Hemp/UPE -
13M %
Gm" ,1,“ w ”I" . ‘, 'Ipiffir
Hemp/UPE - .
17m 15 ‘
Green
Hemp/UPE -
25M 96

Unprocessed
Green

Figure 2.15. Specific tensile strength of biocomposite material systems

2.3.5 Impact Test Results

Results from the impact strength testing are provided in Table 2.12 and shown in
Figure 2.16. The test data shows that, on average, the E-glass material system had an
impact strength 5.6 times that of the bio composite material systems (Table 2.12).

The test results, excluding the E-glass material system, are shown in Figure 2.17

to better compare the results from the biocomposite materials. The results from the

31

biocomposite material systems show increased impact strength with increase in fiber
weight fraction. For example, as the fiber weight fraction of the green hemp material
systems increased from 13% to 25%, the impact strength increased from 6.86 J/m to
10.40 J/m, an increase of 52% (Table 2.12). This trend of increasing impact strength
with increased fiber weight fraction is expected as the fibers provide more impact
strength than the matrix material. The flax biocomposite material system showed the
lowest impact strength, at 5.28 J/m (Table 2.12), which may be due to the short fiber

length and the 50% core material contained with the flax fibers.

Table 2.12. Impact strength (Izod Test) test results for biocomposite material systems

 

Impact Str. Std. Dev. Coeff. of

 

SAMPLE ID (J/m) (J/mL Variation
Green Hemp/UPE - 13wt.% P-01-13% 6.86 1.88 27%
Green Hemp/UPE - 17wt.% P-01-17% 7.54 2.06 27%
Green Hemp/UPE - 25wt.% P-01-25% 10.40 2.07 20%
Unprocessed Green Hemp/UPE P-02-25% 17.45 3.41 20%
Raw Hemp/UPE P-03-20% 13.52 5.01 37%
Flax/UPE P-04-33% 5.28 0.76 14%
E-GIass/UPE P-05-29% 67.29 7.55 1 1%

 

32

 

 

w m w w .o. m m m
E... :Pozmmpm 55......

 

,. mtg-020$

, mug—u.

mask—Eu:
3am

menace:
:85

, 33803::

«Quinn

, ,, . maniac:

c026

$.Ew

, . washes.

.520

$.52
. manage...
:85

Figure 2.16. Impact strength (Izod Test) test results for biocomposite material systems

 

 

3,, manna“.

,, ,_ . ,._ maniac:

in:

maniac:
50.6

,,, toaeoooaca

o\o.<>mN

., _ , . maniac:

c026

$.52
- maniac:
:35

$.36 w

, . maniac:

A..............u..«Ji<J.

m e m s
.25 :hoszm 55......

 

520

Figure 2.17. Impact strength (Izod Test) test results for biocomposite materials only

33

2.3.6 Thermal Expansion Test Results

Results from the coefficient of thermal expansion (CTE) tests are provided in
Table 2.13 and shown graphically in Figure 2.18. The results from the CTE tests show
that the biocomposite material systems, on average, have a CTE 30% higher than the E-
glass material system (Table 2.13). In addition, the results from the green hemp material
systems show that the CTE decreases as the fiber weight fraction increases (Table 2.13).
For example, as the fiber weight fraction of the green hemp material systems increased
from 13% to 25%, the CTE decreased from 51.0 um/m°C to 34.1 um/m°C, a decrease of
58% (Table 2.13). This trend verifies that the CTE of I the matrix material is higher than
that of the natural fibers, thus, the CTE of the composite material system can be reduced

by increasing the content of natural fibers.

Table 2.13. Coefficient of therml expansion (CTE) test results for biocomposite material systems

 

CTE Std. Dev. Coeff. of

 

SAMPLE ID (um/m°C) (um/m°C) Variation
Green Hemp/UPE - 13wt.% P-01-13% 59.0 10.7 18%
Green Hemp/UPE - 17wt.% P-01-17% 45.0 7.2 16%
Green Hemp/UPE - 25wt.% P-01-25% 34.1 3.4 10%
Unprocessed Green Hemp/UPE P-02-25% 35.2 3.4 10%
Raw Hemp/UPE P-03-20% 23.2 2.8 12%
Flax/UPE P-04-33% 67.6 10.8 16%
E-Glass/UPE P-05-29% 30.7 8.8 29%

 

34

 

 

 

Green 1 M
Hemp/UPE- j'i ~

13wt.%

Green
Hemp/UPE- .1};

17wt.%

Green
Hemp/UPE- I

25wt% '

Green

Hemp/UPE
Raw
Hemp/UPE I
Flax/UPE
E-GlaseIUPE

Unprocessed m A ;

Figure 2.18. Coefficient of thermal expansion (CTE) test results for biocomposite material systems

The CTE of the green hemp material system with a fiber weight fraction of 25%
is compared with conventional building materials in Figure 2.19. The comparison shows
that the CTE of the biocomposite material is approximately 3 times larger than that of
steel, and concrete (Figure 2.19). This difference in CTE of the biocomposite material
system with conventional construction materials, and the resulting differential thermal
effects, must be considered if biocomposite materials are used with conventional building

materials for structural applications.

35

 

 

 

 

Steel
Aluminum
Douglas Fur

(perpendicular
to the grain)

Douglas Fur
(parallel to the
grain)

I. r

- 3

£33 E; s
0

Figure 2.19. Comparison of Coefficient of thermal expansion (CTE) with conventional materials

2.3. 7 Moisture Absorption Test Results

The moisture absorption test results are given in Table 2.14 and shown in Figure
2.20. The moisture absorbed by the material systems was measured over a period of 50
days as show in the plot of percent weight absorbed versus the square root of time in
Figure 2.21. The moisture absorption of all material systems stabilized after
approximately 33 days (800 hours) in the environmental chamber.

Results from the moisture absorption test show that the biocomposite material
systems, on average, absorbed 4 times as much moisture as the E-glass material system
(Table 2.14). In addition, the results from the green hemp material systems show that
moisture absorption increases as the fiber weight fraction increases (Table 2.14). For
example, as the fiber weight fraction of the green hemp material systems increased from

13% to 25%, the moisture absorption increased from 0.96% to 1.31%, an increase of 36%

36

 

(Table 2.14). This trend is expected as the fibers absorb more moisture than the matrix
material. The trend of increasing moisture absorption with increasing fiber weight
fraction is also seen with other material systems. For example, the raw hemp material
with a fiber weight fraction of 20% absorbed 1.37% moisture, while the unprocessed
green hemp material with a fiber weight fraction of 25% absorbed 1.44%, and the flax
material with a fiber weight fraction of 33% absorbed 1.71% (Table 2.14).

The results from the moisture absorption test also show that the biocomposite
material systems absorb less moisture than the natural fiber constituents. The hemp
biocomposite material systems (green, unprocessed, and raw) absorbed between 0.96% to
1.44% moisture, while the hemp fibers alone absorb 8% moisture (see Table 1.1). The
flax biocomposite material system absorbed 1.71% moisture, while the flax fibers alone
absorb 7% moisture (see Table 1.1). This trend suggests that the matrix material acts as a
moisture barrier and helps reduce the moisture absorbed by the composite material

system.

Table 2.14. Moisture absorption test results for biocomposite material systems

 

 

 

Moisture

Absorption Std. Dev. Coeff. of

SAMPLE ID (%) (%) Variation
Green Hemp/UPE - 13wt.% P-01-13% 0.96% 0.03% 3%
Green Hemp/UPE - 17wt.% P-01-17% 1.05% 0.06% 6%
Green Hemp/UPE - 25wt.% P-01-25% 1.31% 0.04% 3%
Unprocessed Green Hemp/UPE P-02-25% 1.44% 0.08% 6%
Raw Hemp/UPE P-03-20% 1.37% 0.21% 15%
Flax/UPE P-04-33% 1 .71 % 0.09% 5%
E-Glass/UPE P-05-29% 0.34% 0.02% 6%

 

37

 

MOISTURE ABSORPTION (%)

 

 

Cale? ails! 92$ cg: :2 E g

Figure 2.20. Moisture absorption test results for biocomposite material system

2.5%

 

 

+ P-o1-13% - Green Hemp/UPE Test Conditions:

: —a— P-01-17% - Green Hemp/UPE Temperature = 30°C
. —a~— P-o1-25% - Green Hemp/UPE Humidity = 90%

- —e— P-02-25% - Unprocessed GHIUPE
‘ + Poe-20% - Raw HempIUPE Flax
: —l— Pot-33% - FIax/UPE
_ —+— P-05-29% - Glass/UPE

g

    

 

 

z';
a!

Percent oi Weight Absorbed (%)
3

 

 

o.o%-' - .. .....
1 e 11 1e 21 26 31 as

Square Root of Time (hrm’)

 

Figure 2.21. Percent weight absorbed versus square root of time for biocomposite material system

38

2.4 Hybrid Material System

Improvement of the structural performance of biocomposite material systems and
cellular structures through the use of hybrid material systems was investigated at both the
material and structural level. A hybrid material is defined here as combination of
chopped natural fibers with synthetic or natural fiber fabrics. In the hybrid material
system the chopped natural fibers are used as the core material, sandwiched between face
sheets of synthetic or natural fiber fabrics (Figure 2.22). These hybrid material systems
could be used as face sheet, or cell wall materials in cellular beams and plates. The use
of fabric face sheets also improves the ability to handle the natural fibers during
manufacturing. Thus, the hybrid material systems are ideal for use as pro-forms for

automated manufacturing.

Fabric Face A7 , _ :1? , an, ,
Sheets ”5.5:; 7'" 4
II, 7 521/ //A;':D:

’5’; é%/; / .. ”
’sz” *— —v"»,é‘/—//y(4’ %.'
9, ‘3' Q g. VI! ff‘
4 ’ %_ "f’ /;’5’ ’
m Lcooose F. bef

”’H”

 

Figure 2.22. Hybrid material system

In addition, single composite laminates using chopped strand E-glass,
unidirectional carbon, and woven jute mats were investigated to provide the necessary
material characterization for the face sheet materials used in the structural experimental

studies (Chapter 3). These material systems were only tested for tensile properties.

39

2.4.1 Materials and Constituents

Two types of fiber fabrics (Figure 2.23) were used for the hybrid material
systems: (a) randomly oriented glass, and (b) woven jute. These fabrics, along with
unidirectional carbon (Figure 2.23c), were considered as face sheet materials for hybrid
cellular plates (see Chapter 3). Engineered hybrid fabric/biofiber blends for the hybrid
material systems consisted of: (a) top and bottom chopped glass strand mat face sheets
with raw hemp fiber core, and (b) top and bottom woven jute fabric with a raw hemp
fiber core. The woven jute hybrid composite material contained 60% core fibers by total
fiber weight in the sample, while the glass strand mat hybrid material contained 51% core
fibers by total fiber weight. Both hybrid materials were manufactured with a fiber weight
fraction of 15%. Due to resin squeeze-out during pressing, the actual fiber weight
fraction was higher than 15%. The measured fiber weight fraction of the jute hybrid
material was 27%, while the glass strand mat hybrid material was 31% (Table 2.15). The
resin system used was unsaturated polyester (UPE), as shown in Table 2.1.

The composite systems for single natural and synthetic fabrics (Figure 2.23) were
composed of one layer of fabric material and unsaturated polyester (UPE). The fabric
material systems were manufactured with a fiber weight fraction of 30% to match the
estimated face sheet material composition used in the structural experimental studies

(Chapter 3).

40

         

(a) Woven jute (b) Randomly orientated glass (c) Unidirectional carbon

Figure 2.23. Fiber mat fabrirs used as face sheets for hybrid material system

Table 2.15. Average properties of the hybrid material system with biocomposite and E-glass

 

 

material system
Fiber Weight Density MOE Tensile Str.
SAMPLE Fraction (gloma) (GPa) (MPa)
Green Hemp/UPE T&B Jute 27% 1.17 10.78 28.19
Green Hemp/UPE T&B Glass 31% 1.29 11.92 43.37
Green Hemp 25% 1.22 5.23 16.79
Eglass a 1.3_§ 7.8 34.7

 

2.4.2 Manufacturing

The characterization of the hybrid material systems was done by manufacturing
compression molded plates of sandwiched material systems as well as composite systems
of the individual natural and synthetic fabrics. The typical manufacturing process of the
hybrid compression molded plates is summarized in the photograph sequence shown in
Figure 2.24. The manufacturing of the hybrid material systems was very similar to the
biocomposite compression molded plates (Figure 2.9) with the only difference being that
face sheets were used for the hybrid material systems. The face sheets were added to the
system by placing a bottom layer of fabric (Figure 2.24d), followed by the impregnated

chopped hemp fiber core (Figure 2.24e), and then the top face sheet (Figure 2240. The

41

face sheets were wet out with resin by pouring and spreading resin over the fabric instead
of the hand mixing used to wet out the loose natural fibers.

The manufacturing of the fabric material systems was also very similar to the
biocomposite compression molded plates (Figure 2.9), with the only differences being the
procedure for mixing the fibers and resin and the size of the picture frame mold. The
fibers and resin were mixed by simply pouring and spreading the resin over the fibers
(Figure 2.24d). A smaller 178mm x 127 mm picture frame mold was used because the
fabric material systems were only tested in tension. Thus, a smaller amount of material
was required for the test specimens.

All the hybrid material system plates were cured under 550 kPa of pressure for a
total of 4 hours. During the 4 hours the following temperatures were used: 100°C for 2
hours, followed by 150°C for 2 hours.

The two hybrid compression molded plates manufactured for the experimental
program are shown in Figure 2.25. The compression molded plates of the fiber mat

materials are shown in Figure 2.26.

42

         
  

 

(b) Tray with hemp fibers

  

(1) Wet out of bottom mat layer

   

(g) Hybrid fabric/biofibcr blend (h) Compression curing of sample

Figure 2.24. Manufacturing process of compression molded hybrid plates

43

  
  

   

(a) Hemp core with jute fabric face sheets (b) Hemp core with glass strand mat face sheets

Figure 2.25. Hybrid compression molded plates

   

(a) Unidirectional Carbon (b) Randomly orientated E-glass (c) Woven jute

Figure 2.26. Compression molded plates of fiber mat materials and polyester resin

2.4.3 Test Matrix

Each material type tested was given a sample identification label (ID) depending
on the material type and the fiber weight fraction used as shown in Table 2.16. The first
letter of the ID, P, stands for compression molded plate. The next two digits of the 1])
stand for the material type. The final two digits stand for the sample number of the
particular material type. The results from the previous green hemp and E-glass material

tests are used for comparison with the results from the hybrid plates throughout this

discussion. The green hemp and E-glass materials used for comparison have similar fiber
weight fractions to the hybrid plates. The fiber mat material systems were only tested for

tensile properties.

Table 2.16. Hybrid material system plate naming convention

 

 

ID Label Plate Type Constituents
11 Hybrid plate Top and bottom glass face sheets
With green hemp core
31 Hybrid plate Top and bottom jute fabric face sheets

With green hemp core
10 Randomly orientated glass strand mat Randomly orientated glass strand mat
20 Unidirectional carbon fiber mat Unidirectional carbon with nylon stitching
30 Woven jute fabric mat Woven jute fabric

2.4.4 Tensile Test Results

The tensile stress-strain response of each hybrid material system was measured
using ASTM D638 and used to determine the modulus of elasticity, ultimate tensile
strength, and the elongation at break. The results for each of these tensile properties are
presented in the following sections. The specific modulus and strength of the hybrid

material systems are presented at the end of the section.

2.4.4.1 Tensile Characteristic Values

A summary of the characteristic tensile properties for the hybrid material systems
including modulus of elasticity (MOE), ultimate tensile strength, and elongation at break
are shown in Table 2.17. The results from each of these tensile properties are presented

in detail and discussed in the following sections.

45

Table 2.17. Tensile test results summary for hybrid material system

 

MOE Tensile Str. Tensile

 

SAMPLE ID (G Pa) (M Pa) Elong (%)
Green Hemp/UPE T&B Jute - 27 wt.% P-31-02 10.78 28.19 0.39
Green Hemp/UPE T&B Glass - 31 wt.% P-11-02 11.92 43.37 0.61
Green Hemp/UPE - 25 wt.% P-01-25% 5.23 16.79 0.33
E-Glass/UPE - 29 wt.% P-05-29% 7.80 34.70 0.50
Glass Strand Mat/UPE P-10-01 13.46 67.02 0.91
Unidirectional Carbon/UPE P-20-01 37.92 585.12 1.67
Woven Jute/UPE P-30-01 4.03 22.01 0.68

 

2.4.4.2 Modulus of Elasticity

Results for the modulus of elasticity are given in Table 2.18 and shown in Figure
2.27 and Figure 2.28. The addition of the fabric face sheets in the hybrid material
systems clearly improved the modulus of elasticity over the solid material systems
(Figure 2.27). By adding jute fabric face sheets to the hemp core the modulus was
increased by 106% over the all hemp material, and by adding glass fabric face sheets the
modulus was increased by 128% over the all-hemp sample (Table 2.18). The glass
hybrid material system also showed an improvement in the modulus of 53% over the all-
glass material (Table 2.18). Despite these improvements in modulus, the hybrid material
systems have more deviation in the results compared to the all-biocomposite material
plates (Table 2.18). This may be due to voids in the green hemp core material from non-
uniform placement of the material during manufacturing.

The unidirectional carbon fabric material system showed by far the highest
modulus of the face sheet fabric material systems (Figure 2.28). The modulus of the
carbon fabric material system was 2.8 times the modulus of the glass fabric material, and

9.4 times the modulus of the jute fabric material (Table 2.18).

46

Table 2.18. Modulus of elasticity (MOE) test results for hybrid material systems

 

MOE Std. Dev. Coeff. of

 

SAMPLE ID (G Pa) (G Pa) Variation
Green Hemp/UPE T&B Jute - 27 wt.% P-31-02 10.78 6.12 57%
Green Hemp/UPE T&B Glass - 31 wt.% P-11-02 11.92 7.30 61%
Green Hemp/UPE - 25 wt.% P-01-25% 5.23 0.82 16%
E-Glass/UPE - 29 wt.% P-05-29% 7.80 1.30 17%
Glass Strand Mat/UPE P-10-01 13.46 10.74 80%
Unidirectional Carbon/UPE P-20-01 37.92 3.69 10%
Woven Jute/UPE P-30-01 4.03 0.64 16%

 

 

N
01

3
A A A #4 A A A A L A
U I

uoouws or ELASTICITY (GPa)
3

 

 

 

 

 

5 i-
0
Green HemlePE- Green l-lempllPE- E-GlasellPE- Green l-lempllPE-
25wt.% T&B Jute Mat - 29wt.% T&B Glace Mat -
27wt.% 31wt.%

Figure 2.27. Modulus of elasticity (MOE) test results for hybrid material system

47

 

 

 

MODULUS OF ELASTICITY (GPa)

 

 

Glace Strand Mat/UPE - Unidirectional Carbon Woven Jute Mat/UPE -
23M.% Mat/UPE - 23M.% 31 wt.%

Figure 2.28. Modulus of elasticity (MOE) test results for face sheet material system

2.4.4.3 Ultimate Tensile Strength

The results for the ultimate tensile strength are provided in Table 2.19 and shown
in Figure 2.29 and Figure 2.30. When loaded up to failure in tension the two hybrid
material systems failed in different manners. The jute hybrid material system failed in a
sudden brittle manner, with complete fracture of the material and little noise from fiber
cracking before failure. The glass hybrid material system also failed in a brittle manner,
but with partial fracture of the material and fiber cracking noise before failure. At failure,
the hemp core fractured, while the glass fibers in the face sheets had began to pull out of
the matrix but did not completely fracture.

The face sheet fabric material systems also failed in different manners under
tensile loading. Both the glass and jute fabric material systems failed in a sudden brittle

mode with noise from fiber cracking near failure. The unidirectional carbon material

48

system failed in a gradual manner, with considerable matrix cracking and debonding near
failure. The matrix cracking was accompanied by loud cracking noises and an explosive
noise at failure.

The addition of the fabric face sheets in the hybrid material systems improved the
ultimate tensile strength over the solid material systems. By adding jute fabric face
sheets to the hemp core the tensile strength was increased by 68% over the all-hemp
material, and by adding glass fabric face sheets the tensile strength was increased by
158% over the all-hemp sample (Table 2.19). The glass hybrid material system also

showed an improvement in tensile strength of 25% over the all-glass material (Table

 

 

2.19).
Table 2.19. Ultimate tensile strength (UTS) test results for hybrid material system
Tensile Str. Std. Dev. Coefl. of
SAMPLE ID (M Pa) (M Pa) Variation
Green Hemp/UPE T&B Jute - 27% wt. P-31-02 28.20 4.5 16%
Green Hemp/UPE T&B Glass - 31% wt. P-11-02 43.4 3.3 8%
Green Hemp/UPE - 25 wt.% P-01-25% 16.79 4.19 25%
E-GIass/UPE - 29 wt.% P-05-29% 34.70 17.49 50%
Glass Strand Mat/UPE P-10-01 67.0 12.2 18%
Unidirectional Carbon/UPE P-20-01 585.1 152.8 26%
Woven Jute/UPE P-30-01 22.0 1.3 6%

 

49

 

14_.r_.

r I

Jllll

TENSILE STRENGTH (MPa)
3 8

 

 

 

1o
o 2
Green Hemp/UPE - Green Hemp/UPE - E-GlaeelUPE - Green Hemp/UPS -
25wt.% T&B Jute Mat - 29mm, T&B Glass Mat -
27wt.% 31m.%

Figure 2.29. Ultimate tensile strength (UTS) test results for hybrid material system

 

TENSILE STRENGTH (MPa)

 

 

 

Glass Strand Mat/UPE - Unidirectional Carbon Woven Jute Mat/UPE -
3wt.% Mat/UPE - 23m.% 31 wt.%

Figure 2.30. Ultimate tensile strength (UTS) test results for face sheet material system

50

2.4.4.4 Elongation at Break

Results for the tensile elongation at break are provided in Table 2.20 and shown
in Figure 2.31 and Figure 2.32. The hybrid material systems showed improvement in
elongation at break compared to the hemp and glass material systems, with trends similar
to the results shown for the modulus of elasticity and ultimate tensile strength. The
increase in both the elongation at break and tensile strength of the hybrid material
systems may be attributable to the face sheet materials bridging and preventing cracks in
the hemp core material, which lead to sudden crack propagation and failure. The
elongation at break of the hybrid materials show a high variation (Table 2.20), which, as

previously described, may be due to manufacturing defects in the samples.

Table 2.20. Tensile elongation at break test results for hybrid material system

 

Tensile Std. Dev. Coeff. of

 

SAMPLE ID Elong (‘73 (%) Variation

Green Hemp/UPE T&B Jute - 27% wt. P-31-02 0.39 0.24 60%
Green Hemp/UPE T&B Glass - 31 % wt. P-11-02 0.61 0.32 53%
Green Hemp/UPE - 25 wt.% P-01-25% 0.33 0.07 22%
E-Glass/UPE - 29 wt.% P-05-29% 0.50 0.21 41%
Glass Strand Mat/UPE P-10-01 0.91 0.60 66%
Unidirectional Carbon/UPE P-20-01 1 .67 0.1 1 6%

Woven Jute/UPE P-30-01 0.68 0.07 11%

 

51

 

   

Green Hemp/UPE - Green Hemp/UPE - E-GlaeelUPE - Green Hemp/UPE -
25wt.% T&B Jute Mat — ”wt.% T&B Glace Met -
27wt.% 31wt.%

Figure 2.31. Tensile elongation at break test results for hybrid material system

 

 

 

 

     

2.00
1.80
1.60
A
55 1.40
2
1 .20
3
m 1.00
Ill
=| 0.80
9
m 0.60
l-
0.40
0.20
Glass Strand Mat/UPE - Unidirectional Carbon Woven Jute Mat/UPE -
23m.% Mat/UPE - 23wt.% 31 wt.%

Figure 2.32. Tensile elongation at break test results for face sheet material system

 

2.4.4.5 Specific Tensile Properties

Specific tensile stiffness values are provided in Table 2.21 and shown graphically
in Figure 2.33, while specific tensile strengths are proved in Table 2.22 and shown in
Figure 2.34. The hybrid material systems showed higher specific stiffness and specific
strength (Table 2.21 and Table 2.22) with trends similar to the results shown for the
modulus of elasticity and ultimate tensile strength. Comparing the two hybrid material
systems, the specific strength of the glass/hemp material system was higher than that of
the jute material system (Table 2.22). However, the specific stiffness of the jute hybrid
material system was equal to the specific stiffness of the glass hybrid material system
(Table 2.21). This shows the improved performance that can be achieved with the use of

natural fiber reinforcement in place of synthetic fiber reinforcement.

Table 2.21. Specific tensile stiffness of hybrid material system

 

 

Density MOE E/p

SAMPLE ID (g/cma) (G Pa) [G Pa/(g/cm3n
Green Hemp/UPE T&B Jute - 27 wt.% P-31-02 1.17 10.78 9.22
Green Hemp/UPE T&B Glass - 31 wt.% P-11-02 1.29 11.92 9.21
Green Hemp/UPE - 25 wt.% P-01-25% 1.22 5.23 4.28
E-Glass/UPE - 29 wt.% P-05-29°/o 1.36 7.80 5.74

 

53

 

.A
MflbUlO‘lOOO

SPECIFIC STIFFNESS [GPa/(g/cm‘n

d

0 ,
Green Hemp/UPE - Green Hemp/UPE - E-GiaeeIUPE -
25wt.% T&B Jute Mat -
27wt.%

29Wt.%

    

 

Green Hemp/UPE -
T&B Glass Met -

31m.%

Figure 2.33. Specific tensile stiffness of hybrid material system

Table 2.22. Specific tensile strength of hybrid materials system

 

 

Density Tensile Str. a/p
SAMPLE ID cm3 MPa MPa/ cm
Green Hemp/UPE T&B Jute - 27 wt.% P-31-02 1.17 28.19 24.11
Green Hemp/UPE T&B Glass - 31 wt.% P-11-02 1.29 43.37 33.51
Green Hemp/UPE - 25 wt.% P-01-25% 1.22 16.79 13.73
E-Glass/UPE - 29 wt.% P-§2_9% 1.3g 34.70 2_5.51

 

54

 

88881888

  

SPECIFIC STRENGTH [MPa/(g/cni’n

   

 

5
o . , r ' .
Green Hemp/UPE - Green Hemp/UPE - E-GleeelUPE - Green Hemp/UPE -
25wt.% T&B Jute Mat - 29wt.% T&B Glace Mat -
27wt.% 31wt.%

Figure 2.34. Specific tensile strength of hybrid materials system

2.4.5 Impact Test Results

Impact strength test results are provided in Table 2.23 and shown in Figure 2.35.
The hybrid material systems showed improved impact strength over the all—hemp and all-
glass material systems (Table 2.23). By adding jute fabric face sheets to the hemp core
the impact strength was increased by 9% over the all-hemp material, while the impact
strength was increased by 86% by adding glass fabric face sheets (Table 2.23). The
glass/hemp hybrid material system also showed an improvement in tensile strength of 7%
over the all-glass material (Table 2.23). The increase in impact strength of the glass
hybrid material system over the all-glass material system may be due to the longer length

of fibers used in the glass strand fabric.

55

Table 2.23. Impact strength (notched Izod) test results for hybrid material system

 

Impact Str. Std. Dev. Coeff. of

 

SAMPLE ID (J/_m) (J/m) Variation
Green Hemp/UPE T&B Jute - 27% wt. P-31-02 11.28 2.57 23%
Green Hemp/UPE T&B Glass - 31% wt. P-11-02 72.11 7.82 11%
Green Hemp/UPE 25 wt.% P-01-25% 10.40 2.07 20%
E-GIass/UPE 29 wt.% P-05-§l% 67.29 7.55 11%

 

 

40
30
20

IMPACT STRENGTH (Jlm)

10

    

 

Green l-lempIUPE - G'een HempIUPE - E—Gaee/UPE - G-een Hemp/UPE -
25wt.% T&B Jute Mat - 29wt.% T&B Gees Mat -
27m.% 31 wt.%

Figure 2.35. Impact strength (notched Izod) test results for hybrid material system

2.4.6 Thermal Expansion Test Results

Results from the coefficient of thermal expansion (CTE) tests are provided in
Table 2.24 and shown in Figure 2.36. In general, the hybrid material systems showed
little or no difference in CTE compared to the all-hemp and all-glass material systems.
The jute hybrid material systems show no difference in the CTE compared to the all-
hemp material system (Table 2.24), which may be due to the hemp fibers and jute fibers

having similar CTE values. The glass hybrid material system showed an increase in CTE

56

over the all-glass material system (10%, Table 2.24), which may be due to the core hemp

fibers having a larger CTE than the glass fabric fibers.

Table 2.24. Coefficient of thermal expansion (CTE) test results for hybrid material system

 

CTE Std. Dev. Coeff. of

SAMPLE ID (Elm?) (w/m°Cj Variation
Green Hemp/UPE T&B Jute - 27% wt. P-31-02 33.38 4.86 15%
Green Hemp/UPE T&B Glass - 31% wt. P-11-02 33.93 8.15 24%

 

 

Green Hemp/UPE - 25% wt. P-01-25% 34.10 3.37 10%
E-Glass/UPE - 29% wt. P-05-2_9£ 30.73 8.81 E6—
45
4o 3
35 I
30 :.

CTE (pm/m°C)
s a

 

 

 

15 _~
10 >
5 L
0
Green l-iempllPE- Green HempIIPE- E-Gleee/UPE- Green Hemp/LPE-
25wt.% T&B Jute Mat - 29wt.% T&B Glaee Mat -
27wt.% 31wt.%

Figure 2.36. Coefficient of thermal expansion (CTE) test results for hybrid material system

2.4.7 Moisture Absorption Test Results

Moisture absorption test results are given in Table 2.25 and shown in Figure 2.27.
The moisture absorbed by the material system was measured over a period of 50 days, as

shown in the plot of percent weight absorbed versus the square root of time in Figure

57

2.38. The moisture absorption of all material systems stabilized after approximately 30
days (675 hours) in the environmental chamber.

Moisture absorption for the jute hybrid material system was 1.25%, while the
absorption of the all-hemp material system was 1.31% (Table 2.25). Thus, the results
from the absorption tests show that the jute hybrid material system had little effect on
moisture absorption (Figure 2.27). However, the glass hybrid material system showed a
much lower moisture absorption than the all-hemp material system (Figure 2.27). The
moisture absorption of the glass hybrid material system was 0.41%, while the all-hemp
material system was 1.31% (Table 2.25). Results for the glass hybrid material were more
comparable to the all-glass material system with a moisture absorption of 0.34% (Table
2.25). It is expected that the glass hybrid material system would have a slightly larger
moisture absorption than the all-glass material system since the hemp fibers used in the
hybrid system are more susceptible to moisture absorption. The results suggest that the
use of glass-fiber face sheets, which have higher resistance to moisture absorption, act as
a barrier for the hemp fibers in the core the hybrid material system. This hybrid
arrangement of material is thus efficient for reducing moisture absorption in the

biocomposite core material system.

Table 2.25. Moisture absorption test results for hybrid material systems

 

Moisture
Absorption Std. Dev. Coeff. of
SAMPLE ID (%) (%) Variation

 

Green Hemp/UPE T&B Jute - 27wt.% P-31-02 1.25% 0.06% 5%
Green Hemp/UPE T&B Glass - 31wt.% P-11-02 0.42% 0.05% 12%
Green Hemp/UPE - 25wt.% P-01-25% 1.31% 0.04% 3%
E-Glass/UPE - 29wt.% P-05-29% 0.34% 0.02% 6%

58

MOIFS’TUILE ABPSORETION (%)d _.
i a a a is a i: i:

9
§

Percent of Weight Absoer (%)

 

    

MW- GremI-WLPETSIB E-GWLPE-m‘le Grele’ETGB
25113.96 Jae-THINK: Ghee-31M.%

Figure 2.37. Moisture absorption test results for hybrid material system

 

§~ +P-11-02:TopandBottomGeee
3_ TeetCondltiona:

§ 3-5

 

 

+ Pa31-02: Topand Bottchute

 

WE with
T&B Jute Layers

 

 

 

I I I I J J
11111 r v I v I r v v v v v r 1 1 v I r v v r v r

16 21 26
Square Root ofTIne (la-1’")

Figure 2.38. Percent weight absorbed versus square root of time for hybrid material system

59

 

CHAPTER 3 CELLULAR BEAMS AND PLATES

3.1 Overview

A structural experimental study was conducted with the aim of developing and
evaluating hierarchical and hybrid sandwich structures that would allow biocomposites to
be used for load-bearing structural components. The studies included development,
manufacturing, and testing of cellular beams and plates with varying cellular topologies
and hybrid material systems. In the following sections, the structural forms investigated
in the experimental studies are described, followed by the manufacturing and testing of
the test units. The results of the flexural tests are provided with emphasis on comparing
the performance of the various structural forms investigated in this study. The chapter is
concluded with an experimental comparison of wood materials with the biocomposite

cellular beams and plates tested in this study.

3.2 Structural Form

Three types of high-performance structural forms were investigated for the
structural experiments: (i) cellular sandwich structures, (ii) hierarchical cellular
structures, and (iii) hybrid cellular structures. A brief description of each structural form

is provided in the following sections.

3.2.1 Cellular Sandwich Structures

A sandwich structure is typically a low-density core material sandwiched between
two high modulus face sheets to produce a lightweight panel with excellent stiffness

(Allen 1969). The face sheets act like the flanges of an I-beam carrying the tensile and

60

compressive loads due to bending, while the core acts like the web of an I-beam carrying
the shear loads. The core material provides the panel thickness, and is typically made of
a cellular structure to minimize the weight of the panel. The thicker the core of the panel,
the higher the stiffness and strength for minimum weight gain. These conceps of cellular
and sandwich structures are thus used to provide an efficient material arrangement for
biocomposite structures.

The most important characteristic of a cellular structure affecting its mechanical
properties is its relative density, p*/p_., ie. the density of the cellular material, pi, divided
by the density of the solid from which the cells are made, p... This concept will be
discussed further in the analytical studies chapter (Chapter 4). However, the relative
density is used in this chapter to distinguish between the different cellular structures
studied in the experimental program.

In the structural experimental program, sandwich beams and plates with cellular
cores were investigated. A cellular core comprised of circular periodic cells in the
longitudinal direction of the structure (Figure 3.1) was selected for the cellular beams and

plates based on a similar study by Huang and Gibson (1995).

 

Cellular Structure /

 

 

 

.siiOOI
LOO 13ml; 000000000000
|<——>i Le 102mm >!

25 mm

 

 

 

 

 

 

 

 

Figure 3.1. Periodic cellular sandwich bema and plate

61

3.2.2 Hierarchical Cellular Structures

The concept of hierarchical cellular structures, which is commonly found in
nature, can be used to further increase the stiffness and efficiency of a cellular sandwich
structure (Gunderson and Thorp 1993). Hierarchical cellular structures are complex
rnicrostructures that are designed to place material where most needed, i.e. in areas of
high stress, and which feature multiple levels of embedded complexity in their geometric
or material morphology. For structures primarily subjected to bending loads, the material
in the section is most densely arranged near the top and bottom of the section where the
stresses are high, and less densely arranged near the center of the section where the
stresses are low.

The hierarchical cellular structures developed for the structural experimental
program consisted circular cells, in the longitudinal direction of the structure, arranged in

varying sizes through the depth of the structure (Figure 3.2).

Hierarchical Cellular Structure ——7

 

 

 

000000
0000
DO oeooooooooooooooooooooooooooooooooooeoo
000 06°COZ°°G°O°G°G°°O
0000 OooooQoOooooooooooO
°°°°°° ooooooooooooooooooooooooooooooooooeooeo

 

 

 

 

\— Hierarchical Longitudinal Void Arrangement ]

Figure 3.2. Hierarchical cellular sandwich beam and plate

62

3.2.3 11be Cellular Structures

The hybrid cellular structures investigated in the structural experimental program
were either cellular sandwich structures, or hierarchical cellular structures with integral
bottom, or top and bottom, face sheets (Figure 3.3). The face sheets used consisted of
natural and synthetic fiber mats, as described in the material experimental program of
Chapter 2. The face sheets were integrally cured with the cellular structure to avoid

delamination problems.

A Integral Face Sheet

 

Cellular Structure /

 

 

 

0:
Q.

 

 

000000000000

Longitudinal Void Arrangement —/
Figure 3.3. Hybrid cellular beam and plate

 

 

 

 

P8

3.3 Experimental Studies

A structural experimental study was conducted with the aim of developing and
evaluating hierarchical and hybrid sandwich structures that would allow biocomposites to
be used for load-bearing structural components. The study included manufacturing and
flexural testing of laboratory-scale cellular beams and plates with varying cellular
topologies and hybrid material systems. The dimensions of the test units (Figure 3.1) and

size of the cells were selected based on a similar study by Huang and Gibson (1995),

63

where all-resin cellular beams and plates were manufactured and tested to show the
improvement in performance gained by using cellular structures.

Several manufacturing issues arose during the development of the structural
forms, which limited the level of detail that could be achieved in the cellular structures
and the fiber volume content of the biocomposite material system. As mentioned above
the size of the beams and plates and the cell size were selected base on a study by Huang
and Gibson. In addition, the size of the beams and plates were limited to due cost
concerns for this experimental study. The cells were created using rubber tubes, which
are limited to standard sizes and typically run in 3.18 m (1/8”) intervals. Thus, the
level of cellular and hierarchical structure that could be achieved in the test units was
constrained, since the cell sizes were based on a limited set of dimensions. Coupled with
the size of the cellular beam and plate molds, the dimensions of the cell sizes considered
for the cellular beams and plates were limited to diameters of: 9.53 m (3/8”), 6.35 m
(1/4”), 4.76 mm (3/ 16”), and 3.18 m (1/8”). In addition, the relatively small size of the
mold compared to the size of the cells left very small spaces between the cells. These
small gaps were very difficult to uniformly fill with material during manufacturing. By
trial and error, a fiber weight fraction of 10% (which corresponds to a fiber volume
fraction of approximately 8%) was found to be adequate for manufacturing. Thus, the
cellular beams and plates investigated in the experimental studies were developed based
on the above mentioned manufacturing constraints. These manufacturing constraints
may not arise for larger scale cellular beams and plates.

In the following section the dimensions and details of the developed cellular

layouts are presented in a test matrix, which is followed by the materials and constituents

64

used for the experimental studies. The process used for manufacturing the cellular beams
and plates is then presented, followed by the details of the setup used for the flexural

testing the cellular beams and plates.

3.3.1 Test Matrix

The use of hierarchical cellular and sandwich structures to improve the
performance of biocomposite materials was investigated experimentally through flexural
tests on laboratory-scale beams and plates. The objective of the study was to evaluate the
flexural performance of the biocomposite cellular concepts, and to assess the effects of
hierarchical cellular architectures and hybrid material designs. Three cellular beams with
varying cellular structures, two cellular plates with varying cellular structures, and four
hybrid cellular plates systems were considered for the structural experimental program.
The test matrix with dimensions, cellular structures, and relative densities of all the

samples is schematically depicted in Figure 3.4.

65

 

Cellular Structures

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cellular Beams Cellular Plates
25mm x 25mm x 508mm 13mm x 102mm x 305mm
E g 933‘ D=9.5mm, Typical \ 9E
D=6.4mm -
D=9.5mm, . '—\ D=9.5mm, o o
1...... —\00 ,pr “008 W, ‘000 .7... I00‘000000]
b O /_ Typical ,, / = o 56
D=4.7mrn, J19 O O D=6.4mm. ' 0’ p p5 '
Typical Typical J
P*/Ps = 0.56 P*/Ps = 0-69 p"/ps = 0.57 D=6.4mm, C_:_P2
Typical A O O O O O 0
0.3.2.“... L0§00000o000000fl
Typical p‘/ps = 0.64
Hybrid
Cellular Plates
CP1-G CP1-J CP1-C CP2-C
TOP 8| Bottom Top & Bottom Bottom Top & Bottom
Chopped Glass Strand Woven Jute Unidirectional Carbon Unidirectional Carbon

Figure 3.4. Schematic of the sample matrix for the experimental studies

Each cellular beam and plate type was given a sample identification label (11))
depending on the type of structure (cellular beam or plate), the cellular architecture used,
and the hybrid material system (if used). The first letters of the ID stand for either
cellular beam, CB, or cellular plate, CP. The following digit refers to the layout of the
cellular architecture as shown in Figure 3.4. A final letter is used for the hybrid cellular
structures, which stands for the type of fiber face sheets used in the hybrid structure. The
hybrid material systems and letter IDs are shown in Figure 3.4.

The material design for all the test units is summarized in Table 3.1, noting that
unsaturated polyester resin was used for all materials systems. The influence of the
hierarchical arrangement of the cellular structure was evaluated through different cellular

design as shown in Figure 3.4. Improvement on structural performance through hybrid

66

cellular sandwich plates was investigated by the use of synthetic and natural fibers cured
integrally with the cellular core. The hybrid material systems used for the experimental
study are shown in Figure 3.4. It should be noted that the research program was aimed
towards concept development and feasibility evaluation. Thus, multiple tests of equal
samples were not considered. A pictorial summary of the built test units is given in

Figure 3.5 through Figure 3.8.

Table 3.1. Test matrix of cellular beams and plates with unsaturated polyester resin

 

Material Fiber Volume

 

Fiber ID (%) Test Unit ID

Cellular Beams

Green Hemp 01 8 CB1, CBZ. CBS

Cellular Plates

Green Hemp 01 8 CP1, CP1-G, CP1-J, CP1-C
Flaw Hemp 02 8 CP1, CP2, CP2-C

Ex 03 10 CP1, CP1-J

 

   

(a) Cellular beam 1 (CB 1) (b) Cellular beam 2 (CB2) Cellular beam 3 (033)

Figure 3.5. Pictorial summary of the cellular beam test units

67

 

(c) Cellular plate 1 with top and bottom glass strand (d) Cellular plate 1 with top and bottom woven jute
mat layers (CPI-G) fabric layers (CP1-l)

Figure 3.6. Pictorial summary of cellular plates with green hemp/UPE

 

(b) Cellular plate 2 with top and bottom carbon
layers (CP2-C)

(a) Cellular plate 2 (CP2)

Figure 3.7. Pictorial summary of the cellular plates with raw hemp/UPE

68

 

(b) Cellular plate 1 with top and woven jute fabric
layers (CP1-J)

(a) Cellular plate 1 (CPI)

Figure 3.8. Pictorial summary of the cellular plates with chopped flax/UPE

3.3.2 Materials and Constituents

The cellular beams and plates were manufactured using biocomposite material
systems with either green hemp, raw hemp, or chopped flax fibers, and unsaturated
polyester resin (Table 3.1). The biocomposite material systems are described and the
performed characterization studies are presented in Section 2.3. The resin system used
was an ortho unsaturated polyester (UPE) with methyl ethyl ketone peroxide (MEKP)
catalyst (1% by weight of resin), and cobalt naphthenate (CN) promoter (0.03% by
weight of resin). The resin system used for the cellular beams and plates was the same
system used for the material experimental program, and is shown in tabular form in Table
2.1.

The hybrid cellular plates were manufactured with integral face sheets of either
synthetic or natural fabrics. Three types of fiber fabrics were used for the face sheet
layers: (a) woven jute, (b) randomly orientated glass, and (c) unidirectional carbon
(Figure 2.23). The face sheet material systems and their characterization is presented in
Section 2.4. A summary of the average mechanical properties of the face sheet material

systems used in the hybrid cellular plates is given in Table 2.17.

69

3.3.3 Manufacturing

The cellular beams and plates were manufactured using specially designed molds
with removable faceplates (Figure 3.9). The cellular beam mold was 25.4 mm by 25.4
mm and 508 mm in length, while the cellular plate mold was 304.8 mm by 304.8 mm
with a thickness of 12.7 mm. The removable faceplates allowed different cellular
configurations to be implemented with the same mold.

In order to release the samples from the mold after curing, the molds were lined
with Teflon paper (Figure 3.10a). The cells in the cellular beams and plates were created
by placing rubber tubing through the faceplate (Figure 3.9). Steel rods were placed inside
the rubber tubes to stiffen the tubes, and ensure that the tubes remained straight during
manufacturing (Figure 31%). Tubes smaller than 6.4 m (1/4”) in diameter were held
straight by tensioning the tubing and anchoring it to the mold using small nails. A wood
block was placed inside the plate mold to allow manufacturing of plates with a reduced

width of 101.6 mm (Figure 3%).

   

(a) Cellular beam mold, and rods (b) Cellular plate mold, rods, and Teflon paper

Figure 3.9. Specially designed molds used for manufacturing the cellular beams and plates

70

The biocomposite material system was prepared for manufacturing by pouring the
resin over the fibers (Figure 3.10e) and mixing the two materials together by hand until
the composite material was consistent (Figure 3.101). For the hybrid cellular plates, the
fiber fabric layers were wet out by pouring and spreading the resin over the fabric and
were then placed in the mold. The typical mold and material preparation process for
manufacturing of the cellular beams and plates is summarized in the photograph sequence

shown in Figure 3.10.

71

'e‘liliifll

“If“ ‘

(b) Rubber rods

7 I}:
\ ~. -,

 

(e) Wetting of fibers with resin

Figure 3.10. Preparation of the mold and material system

With the preparation of the mold and material system completed, a layer of
material was placed on the bottom of the mold, followed by a layer of rods (Figure 3.11g-
i). Another layer of material was then placed on top of the rods and compacted in the
voids between the rods (Figure 3.1 l j). This procedure was repeated until all the layers of

rods had been placed in the mold. Because of the small space left between the rods, a

72

low fiber weight fraction was required for the biocomposite material system so that a
properly wet-out material could be evenly distributed throughout the sample. Once all
the material and rods were positioned in the mold, a top layer of Teflon paper was placed
on top of the sample followed by a leveling wood block (Figure 3.11k). Steel plates and
weights (14 kPa of pressure) were then placed on top of the leveling wood block to
uniformly compact the top layer of the sample and obtain a flat surface (Figure 3.111).
All cellular beams and plates were even cured using a cycle of 100°C for 2 hours,
followed by 150°C for 2 hours. The typical manufacturing process of the cellular beams

and plates is summarized in the photograph sequence shown in Figure 3.11.

73

 

(k) Covering and leveling (l) Oven curing with steel weights

Figure 3.1]. Manufacturing process of biocomposite cellular beam and plates

3.3.4 Test Setup

The flexural characterization of the cellular beams and plates was done through
flexural test setups mounted on an MTS loading frame. The cellular beams were tested in

four-point bending with a total span on 457 mm. Load was applied to the beams using

74

the test frame loading ram and a loading fixture that applied two point loads at 102 mm
apart with 178 mm shear spans (Figure 3.12a). The cellular plates were tested in three-
point bending with a shear span of 267 mm (Figure 3.12b). The samples were supported
using a fixture from a steel I-beam and adjustable roller supports. To avoid local
deformations (crushing), aluminum strips were placed underneath each sample at the
supports. The testing setups used for both the cellular beams and plates are shown in
Figure 3.13. All test units were loaded monotonically up to failure in displacement
control at a deformation rate of 0.01 mm/sec. The deflection of the beams and plates was
measured at mid-span using an externally mounted extensometer and the internal LVDT
of the loading frame. The applied load was measured using the internal 12 kN load cell
of the testing frame. The stiffness (force per unit displacement) was measured by taking

the slope of the initial linear portion of the measured load displacement curve.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Four-Porn Loading Test Data: Three-Point Loadhg Tea Data:
Cellular Beams Load Rae=o.o1mrn/s Cellular Plates Load mw.o1m:
m m Load Cell=12KN Load CelbiZKN
p
F J
A, /l A
4 Exteneometer ’ —— /
~—178mm———-1 102mm b—178mmJ[133mm*il:-/133mmé:
457 mm 267 mm
(a) Cellular beam four-point bending test (b) Cellular plate three-point bending test

Figure 3.12. Schematic of flexural test setups

75

   

(a) Cellular beam four-point bending test (b) Cellular plate three-point bending test

Figure 3.13. Pictures of flexural tests on cellular beams and plates

3.4 Cellular Beams

Biocomposite cellular beams with varying cellular and hierarchical topologies
were tested to investigate the effect of hierarchical cellular architectures on the flexural
performance. The observed test behavior and results are presented in the following

section.

3.4.1 Observed Behavior and Test Results: Cellular Beams

The results of the cellular beams with green hemp/unsaturated polyester are
shown in Figure 3.14 and given in Table 3.2. All cellular beams behaved linear-
elastically up to failure exhibiting only small deformations. The failure mode of the
biocomposite cellular beams was a sudden tensile rupture of the bottom surface of the
beam.

The effect of cellular and hierarchical topologies in the performance of the
cellular beams is not evident in the force-displacement response (Figure 3.14). Clearly
from the measured results, the beams with higher relative density, i.e. more solid

material, showed a higher flexural stiffness (Figure 3.14). The effect of hierarchical

76

structure on the performance of the beams must be evaluated by comparing the specific
stiffness of the beams, as shown in Figure 3.15. The specific stiffness of the two
hierarchical cellular beams (CB2 and CB3) is approximately 12% higher than the beam
with periodic cells (CB1). Notice that CB1 and CB3 have the same relative density (use
the same amount of material), but by arranging the material of CB3 in a hierarchical
architecture the stiffness (Table 3.2) and relative stiffness (Figure 3.15) are improved.
Due to manufacturing limitations and constraints on the available cell sizes the
level of hierarchical architecture that could be achieved was limited. Thus, the
improvement in performance shown in the experimental results is not large. Nonetheless,
the experimental results show the improvement in structural performance that can be
achieved by using an efficient hierarchical arrangement of material. More detailed and
efficient hierarchical arrangements should be possible for larger structures with more

available options in cell sizes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

900,
3 —CB1:Cellulu Beam 1 082
80° 1* — 032: Cellular Beam 2
7003 —CB3: Cellular Beam3
3 m: 033 081
600 :. b=25mm, h=25mm, L=457mm
: last:
2 500 §_ 4-Pt Bending
'o : m
g ‘00:? Mid-Span 991 g 93
_| I
. CO 000 000
. 000
30° 1” CO 000 009)
m p'/p‘=O.56 pu/ps-oea pt/ps-057
l Marmalade:
100 3 Green Hemp -1o wt.%
3 Unsaturated Polyester
0‘..s..L....-...4+r-.......4.rs-:.s.
o 1 2 3 4 5 6 7

Displacement (mm)

Figure 3.14. Force versus displacement response at mid-span of cellular biocomposite beams

77

Table 3.2. Flexural test results summary of cellular biocomposite beams

 

Fibers
Beam by Wt. Strength Max Displ. Stiffness
Material Type (%) ID (N) (mm) (N/mflL

 

G.Hemp/UPE CBS1 10 0381-01-03 669.7 6.20 101.9
G.Hemp/UPE 0882 10 0332-01-01 813.2 5.87 135.1
G.HemB/UPE C383 10 CBS3-01-0_2 677.5 55$ 117.0

 

160
Material Data:
Green Hemp - 10 wt.%
155 '_ Unsaturated Polyester

    

Specific Stiffness [(Nlmm)l(glcm’)]

 

 

Cellular Beam 1 Cellular Beam 2 Cellular Beam 3

Figure 3.15. Specific stiffness of cellular biocomposite beam

3.5 Cellular Plates

Results form the flexural tests on cellular plates are divided into the following
groups: (i) cellular and hybrid plates with green hemp/unsaturated polyester, (ii) hybrid
cellular plates with carbon face sheets and the corresponding non—hybrid cellular plates,
and (iii) cellular and hybrid plates with flax/unsaturated polyester. Data from the flax

cellular plates is presented together with results from the green hemp and an all—

78

unsaturated polyester cellular plates for comparison. The observed test behavior and

results are presented in the following section.

3.5.1 Observed Behavior and Test Results: Cellular Beams

The results of the cellular and hybrid plates with green hemp/unsaturated
polyester are shown in Figure 3.16 and provided in Table 3.3. All plates behaved linear-
elastically up to failure (Figure 3.16), failing in a sudden brittle rupture at the bottom
fiber of the plate. The plates exhibited small deformations up to failure, with exception
of the hybrid cellular plate with a bottom layer of unidirectional carbon (Figure 3.16).
This hybrid cellular plate showed large defections near failure due to the large strength
capacity provided by the bottom layer of carbon fibers. The increase in strength is a
result of the high properties of the carbon fibers, efficient placement of the material, and
curing of the carbon fiber layer of fibers integrally with the cellular core of the plate.

The use of hybrid cellular plate systems improved both the stiffness and strength
of the plates as shown in Figure 3.17. The hybrid plates with top and bottom layers of
glass strand mat and jute fabric increased the strength of the cellular plate by 50%, while
their stiffness remained constant (Table 3.3). The increase in strength due to the addition
of top and bottom jute fabric layers is nearly equal to the increase due to the use of glass
strand mat layers (Table 3.3). This suggests the strength of a biocomposite cellular plate
can be increased by adding woven natural fiber layers, which maintains the “green”
appeal of the biocomposite material system. The use of unidirectional glass mat or a
denser weave of natural fiber fabric may result in a stiffer fiber mat, thus increasing the

stiffness of the plate as well as the strength.

79

The results in Figure 3.16 show that both the strength and stiffness are
significantly improved by adding only a bottom layer of carbon fiber to the biocomposite
cellular plate. The carbon fiber layer is most efficient on the bottom of the plate because
it prevents the progressive tensile rupture of the biocomposite material in this region, and
improves the structural stiffness of the plate. Therefore, the hybrid hemp/carbon plate
increased the strength of the base-line cellular plate by a factor of seven, and almost
doubled the stiffness (Table 3.3). The hybrid hemp/carbon cellular plate also improved
the strength of the plate four and a half times over the hybrid hemp/glass and hemp/jute
plates (Table 3.3). The results show that by placing only a small amount of high
performance material in the most efficient location the properties of the plate can be
dramatically increased.

In addition to the improvement in strength and stiffness due to the use of hybrid
systems, the integral casting of the fiber layers with the cellular plate eliminates failure
due to delarrrination. Delamination failure typically occurs in sandwich panels where the
top and bottom face sheets are simply bonded to the core material. The bond between the
face sheet and the core creates a weak plane that can lead to premature failure of the plate
(Jones 1999). None of the samples tested failed by delamination, thus showing the

importance of casting the top and bottom face sheets integrally with the cellular core.

80

 

 

 

 

 

  
   
   

 

 

 

 

 

 

 

12000
‘ —ce1: Cellular Plate 1
‘ —CP1-G: CP1 wrrae Giaea Strand Mat Layers CP1-C
10000 A ——CP1-J: CP1 w/Tae Jute Fabric Layers
i —cp1-c: cm ill/Bottom Carbon Layer /
1 P135 Data:
8000 f b=102mln, h=13mm, L=267lnm
A ~ 19.83;
E j 3-Pt Bending
13 6000 m
3 , Mid-Span
..l .
2000 - CP1-G M teri i .
1 Green Hemp - 10 wt.%
~ Unsaturated Polyester
o:fi.-fi.-..filf....+....:...-
0 5 10 15 20 25

Displacement (mm)

Figure 3.16. Force versus displacement response at mid-span of cellular and hybrid plates with green
hemp fibers, and unsaturated polyester

 

 

. —CP1: Cellular Plate 1

350° ;_ —-—CP1-G: CP1 ill/T&B Glace Strand Mat Layers
* — CP1al: CP1 W/T&B Jute Fabric Layers

—CP1-C: CP1 ill/Bottom Carbon Layer

 

 

 

q
1
3000*

 

 

 

 

 

 

 

: Pia ta:
3 b=102mm,h=13mm, L=267mm
A 2500 M;
E 1 3-Pt Bending CP1“ cp1.G
g 2000 EM".
3 j Mid-Span
150° :” CP1-C CP1
‘ (truncated ,
1000 f P..=11010N)
I MaterialDag:
50° " Green Hemp-10 wt.%
Unsaturated Polyester
O‘r—efi.:-...l....l....:fi..¢-..-l....4...e:....i....
0 1 2 3 4 5 6 7 8 9 10
Displacement (mm)

Figure 3.17. Force versus displacement response at mid-span of cellular and hybrid plates with green
hemp fibers, and unsaturated polyester, zoomed in view

81

Table 3.3. Flexural test results summary of cellular, and hybrid plates with green hemp fibers

 

 

Fibers
Beam by Wt. Strength Max Displ. Stiffness
Material Tme (%) ID (N) (mm) (N/mm)
G.Hemp/UPE CP1 10 CP1-01-O1 1483.3 3.65 406.0
G.Hemp/UPE - T&B Glass CP1 10 CP1-11-01 2236.904 8.42 319.7
G.Hemp/UPE - Bottom Carbon CP1 10 CP1-21-01 10396.8 19.70 591.9
G.Hemp/UPE - T&B Jute CP1 10 CP1-31-01 2196.8 6.95 308.7

 

Results for the hybrid cellular plates with carbon face sheets, and the
corresponding non-hybrid cellular plates are shown in Figure 3.18 and provided in Table
3.4. The response of all the cellular plates were similar to the results previously
mentioned, except for the hybrid cellular plate with top and bottom carbon face sheets.

The results of the plate with top and bottom carbon layers (Figure 3.18) showed a
bi-linear response due to buckling of the top carbon layer. Before buckling of the t0p
layer both carbon layers contribute to the stiffness of the plate, thus the large slope of the
first portion of the plot. After buckling of the top layer of carbon this layer no longer
contributes to the stiffness of the plate, and thus the slope of the load-displacement plot
decreases at this point. Essentially, after buckling of the top layer of carbon the plate is
equivalent to the plate with only a bottom layer of carbon. This is shown in Figure 3.18
by comparing the slope of the plate with top and bottom carbon layers after buckling with
the slope of the plate with only a bottom layer of carbon. The two slopes are almost the
same after buckling of the top carbon layer.

Although the stiffness of the plate with top and bottom carbon layers was
increased by 150% over the plate with only a single bottom carbon layer, the strength of
the plate was not improved. This is due to the buckling of the top layer of carbon fabric,

which occurred at only 50% of the peak load. In structural applications, the large loss of

82

stiffness after buckling could make the structure unusable. Thus, when using hybrid
material systems to improve the performance of biocomposite cellular structures, the

placement of the material should be designed by considering the effect on the complete

response of the structural component.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12000
—CP1: Cellular Plate 1 CP2-C
, —CP2: Cellular Plate 2
10000 .. —CP1-C: CP1 wIBottom Carbon Layer
‘ —CP2-C: CP1 wrrae Carbon Layer:
3 CP1-C
. Plate Data:
8000 j" b=102rnm,h=13mm, L2267mm 931
A . Test:
5 1 Fr Bending loco/o 002001
'0 <>- . t '= ,
a 5000 . mm p p
3 - MIG-Span rip;
m :. @08030803030301
1 pr/p,=0.64
: cpg Material Data:
2000 .t CP1: Green Hemp-10wt.%
j CP2: Flaw Hemp - 10 wt.%
: CP1 Unsaturated Polyester
0‘ -fi.:-...:.......--asfi..... :...¢-...:.e.-:.e--
0 2 4 6 8 10 12 14 16 18 20
Displacement (mm)

Figure 3.18. Force versus displacement response at mid-span of cellular, and hybrid plates with
green hemp, and raw hemp fibers, and unsaturated polyester

Table 3.4. Flexural test results summary of cellular, and hybrid plates with green hemp fibers

 

 

Fibers
Beam by Wt. Strength Max Displ. Stiffness
Material Type (%) ID (N) (mm) (N/mm)
G.Hemp/UPE CP1 10 CP1-01-01 1483.3 3.65 406.0
Raw Hemp/UPE ' CP2 10 0882-01-01 2556.6 5.39 501.5

G.Hemp/UPE - Bottom Carbon CP1 10 CP1-21-01 10396.8 19.70 591.9
Raw Hemp/UPE - T&B Carbon CP2 10 0883-01-02 11010.4 17.56 1480.0

The effect of hierarchical topology on the performance of the cellular plates was
investigated in a plate with periodic cells (CPI) and a plate with a hierarchical

arrangement (CP2) as shown in Figure 3.18. The hierarchical cellular plate showed an

83

increase in specific stiffness of 13% over the periodic cellular plate. As with the cellular
beams, manufacturing limitations constrained the level of topological hierarchy that
could be achieved in the cellular plates. Nevertheless, the experimental results show the
improvement in structural performance that can be achieved by using an efficient

hierarchical arrangement of material.

 

 

. Material Data:

1000 3— CP1:Green Hemp-10M.%
f CP2: Flaw Hemp - 10 wt.%
1300 '- Unsaturated Polyester

 

 

 

Specific Stiffness [(N/mm)l(g/cm3)]

 

 

 

Cellular Plate 1 Cellular Plate 2 CP1 wlBottom CP2 ill/T&B
Carbon Layer Carbon Layers

Figure 3.19. Specific stiffness of cellular and hybrid plates with green hemp fibers, and unsaturated
polyester
The results of the cellular and hybrid plates with flax/unsaturated polyester are
shown in Figure 3.20 and provided in Table 3.5. The flax biocomposite plates are
compared with the green hemp/unsaturated polyester cellular plate and an all-unsaturated
polyester plate. Both flax plates behaved linear-elastically up to failure exhibiting only
small deformations. The failure mode for both cellular plates was a sudden tensile

rupture at the bottom fiber of the plate.

84

Test results showed that the hybrid cellular plate system with top and bottom
woven jute fabric layers improved both the stiffness and strength of the flax plate (Figure
3.20), increasing the stiffness by 32% and the strength by 70% (Table 3.5). These test
results are similar to the test results of the green hemp hybrid plates in this section.

The test results also show that the flax cellular plate performs relatively poorly
compared to the green hemp cellular plate (Figure 3.20). In fact, the stiffness of the flax
cellular plate was equal to the all-unsaturated polyester cellular plate, while the strength
was lower. Thus, the flax fibers are not reinforcing the unsaturated polyester matrix, but
they are rather acting as a pollutant. This behavior is most likely due to the short length
of the flax fibers. Based on these results, the chopped flax fibers are not recommended

for structural applications.

 

 

 

 

 

 

 

 

 

 

 

 

2500
' —CP1: Cellular Plate1 (Green Hemp)
-——-CP1: Cellular Plate 1
. —CP1-J: CP1 WITer Jute Fabric Layers
20m .. ——CP1: Cellular Piate1 (Polyester) CP1-J
: lPlagg Dag:
‘ b2102mm, h=13mm, L=267mm
a1500 L- EL“
a 1 3-Pt Bending cl":
'0 1 we; ( M)
g , Mid-Span
_| 1000 .. CP1 __________._) CP1
: (Green Hemp)
1
500 1* Material Data:
Chopped Flax - 20 wt.%
Unsaturated Polyester
0 ........e...............e.+.e.s.sfis-.fifiier...¥. ..
0 1 2 3 4 5 6 7 8 9 10

Displacement (mm)

Figure 3.20. Force versus displacement response at mid-span of cellular and hybrid plates with flax
fibers, and unsaturated polyester

85

 

Table 3.5. Flexural test results summary of cellular, and hybrid plates with flax fibers

 

 

Fibers
Beam by Wt. Strength Max Displ. Stiffness
Material Type (%) ID (N) (mm) (N/mm)
G.Hemp/UPE CP1 10 CP1-01-01 1483.3 3.65 406.0
FIax/UPE CP1 20 CP1-04-01 1171.7 6.16 196.5
Fiax/UPE - T&B Jute CP1 20 CP1-34-01 1979.7 7.77 259.4
Polyester CP1 -- CP1-03-04 1764.2 9.45 185.8

 

3.6 Comparison with Wood Beams and Plates

One wood beam and three wood plates were tested in flexure and used for

comparison with the cellular beam and plate results from the structural experimental

program. One type of wood, douglas fur, was used for the beam, and three types of

wood, douglas fur (DF), oriented strand board (OSB), and particle board (PB), were used

for the plates. The same dimensions of cellular beams and plates used in the

experimental studies were used for the wood beam and plates. The DF wood beam is

shown in Figure 3.21a, while the DF plate is shown in Figure 3.21b, the OSB plate is

shown in Figure 3.22a, and the PB plate is shown in Figure 3.22b.

if]

(a) Beam: 25.4 mm x 25.4mm x 508 mm

m‘m.

 

(b) Plate: 1

Figure 3.21. Douglas Fur beam and plate test units

86

2.7 mm x 106.4 mm x 304.8 mm

 

(a) Orientated strand board plate: 0.5”x 4"x l2” (b) Particle board plate: 0.5”x 4”x 12"

Figure 3.22. Wood plate test units

The douglas fur wood material is standard stock lumber (solid piece of wood).
The oriented strand board is made by pressing layers of wood chip lamina together and
bonding them with a small amount of resin. Phenolic and diphenyl-methane-di-
isocyanate (MDI) resins are typically used for binding OSB panels (APA Engineered
Wood Association 2003). The particle board is made in a similar manner, by pressing
small wood fibers (saw dust) and bonding them together with a small amount of resin.
Urea-Formaldehyde resins are typically used for binding particle boards (APA
Engineered Wood Association 2003). These types of wood products are commonly used
in the construction industry. It should be noted that while the wood samples are typical
materials used in the construction industry, the biocompostie materials used in the
cellular beams and plates were limited in fiber volume fraction (V, = 8%) due to
limitations in manufacturing. For full-scale biocomposite beams and plates it would be
possible to achieve much higher volume fractions, thus improving the properties of the
material. The comparison of the wood and biocomposite cellular structures must be

viewed in this context.

87

3. 6.1 Observed Behavior and Test Results: Wood Beams

The results of the douglas fur (DF) beam is compared with the biocompoiste
cellular beams in Figure 3.23 and Table 3.6. The DF beam behaved nonlinearly due to
crushing of the material in the top layer of the beam near failure exhibiting large
deformations. The failure mode of the DF beam was a sudden tensile rupture of the
bottom fiber of the beam.

The DF beam out-performed the biocompsotie beams in both stiffness and
strength (Figure 3.23). This is expected, as the wood material is structural grade, while
the manufacturing constrains of the biocomposites material limited the fiber volume
fraction that could be achieved.

Despite the use of hierarchical cellular structures in the biocomposites beams, the
DF beam still out-performed the biocomposite cellular beams (Figure 3.24). The large
specific stiffness of the DF beam is due to the high stiffness of the wood beam to the

biocomposite beams, and the very low density of wood.

88

 

 

 

 

 

 

 

 

 

 

 

 

 

3000
1 —CB1:Celluiar Beam1
j —C32: Cellular Beam 2
2500 .. ——083:CeliularBeam3
. -—DF-B: Douglas Fur Beam DF'3
1 m:
2000 i. b=25mm, h=25mm, L=457mm
A 3 1%:
E . 4-Pt Bending
.0 1500 j_ Lmtlon:
8 ~ MideSpan
.l I
10001r
1 CB2
‘ CB1
500 ._ Mgterlgl bag:
I Green Hemp-10 wt.%
CB3 Unsaturated Polyester
as?
0 5 10 15 20 25
Displacement (mm)

Figure 3.23. Force versus displacement response comparison of cellular, and wood beams

Table 3.6. Flexural test results summary of cellular, and wood beam

 

 

Fibers
Beam by Wt. Strength Max Displ. Stiffness
Material Type (%) IO (N) (mm) (N/mm)
G.Hemp/UPE CBS1 10 0881-01-03 669.7 6.20 101.9
G.Hemp/UPE CBS2 10 0382-01-01 813.2 5.87 135.1
G.Hemp/UPE 0333 10 0883-01-02 677.5 5.58 1 17.0
Douglas Fur DF-B -- DF-B-01 1379.4 5.73 236.6

 

89

 

 

 

Cellular Beam Material Data:
Green Hemp - 10 wt.%
Unsaturated Polyester

J_ll,

Cellular Beam 1 Cellular Beam 2 Cellular Beam 3 Douglas Fur Beam

r

 

 

 

gages

Specific Stiffness [(Nlmm)l(glcm’)]
'2‘ '8‘ fa" s

3

 

 

 

 

Figure 3.24. Specific stiffness of cellular and wood beams

3.6.2 Observed Behavior and Test Results: Wood Plates

The results of the douglas fur (DF), oriented strand board (OSB), and particle
board (PB) wood plates are compared with a green hemp/unsaturated polyester
biocomposite cellular plate in Figure 3.25 and Table 3.7. The DF plate behaved
nonlinearly due to crushing of the material in the top layer of the beam near failure
exhibiting large deformations. The failure mode of the DF plate was tensile rupture of
the bottom fiber of the beam. The fracture of the bottom fiber of the DF plate was not
sudden, but rather gradual cracking along the bottom of the plate (shown by the spikes in
Figure 3.25). The OSB and PB plates behaved linear-elastically up to failure exhibiting
only small deformations. The failure mode of the OSB plate was fracture of the cross-ply
above the bottom fiber of the plate, while the failure mode of the PB plate was a sudden

tensile rupture of the bottom layer of the beam.

90

The DF plate out-performed the biocomposite cellular plate in stiffness, and
strength (Figure 3.25), and specific stiffness (Figure 3.26). These results are similar to
those of the DF beam. However, the biocomposite cellular plate out-performed the OSB
and the PB plates in strength and stiffness (Figure 3.25), even though the wood plates are
considered structural grade. The stiffness of the biocomposite cellular beam was 107%
higher than the OSB plate and 230% higher than the PB plate, while the strength of the
biocomposite cellular beam was 62% higher than the OSB plate and 114% higher that the
PB plate (Table 3.7). The bio composite cellular plate also out-performs the OSB and PB
plates in specific stiffness, despite the very low density of wood (Figure 3.26).

The results of this comparison show that biocomposite cellular plates can out-
perform engineered wood products. Thus, biocomposites cellular structures have the
potential to be used for applications where these engineered wood products are used. The
results would be more in favor of the biocomposite cellular plates if an improved
manufacturing procedure was used to fabricate plates with higher fiber volume fractions,

or if hybrid biocomposite cellular components were used.

91

 

 

 

—CP1: Cellular Plate 1 (Green Hemp)
—— DF-P: Douglas Fur Plate

— OSB-P: Orlentated Strand Bored Plate
— PB—P: Particle Board Plate

 

 

 

 

 

 

 

 

DF-P
Plate Data:
b=102mm, h=13mm, L=267mm
M
3-Pt Bending
Lmtlon:
Mid-Span
CP1
(Green Hemp)
OSB-P Material Data:
”4’ Green Hemp - 10 wt.%
Unsaturated Polyester
0 2 4 6 8 10
Displacement (mm)

 

12

Figure 3.25. Force versus displacement response comparison of cellular, and wood plates

Table 3.7 . Flexural test results summary of cellular, and wood plates

 

 

Fibers
Plate by Wt. Strength Max Displ. Stiffness
Material Type (%) ID (N) (mm) (N/mm)
G.Hemp/UPE CP1 10 CP1-01-01 1483.3 3.65 406.0
Douglas Fur DF-P -- DF-P-01 3494.7 11.19 628.7
Orientated Strand Board OSB-P -- OSB-P-01 916.6 5.56 195.8
Fiber Board PB-P -- FB-P-01 694.3 6.15 123.0

 

92

Specific Stiffness [(Nlmm)l(glcm’)]

5

'5

d

 

.5. 5.5 5

 

 

 

Green Hemp -10 wt.%

Cellular Plate Material Data:
Unsaturated Polyester

 

 

Cellular Plate 1 Douglas Fur Orientated Strand Particle Board

Figure 3.26. Specific stiffness of cellular and wood plates

93

CHAPTER 4 ANALYTICAL STUDIES

4.1 Overview

The material and structural experimental studies of Chapters 2 and 3 were
investigated analytically with the objectives of: (i) validation of the material experimental
studies from Chapter 2, (ii) validation of the structural experimental studies from Chapter
3, and (iii) feasibility studies. The mechanical properties of the biocompostie material
systems are investigated analytically by conducting a survey of available models for the
properties of randomly oriented short fiber composites, and comparing the models with
experimental results. The biocomposite material models are then used with conventional
structural mechanics to analyze the cellular beams and plates, and compare with the
measured results. Next, the performance of the biocomposite cellular structures in
flexure is assessed using indices and factors, which measure the efficiency of a material
and structure in flexure. A more refined analysis procedure is then proposed, which can
be automated in mathematical software to easily analyze hierarchical cellular structures
and used for parametric and optimization studies. The thermal effects due to curing in
hybrid cellular structures are considered, and the resulting prestressing of the section is
investigated analytically on laboratory and full—scale structural components. The
analytical models are then used to predict the performance of full-scale biocomposite
structural components for civil applications. Application of biocomposite structural
components for a highway bridge deck, and commercial and residential flooring systems
is investigated. The concepts discussed above are presented in detail in the following

sections

94

4.2 Properties of Randomly Oriented Short Fiber Composites

The elastic constants, and strength of biocomposite materials can be predicted
using analytical models for randomly oriented short fiber composites. The analytical
models for randomly oriented short fiber composites presented in this section are based
on using the properties of equivalent unidirectional lamina and averaging the properties
over all fiber directions. Thus, the micromechanical properties of unidirectional
orthotropic lamina are presented along with the properties of randomly oriented fiber
composites.

Randomly oriented short fiber-reinforced composites are not as strong or as stiff
as continuous fiber-reinforced composites, but short fiber composites do have several
attractive characteristics that make them attractive for structural applications (Gibson
1994). For example, short fiber composites can be easily mixed with liquid resin and
used to produce parts having complex geometry. The processing methods for short fiber
composites such as resin transfer molding, injection molding, or compression molding
are fast and inexpensive, which makes them attractive for high volume applications.
Composites having randomly oriented short fiber reinforcement are nearly isotropic,
compared to unidirectional continuous fiber composites that are highly anisotropic. The
advantages of low cost, ease of fabricating complex parts, and isotropic behavior make
short fiber composites attractive for many applications.

In the following sections the micromechanical equations for the elastic moduli,
and strength of continuous fiber composites are presented. The continuous unidirectional
equations are needed to model the face sheet materials, e.g. carbon and jute fiber mats,

and to establish limits for the short fiber models. Next, models for the properties of short

95

fiber composites, including a series of models for the isotropic elastic constants and
strength of randomly oriented short fiber composites are presented. The results from the
analytical models for randomly oriented short fiber composites are then compared with

measured results from the material experimental program of Chapter 2.

4.2.1 Continuous Fiber Composites

A mechanics of materials approach can be used as a basic method to model the
micromechanics of continuous fiber-reinforced composite materials (Jones 1999). The
mechanics of materials approach implies that certain simplifying assumptions must be
made in order to arrive at an effective solution. The most important assumption made is
that the strains in the fiber direction of the composite are the same in the fibers as in the
matrix. In other words, when stressed in the l-axis, or material longitudinal axis, the
plane sections normal to the l-axis will remain plane after stressing. This assumption is
similar to the “plane sections remain plane” assumption made in the mechanics of
materials approach to beam theory. Other important assumptions in the mechanics of
materials approach include: (i) the fibers are all uniform in properties and diameter,
continuous, and parallel through the media, and (ii) there is a perfect bond between the
fibers and matrix so that no slippage can occur.

In the following sections the mechanics of materials approach is used to derive
the orthotropic moduli, E1, E2, V12, and 012, and strength of a continuous unidirectional
fiber-reinforced composite (Figure 4.1). The derived properties are obtained by
analyzing a representative volume element (Figure 4.1). Other approaches to the

mechanical properties, such as the theory of elasticity, are available. However, the

96

results from the mechanics of materials approach provide simple and practical equations

that accurately predict experimental behavior (Jones 1999, Gibson 1994).

 

 

I
g
l.

 

 

 

 

 

 

Figure 4.1. Orthotropic lamina, and representative volume element

4.2.1.1 Properties for Continuous Fiber Composites: Stifiiless

Consider first the modulus of a unidirectional fiber-reinforced composite in the 1-

direction (Figure 4.1). From Figure 4.2, the strain in the l-direction is:
A
6, = —LL-, (4.2.1)

where the strain is the same in both the fiber and matrix. If both the fibers and matrix are

elastic the stresses in the l-direction are:
0', = E ,6, 0' = E 6,. (4.2.2a,b)
In Figure 4.2 the stress or acts over the entire cross-sectional area A of the representative
volume element, while 0‘; acts over the cross-sectional area of the fibers Af, and 0'," acts
over the cross-sectional area of the matrix Am. Accordingly, the resultant force in the
representative volume element is:
P=01A=crfAf +amAm. (4.2.3)
Using Equation (4.2.2) with Equation (4.2.3), and recalling that 0'=E6‘, the elastic

modulus in the l-direction can be written as

97

A A
E1=Ef 7H5", 7'". (4.2.4)

Assuming no voids in the composite, the fiber volume fractions of the fibers and matrix
can be written as

A A
v, = 7’ Vm = 7'". (4.2.5a,b)

Substituting Equation (4.2.5) into (4.2.4) leads to
E1 = E fo + Eme, (4.2.6)
which is known as the rule of mixtures for the apparent Young’s modulus of the

composite material in the 1-direction (Jones 1999).

 

/ matrix '
I /
l
<— / Matrix / ——-
F ----------------- -( /

 

 

 

 

L._.—.i is...

Figure 4.2. Representative volume element for loading in the l-direction

The rule of mixtures predicts a linear variation of the modulus in the l-direction
for a given fiber volume fraction. Despite the simple form of Equation (4.2.6), its
predictions agree well with experimental tensile tests (Jones 1999, Gibson 1994). The
results from the rule of mixtures also agree well with the unidirectional
carbon/unsaturated polyester material tested in the experimental program of this study

(Figure 4.3).

98

 

1'5

  
  
  
   

 

—a— Rule of Mixtures
O Carbon/Polyester (Measured)

ii
‘1

.L

 

 

 

d g

‘

 

Material Dag;
Carbon/Polyester

E, =2306Pl
5... =1.5 Gpa
V, =O.16

Longitudinal Elastic Modulus (GPa)

 

 

 

 

 

 

1 J

‘j‘ "Yj‘fi VrYTTfiY '—T '1' j' f YTY

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.30 0.90 1.00
Fiber Volume Fraction ( —)

Figure 4.3. Predicted versus measured E1

Consider next the apparent Young’s modulus in the 2-direction (Figure 4.1). In
the mechanics of materials approach the transverse stress, 0'2, is assumed to be applied to
both the fiber and matrix (Figure 4.4). However, unlike the modulus in the l-direction, it
cannot be assumed that the strains in the fiber and matrix are the same. Thus, the strains
in the fiber and matrix are found from the stresses, that is:

a, =—g-:- em = g? (4.2.7,b)

It is assumed that that transverse dimension over which, 6} acts is VfW, while 8," acts on
VmW (Figure 4.4). Accordingly, the total transverse deformation is:

AW = 52W = VfW€f + VmWe'm , (4.2.8)

or dividing through by the width W:

e, =Vf8f +Vmsm. (4.2.9)

99

Using Equation (4.2.9), and substituting Equation (4.2.7) for the fiber and matrix strains

the total strain in the 2-direction becomes:

82 = V, % + Vm 3:. (4.2.10)
Using Equation (4.2.10), and noting that
0'2 = E282 , (4.2.11)
the stress in the 2-direction is:
0' 0'
0'2 2 E2[V, 1:72 + V," 5:]. (4.2.12)

By adding the bracketed terms and dividing through by 02, the apparent Young’s
modulus of the composite material in the 2-direction is:
: EfEm
2 .
VmE, +Vf Em
/"’

/ Matrix
W Fiber \ 1
/ Matrix

/
A.

Figure 4.4. Representative volume element for loading in the 2-direction

 

(4.2.13)

 

 

 

 

 

 

Predicted results for both E1 and E2 are plotted with fiber volume fraction in
Figure 4.5. It can be seen from Figure 4.5 that that fiber stiffness has very little effect on
the modulus in the 2-direction. Notice that 55% fiber volume fraction is required to raise
E1/Em to 2, while only 10% is needed to raise E1/Em to 2. Thus, the composite modulus

E2 is a matrix-dominated material property.

100

 

.s
O

 

 

 

 

NUDU‘IONOD

 

 

 

Figure 4.5. Variation of composite moduli with fiber volume fraction

Consider next the major Poisson’s ration v12 (Figure 4.1). The Poisson’s ratio in

the 12-direction is defined as:

v1, = ——, (4.2.14)

for an applied stress 0'1 in the l-direction, with all other stresses equal to zero. The major
Poisson’s ratio of the composite material can be obtained using an approach similar to the
derivation used for E) (Figure 4.6). Again, the strains in the 1-direction are assumed to
be equal in the fiber and matrix. For the representative volume element shown in Figure
4.6, the deformation in the transverse direction, Aw, is

AW 2 —Ws2 = Wvlzsl. (4.2.15)
Also, on the microscopic level, the deformation in the transverse direction is

(4.2.16)

101

It is assumed that that transverse dimension over which, .6} acts is VfW, while 8," acts on
VmW (Figure 4.6), as assumed in the derivation of E2. Thus, the transverse deformations
in the fiber and matrix are:

Amw = Wvame, AM = WV, v, 8,. (4.2.l7a,b)

Using Equations (4.2.15) — (4.2.17), and dividing through by 81W the major Poisson’s
ratio of the composite material is:

v,2 = vme + v,V,. (4.2.18)

The result is a rule of mixture for the major Poisson’s ratio, which gives a linear variation

with fiber volume fraction. Because the Poisson’s ratio of the fiber 1!; and matrix Vm are

not considerably different, the major Poisson’s ratio of the composite material is neither

matrix, or fiber-dominated.

 

 

, Matrix ,
a, x Fiber , j a,
w/ Matrix L W ,/—>
f j —————————— '— ______ -( l/
W i i/
Q L ———————————————— —l

 

 

 

Figure 4.6. Representative volume element loaded in the l-direction

Consider now the in-plane shear modulus, 012, of the composite material. The
shearing stresses on the on the fiber and matrix are the same. Because of the difference
in shear modulus between the fiber and matrix the shear deformations occur as shown in
Figure 4.7. Assuming a linear stress-strain behavior, the shear strains in the matrix and

fiber are:

T ‘l'
: — = —- , 4.2.193,b
yin G Yf Gf ( )

m

102

Following the usual approach, the total shearing deformation is
approximately
Am =Wmem AN =WV,}/,.
Since the total deformation is A = A ,, + A ,, division by W gives:
r=nn+%n.

or using Equation (4.2.19), and noting that

y = — ’
612
Equation (4.2.21) can be written as:
L = V", — + V, -—T—
GIZ m Gf

composed of

(4.2.203,b)

(4.2.21)

(4.2.22)

(4.2.23)

Solving Equation (4.2.23) for 012, the in-plane shear modulus of the composite material

is:

012 = ’
VmG, +V,Gm

 

(4.2.24)

which is similar to the expression found for the Young’s modulus in the 2-direction E2.

Thus, as with E2, the in-plane shear modulus G12 is a matrix-dominated composite

 

 

 

 

 

 

-— Aer/2

 

 

 

 

material property.
/ 2 A.
T , 5,, Matrix

,z’ Matrix , ’ : Fibe

Fiber _’,I 1 r

1’ Matrix 1,” ,H Matrix
.1.
(a) Shear loading in the 12-direction (b) Shear deformati

Figure 4.7. Representative volume element loaded in shear

103

0118

The results from the mechanics of materials approach to the four elastic moduli of
a unidirectional continuous fiber reinforced lamina are shown in Table 4.1. The
equations are simple in form and only depend on the moduli of the fiber and matrix, and
the fiber volume fraction. These equations are based on the assumption that the fibers are
continuous in length. Thus, to model composite materials reinforced with fibers that are
discrete in length other models must be used. However, the continuous unidirectional
equations are needed to model the face sheet materials used in the experimental studies of
Chapters 2 and 3, i.e. the carbon and jute fiber mats, and to establish the upper limit for

the short fiber models.

Table 4.1. Orthotropic moduli of a unidirectional fiber-reinforced composite by mechanics of

 

 

 

 

materials
property Expression
Modulus in the 1-direction E, = E ,V, + Eme
E E, E”,
Modulus in the 2-direction =
2 VmE, +V,E,,
Major Poisson’s ratio (12-direction) v12 = Vme + v, V,
od i G G’ G'"
In- lane shear in u us 12-directino =
p ( l 12 ,1me + ,1me

 

4.2.1.2 Properties for Continuous Fiber Composites: Strength

Consider now the strength of a continuous fiber-reinforced composite material in
the direction of the fiber. In the mechanics of materials approach it is assumed that all
fibers have the same strength and are relatively brittle in comparison to the matrix

material, and both the fibers and matrix behave linear elastically (Jones 1999). If the

104

composite material has more that a certain minimum volume fraction of fibers, the
ultimate strength of the composite material is reached when the fibers are strained to their
maximum stress, i.e.

e . = emx. (4.2.25)

Because the fibers are usually more brittle than the matrix, they cannot elongate as much
as the matrix, and thus become the weak component in the composite material. Thus,
assuming the fiber strain is equal to the matrix strain in the direction of the fibers, the
strength of the composite material is:
0',max = 0',MV, + Em (0', m IE, )Vm, (4.2.26)
where ”firm is the strength of the fiber, Em is the modulus of the matrix, Ef is the
longitudinal modulus of the fiber, and Vf and V”, are the volume fractions of the fibers and
matrix, respectively. Notice that the term amlEfis equal to the failure strain in the fiber,
gum, and that multiplying by Em gives the stress in the matrix at the ultimate fibers strain.
In order for the fibers to have a strengthening effect on the composite material,

the fiber volume fraction must be above the critical volume fraction (Jones 1999):

Ummax —Em(0fmax lEf)
V,m.,,m, = . (4.2.27)
' afmx—Em(afmx/Ef)

 

For fiber volume fractions below the critical value Equation (4.2.26) is no longer valid.
The matrix is then the only contributor to the composite strength, and the fibers act as
holes in the matrix. In this case, the strength of the composite is:

0 ,=0'

(ma

(1 — V, ). (4.2.28)

mmax

105

 

 

4.2.2 Aligned Short Fiber Composites

The elastic moduli of an aligned short fiber composite can be estimated using the
Halpin-Tsai equations (1969) with a modification proposed by Halpin (1969) that

accounts for the fiber aspect ratio (l/d). The stiffness in the fiber direction is given by:

E = ____1+ 577‘]! (4.2.29)
Em l-nV, ’
n E’ ”5’" _l (4230)
W ere =—— . .
77 E,/Em+§
g = 2(l/d). (4.2.31)

The predictions from these equations give good agreement with experimental data
(Halpin 1969). Halpin concluded that the in-plane shear modulus is not significantly

sensitive to the fiber aspect ratio, and may be approximated as:

.912 = ..__._.1+ 5W! (4.2.32)
6,, 1— 77V, ’
G /G -1
where n = —-’—"‘—— (4.2.33)
G, I Gm + 4‘
5 =1.

The same argument is made is made for the transverse modulus E2, and the major

Poisson’s ratio v12. The transverse modulus is then approximated as:

 

55 =——l+f'7vf , (4.2.34)
Em l-nV,
h E’IE'" _1 (4235)
W ere =——-—-— . .
’7 E,/Em+lf
(i=2

106

The major Poisson’s ratio is close to the result given by the rule of mixtures, and can be
approximated using Equation (4.2.18).

The variation of the longitudinal modulus with fiber aspect ratio for a green
hemp-polyester composite using the Halpin-Tsai equations is shown in Figure 4.8. For
low fiber aspect ratios, there is a significant reduction in the longitudinal modulus, but as
the aspect ratio increases the modulus approaches the rule of mixtures. For natural fibers,
typical aspect ratios range from 50 to 500 depending on the length the fibers are chopped

to during processing (Mohanty et al. 2000).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10,

A 1 +Rule ot MlxItures

a 1 +Ha ln-Tsa

n.9i 'p _ A _ _

9.8.. ' ' : '

“’ i

2 .

37::

is;

.2 E

553‘

mi.

3 :

533 l Materthata:

a f GreenHemp/Polyester

'52} E,=706Pa

g I Em=1.5GPI

41; 1 v,=0.10

0‘ 2..
0 100 200 300 400 500

Fiber Aspect Ratio (Ild)

Figure 4.8. Variation of the longitudinal modulus of an aligned short fiber composite with fiber
aspect ratio, using the Halpin-Tsai equation
The strength of an aligned short fiber composite can be estimated using the model
by Pi ggott (1980), which accounts for the fiber aspect ratio and the fiber-matrix
interfacial shear stress. Consider the geometry of deformation in the representative

volume element shown in Figure 4.9. The stiffness mismatch between fiber and matrix

107

(Em << Ef) leads to large shear deformations near the fiber ends but no shear deformations
at the rnide of the fiber (Figure 4.9). The stress transfer between the matrix and fiber
occurs primarily through interfacial shear, which is greatest near the fiber ends, and zero
near the center (Figure 4.10). Conversely, the normal stress in the fiber ramps from a
minimum near the ends to a maximum near the center of the fiber (Figure 4.10). Piggott
developed prediction models for the strength of a short fiber composite and the stress-

strain behavior considering these concepts.

 

    

Fiber Length

Figure 4.10. Single fiber composite element showing the fiber-matrix interfacial stress and the fiber
internal stress for stress transfer by slip

108

As with the strength of a continuous fiber-reinforced composite, in order for the
fibers to have a strengthening effect on the composite material the fiber volume fraction
must be above a critical fraction. The critical fiber aspect ratio for the model by Piggott
is given by:

SC = 0",“, / 22', , (4.2.36)
where z;- is the fiber-matrix interfacial shear stress. The interfacial shear stress may be
assumed to be r,- = O'mM/ 2 when experimental data is not available (Piggott 1980). This
assumption for the interfacial shear stress may lead to an over estimate of the actual
value, but is sufficient for estimations (Piggott 1980).

The composite strength for fiber volume fractions greater that or equal to the

critical ratio is given by:

0 =0',max[V,(1—s, /23)+V,,E,,, /E,], (4.2.37)

where s is the actual fiber aspect ratio (M. For fiber aspect ratios less than the critical
ratio, the composite strength is given by:

0cm = V,sr,. + Vmamm . (4.2.38)

The variation of the longitudinal strength with fiber aspect ratio for a green hemp-

polyester composite using Piggott’s model is shown in Figure 4.11. For low fiber aspect

ratios there is a significant reduction in the longitudinal strength, but as the aspect ratio

increases the modulus approaches the rule of mixtures prediction.

109

 

\l
O

 

 

 

 

 

 

  
     

 

 

 

 

 

 

 

 

. + Rule of Mixtures
: + Piggott Model
A60 3 3 3 :‘r 3
a : .
§ 1
v50 :1,
5 .
2’ i
40 «~
3 ;
to .
£30 :
'o I
,3 20 j 1 Material Data-
!» Green Hemp/Polyester
5 i a... = 500 MPa
4 10 if on", = 70 up.
i , v, = 0.10
o 1 Am 1 Q I
0 100 200 300 400 500
Fiber Aspect Ratio (Ild)

Figure 4.11. Variation of the longitudinal strength of an aligned short fiber composite with fiber
aspect ratio, using the model by Piggott

Piggott also derived a model for the stress-strain curve of an aligned short fiber
composite, which accounts for the fiber aspect ratio and the fiber-matrix interfacial slip.

The stress-strain relation is given as:

2 2
0' ‘(VE +VE )8 -M (4239)
l_ f f m m 1 47.8 1 . .

where 0'1 and 81 are the composite stress and strain in the longitudinal direction,
respectively. Stress-strain curves for varying fiber aspect ratios are plotted in Figure
4.12. For lower fiber aspect ratios both strength and stiffness decrease, and the stress-
strain curve becomes non-linear near failure due to slip between the fibers and matrix
(Figure 4.12). For large aspect ratios (l/d>1000) the results approach a linear curve, as

predicted by the rule of mixtures.

110

 

 

 

   

 

 

 

 

 

 

 

1000 ,
lI—e—uesiooo —e—i/ci=50 ——a—ild=1ol
900 j-
i
800 i-
700 3-
? :
o. 6002»
a :
3 500 gr
g 400}
(I) :
300 ‘5 Mamlal 32:
f Unidirectional FRP
2” r E,=250 GP.
:_ Em=1.SGpa
10° : v,.50%
o‘....: ................... :....:....:..sr
0.000 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008

Strain (mm/mm)

Figure 4.12. Effect of fiber aspect ratio on the stress-strain response of short fiber composites

4.2.3 Randomly Oriented Short Fiber Composites

The fiber alignment in a randomly oriented fiber composites can by truly random
in a three-dimensional sense, or random in only two-dimensions. A three-dimensional
orientation is likely to exist when the length is much less than the thickness of the part, or
when low aspect ratio reinforcement such as whisker and microfibers are used. However,
a two-dimensional fiber orientation is more common in many short fiber composites
where the fiber length is much longer that the thickness of the part. In this situation fiber
orientation in the thickness direction is not possible. The analysis of the isotropic elastic
constants E, V, G, and strength 0', for the two-dimensional case will be presented, as
biocomposite materials tested in the material experimental program fall into this category

of fiber orientation.

111

 

4.2.3.1 Properties for Randomly Oriented Short Fiber Composites: Stijffiiess

Halpin and Pagano (1969) have obtained the isotropic elastic constants of a
randomly oriented fiber composite by approximating the composite as a quasi-isotropic
laminate. Using the invariant properties of the laminate stiffness matrix (Jones 1999),

Halpin and Pagano expressed the elastic constants of the random fiber composite as:

E = 4U, (Ul — U5)/U, (4.2.40)
v = (U, — 2U,)/U, (4.2.41)
G = US , (4.2.42)

where the invariants are in terms of the engineering constants of the equivalent

unidirectional material from which the laminate is constructed. The invariant terms are:

U, = (30,, + 3Q22 + 2Q,2 + 49,, ) / 8 (4.2.43)

U5 = (Q,, + Q22 -— 2Q,2 + 4Q“)/8 , (4.2.44)

with lamina stiffness coefficients:

Qll = El /(1- vnv21 ) (4.2.45)
Q22 = E2 /(1- v,,v,,) (4.2.46)
Q,2 = v,,Q,, = v,,Q,, (4.2.47)
Q...S = 0,,, (4.2.48)

for an orthotropic material. The orthotropic elastic constants E1, E2, V12, and 612 of a
randomly oriented fiber composite can be deterrrrined using a micromechanics approach.
Thus, to model a randomly oriented short fiber composite, rrricromechanics equations that
account for the fiber aspect ratio are used. Halpin and Pagano used the Halpin-Tsai

equations to model the orthotropic elastic properties of a randomly oriented short fiber

112

composite and found the predictions to be in good agreement with experimental results
for a boron/epoxy composite.
Tsai and Pagano (1968) used Equations (4.2.40)-(4.2.42), and related them to the

moduli of an orthotropic material to develop the following approximate expressions:

3 5
E = E El ‘1’ g E2 (4.249)
1 1

These simple equations can be used with micromechanics equations for E1, E2, V12, and
012 that account for the fiber aspect ratio to model a randomly oriented short fiber
composite.

The isotropic elastic modulus of a randomly oriented fiber composite can be
modeled by assuming the composite consists of a large number of superimposed
unidirectional layers. Each layer is characterized by its fiber orientation Band the layer
behavior is fully additive over all layers. Thus, the elastic modulus can be modeled by
geometrically averaging the properties of the unidirectional layers over all the fiber
orientations (Gibson 1994). This approach is equivalent to the “rule of mixtures” for a
randomly oriented fiber composite. Using geometric averaging the modulus is given by

7:12
= [o E(6)d6
1:516

 

, (4.2.51)

where E(6) is the one-dimensional elastic modulus of the B-ply. Assuming a uniform

distribution of fibers over all orientations, Equation (4.2.51) reduces to:

2 n/Z
E = -— j E(6)d6. (4.2.52)
72' 0

113

For situations where the fiber orientation is known to be non-uniforrn, fiber orientation
density functions have been developed for non-uniform orientations (Giurgiutiu and
Reifsnider 1994). Using the one-dimensional macromechanics analysis of layered

composites (Jones 1999), the off-angle lamina elastic modulus is:

—1

E(6) = [Leos4 0 + —l—sin" 0 + {—1— — 233-]0032 (93in2 0] , (4.2.53)
E, E 2 ,2 E,

where the orthotropic moduli for short fiber composites can be used to model a randomly

oriented short fiber composite.

While the previous approach to modeling a randomly oriented fiber composite as
quasi-isotropic larrrinate led to practical equations, the geometric averaging approach
leads to a complicated integral for composite stiffness. The integration of Equation
(4.2.52) can be evaluated using mathematical software (e.g., MATLAB [2003], Mathcad
[2001], Mathematica [2002], etc.). Despite this, geometric averaging can only be used to
model the extensional elastic modulus of a randomly oriented fiber composite, and
cannot model the shear modulus, G, or the Poisson’s ratio, v, as in the quasi-isotropic
model.

Christensen and Waals (1972) also used the geometric averaging approach to
develop closed form equations for the isotropic elastic constants for a randomly oriented

fiber composite. Using a geometric averaging approach, Christensen and Waals found

the elastic modulus and Poisson’s ratio of the randomly oriented fiber composite to be:

E = Ill—("‘2 — uj) (4.2.54)
1
v = “‘ , (4.2.55)

 

2
o

114

where:

0,, + (3 + 2v,2 + 31!,2,)G,,K23

 

 

 

u = 3/8 E + 4.2.56

‘ ( ) ‘ 2 2(G,,+K,,) ( )
2

u, = (1/8)E, - 02” + (1 “5"” + V12)GZ3K” , (4.2.57)

2(G23 + K 23)
and G23 is the shear modulus in the 2-3 plane, and K23 is the bulk modulus in the 2-3
plane with an = 0. Christensen and Waals used micromechanics equations based on

composite cylinders to calculate G23, and K23 as:

 

 

 

 

623 VI
' = 1+ 4. .
0.. Gm +[km +(7/3)G,,,]V,,, ( 258)
o, + 0,, 2(km + (4/3)G,,,)
V!
K,, = km + Gm I3 + 1 v . (4.2.59)

+
k, —k,,, +(1/3)(o, -o,,,) k,, +(4/3)G,,,

 

 

Equations (4.2.54) and (4.2.55), or the equations shown in the previous sections can be
used to calculate the remaining elastic constants E1 and v12. The shear modulus for the
isotropic randomly oriented fiber composite can then be calculated using the extensional
elastic modulus and the Poisson’s ratio by the conventional elasticity relation (Jones
1999):

E

G = . (4.2.60)
2(1+ v)

 

4.2.3.2 Properties for Randomly Oriented Short Fiber Composites: Strength

The geometric averaging approach was also used by Baxter (1992) to model the
strength of a randomly oriented fiber composites. The model is based on the Tsai-Hill

equation, which describes the strength of the B-ply of an aligned fiber composite as:

115

-1/2
0(6) 2 I:—12—cos4 l9 +—li—sin4 6 + [:2 ——1—]cos2 65in2 6] (4.2.61)
0', 0'2 7 0',

where 0'1 is the strength in the fiber direction, 02 is the transverse strength, and ris the
shear strength of the composite. A short fiber composite can be modeled by using an
equation for 01, which accounts for the fiber aspect ratio. The upper limit of composite
strength is calculated by assuming a strong interfacial bond is formed such that the shear
strength and interfacial shear strength are equal to the matrix shear strength:

2' = 2'. = r . (4.2.62)

The same assumption is made for the transverse strength 0'2:

0‘2 = 0mm . (4.2.63)
A more realistic assumption for the transverse strength, which accounts for slip between
the fibers and matrix, is to approximate the fiber as a cylindrical hole in the matrix:

a, = 0 11— 2(V, In)“ J. (4.2.64)

The strength of a randomly oriented short fiber composite is then calculated using
Equations (4.2.62)-( 4.2.64) and a geometric averaging approach similar to Equation

(4.2.52). Thus, the strength is given by:
2 :r/Z
a, =— j 0(6)d6. (4.2.65)
71' 0

The integration of Equation (4.2.65) can be evaluated using mathematical software (e.g.,
MATLAB [2003], Mathcad [2001], Mathematica [2002], etc.).
Lees (1968) also used the geometric averaging approach to develop models for

the strength of randomly oriented fiber composites. Lees assumed there are three failure

116

 

mechanisms according to the Maximum Stress Criterion, each operating over an angle

range as follows:

 

 

for o s 6 s 9,: 0, = a; , (4.2.66)
cos 0
for a, s 6 .<_ 6,: 0,, = —_—T———, (4.2.67)
sm 0 cos (9
02
andfor 023057272: 0( = _ 2 . (4.2.68)
srn 6

Lees then assumed that, for a randomly oriented fiber composite, the average strength

over all angles is given by:

2i
0,:—
72

After evaluating the above integral, Lees found the composite strength to be:

0 =35[1+ ”2 +ln-0—2—glfl]. (4.2.70)

 

 

[9' a; 210+ j”’___.f___do+ I” .022 (16. (4.2.69)
0 cos 6 9: smflcosfl 92 sm 6

 

‘ 7: 0 22

mrmx

Chen (1971) modified Lees’ equation by using a strength efficiency factor to
account for the effect of the fiber aspect ratio on the composite strength. The modified

equation is:

0 =3€[1+—92—+lnmj, (4.2.71)

where the strength reduction factor wis the ratio of the strength of an aligned short fiber

composite to the strength of an equivalent continuous unidirectional composite.

117

4.2.4 Comparison with Measured Material Properties

The randomly oriented short fiber composite models presented in the previous
sections are used to model the properties of the biocomposite material systems tested in
the material experimental program of Chapter 2. The stiffness of the biocomposite
material systems was predicted using the following analytical models: (i) the Halpin-
Pagano quasi-isotropic approximation, (ii)\ the Tsai-Pagano approximate equation, (iii)
the geometric averaging approach, and (iv) the model by Christensen-Waals. For the
stiffness models, the orthotropic elastic constants were calculated using the Halpin-Tsai
equations to account for the fiber aspect ratio. The strength of the biocomposite material
systems was predicted using the following analytical models: (i) geometric averaging
approach, and (ii) the model by Lees. For the strength models, the strength in the fiber
direction was calculated using the model by Piggott to account for the fiber aspect ratio.

The green hemp/polyester biocomposite material system from Chapter 2, which
was manufacturing at three fiber volume fractions (10%, 13%, and 20%), was first used
to evaluate the analytical models for stiffness and strength. Based on the performance of
the analytical models, one model was selected and used to predict the properties of the

biocomposite material systems.

4.2.4.1 Evaluation of Analytical Models: Stiffness

Results from the modulus of elasticity tests for the green hemp/polyester material
system of Chapter 2 are compared with analytical predictions in Figure 4.13 and Figure
4.14. In general, the results from all analytical models are in agreement with the
measured results (Figure 4.15) and show similar trends for increasing fiber volume

fractions (Figure 4.13). More specifically, the prediction of the geometric averaging

118

approach underestimates the measured results, whereas the predictions of the
Christensen-Waals model overestimates the measured results. Both the Halpin—Pagano
model and the Tsai-Pagano model agree well with the measured results, with the Halpin-
Pagano model matching the best for all fiber volume fractions (Figure 4.15). Based on
these results, the Halpin-Pagano model was chosen to model the modulus of elasticity the

remaining biocomposite material systems from Chapter 2.

   
   
 
 
 
 
 

 

 

 

 

 

  

 

 

 

N
O

70
I —o— Halpin-Pagano (Quasi-isotropic Approx.)
I +Teal-Pagano (Approx. Expression)
’3 5° 3’ —e-— Geometric Averaging
a, j —0— Christensen-Waals (2-D Orientation)
g 50 j_ 0 Measured Values
E 1 Material Data:
0 3 Green Hemp/Polyester
:92 4° 2‘ E, -.- 70 GPa
m E... = 1.5 GP.
I Ud 60
'5 3°. =
a
2
i
E

 

 

0.00 0.20 0.40 0.60 0.00 1.00
Fiber Volume Fraction ( -)

Figure 4.13. Comparison of measured elastic modulus of a green hemp/polyester biocomposite with
analytical predictions

119

 

 

O
.

 

—o— Halpin-Pagano (Quasi-Isotropic Approx.)
+Tsal- -Pagano (Approx. Expression)
+ Geometric Averaging
+ Christensen-Waals (2-D Orientation)
0 Measured Values

‘1
1......
v

 

 

 

0|

 

0
.;i

Modulus of Elasticity (GPa)
'0 fi

 

O ‘ , . A . 5

  
 
   

 

  

 

 
 

Material Data:
Green Hemp/Polyester
E , = 70 GPa
E," = 1.5 Gpa

le = 60

 

 

 

 

0.00 0.05 0.10.15

Fiber Volume Fraction ( )

0.20 0.25

Figure 4.14. Comparison of measured elastic modulus of a green hemp/polyester biocomposite with
analytical predictions, zoomed in View

 

 

  

I Measured
Habit-Pager» (OuasHsotropic Approx.)
D Tsai-Pagano (Approx. Expression)
[3 Geometric Averaghg

El Christensen-Wash (2-D Orientation)

 

 

 

Modulus of Elasticity (GPa)

d

 

Vf=0.10 W=0.13

  
     
    
  

 

 

 

Figure 4.15. Results of the analytical models compared with the measured elastic modulus of a green

hemp/polyester biocomposite

120

The Halpin-Pagano model was used with the Halpin-Tsai equations to predict the
elastic modulus of the biocomposite material systems from Chapter 2. The measured and
predicted results are provided in Table 4.2 and shown in Figure 4.16. The analytical
predictions show good agreement with the measured results, except for the E-
glass/polyester composite. The E-glass fibers used were sized (treaded for improved
adhesion) for polyester, which may account for the underestimation by the analytical
model. The analytical predictions also show some deviation from the results of the
unprocessed green hemp/polyester and the raw hemp/polyester composites. These results
may be due to inaccurate measurements of the fiber lengths and diameters, or the

randomness in the properties of the natural fibers.

Table 4.2. Measured modulus of elasticity of biocomposite material systems compared with results
from the Halpin-Pagano quasi-isotropic approximation

 

 

E measured V, L/d E predicted % Error

Sample (G Pa) (%) ( --) (G Pa) (%)
Green Hemp/UPE - 13wt.% 3.81 10 60 3.29 14%
Green Hemp/UPE - 17wt.% 4.40 13 60 3.87 12%
Green Hemp/UPE - 25wt.% 5.23 20 60 5.27 1%
Unprocessed Green Hemp/UPE 8.65 25 125 6.90 20%
Flaw Hemp/UPE 6.15 15 100 4.54 26%
Fiax/UPE 3.37 30 15 3.20 5%
E-Giass/UPE 7.80 15 467 4.96 36%

 

121

 

 

 

3 [ I Measured Halpin-Pagano (Quasi-Isotropic Approx) 1

 

Modulus of Elasticity (GPa)
0 d M w & 0| a: Q G u:

 

 

 

33222
£= :3

Green
HemplUPE ' l‘. ;' 9.51:1}!
25wt.% ll‘n‘ liflwi‘t.‘

age! :232 '5
8 ' 3‘
GEE ii; §§
5‘3 :3 E
D

Figure 4.16. Measured modulus of elasticity of biocomposite material systems compared with results
from the Halpin-Pagano quasi-isotropic approximation

4.2.4.2 Evaluation of Analytical Models: Strength

The results from the tensile strength tests of the green hemp/polyester material
system of Chapter 2 are compared with analytical predictions using geometric averaging
(Baxter 1992) and Lees’ model (1968) in Figure 4.17. In general, the results show
similar trends for increasing fiber volume fractions, but both analytical models
overestimate the measured results (Figure 4.17). The overestimation in the strength may
be due to defects in the surface of the samples, which can lead to premature failure. In
addition, the measured tensile strength may be low due to impurities, such as the wood
like core material present in the natural fibers. These impurities may also lead to
premature failure, or poor fiber-matrix adhesion, which reduce the composite material
strength. The overestimation of the analytical predictions may also be due to the value of

the interfacial shear strength assumed in the model by Piggott. The shear strength at the

122

fiber-matrix interface was assumed to be the tensile strength of the matrix divided by 2,
as recommended by Pi ggott. However, this assumption may have overestimated the
actual interfacial shear strength, which could lead to the overestimation in the tensile
strength of the green hemp/polyester biocomposite.

Even though both analytical models overestimated the measured strength, the
predictions of the geometric averaging approach were closest to the measured strength.
Thus, the geometric averaging approach was selected to model the remaining

biocomposite material systems from Chapter 2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

35
I + Geometric Averaging
: +1.”! . . I I
30.- O MeasuredVaIues . _ . . - ' ' (
A I ' . -
a25' - ' .
E j '
V I
.c 2
620" T"
c .
g
(015 .?
2
310 . MateLIgiDgg:
I2 " ' GmnHemp/Polyester
0~=500MPO
5 am=70Mpl
i Ud=80
o..-.......ififi.-.fi.-..-f..
0.00 0.20 0.40 0.60 0.80 1 .00
Fiber Volume Fraction ( -)

Figure 4.17. Comparison of measured strength of a green hemp/polyester biocomposite with
analytical predictions

The geometric averaging approach was used with the Piggott model to predict the
tensile strength of the biocomposite material systems from Chapter 2. The measured and

predicted results are provided in Table 4.3 and shown in Figure 4.18. Again, the

123

 

analytical predictions show overestimation of the measured results (Figure 4.18), as

expected from the results for the green hemp/polyester results shown in Figure 4.17.

Table 4.3. Measured tensile strength of biocomposite material systems compared with results from

 

 

 

geometric averaging
0.," measured V, L/d 0",, predicted % Error
Sample (MPa) (%) (-- ) (MPa) (%)
Green Hemp/UPE 14.47 10 60 22.03 52%
Green Hemp/UPE 14.91 13 60 22.77 53%
Green Hemp/UPE 16.79 20 60 25.06 49%
Unprocessed Green Hemp/U PE 10.08 25 125 24.86 147%
Raw Hemp/UPE 19.49 15 100 23.26 19%
Flax/UPE 13.31 30 15 22.55 69%
E-GIass/UPE 34.70 1 5 467 2_9.18 16%

 

 

40
I Ilvbasured
35:) IGeometricAveraghg
A .
a .
n. 30»
E i
V I
s 25 r
2’
2 2° 3'
a .
2
To
C
0
l-

 

 

 

Green
HempIUPE- g
17wt.% “H“.
Green
Hemp/UPE-
25M.%

E... E E

Figure 4.18. Measured tensile strength of biocomposite material systems compared with results from
the geometric averaging approach

124

4.3 Cellular Sandwich Structures

The hierarchical, and hybrid sandwich structures tested in the structural
experimental study (Chapter 3) were modeled analytically using conventional mechanics
of materials, and the micromechanical models for the biocomposite material systems. In
the following sections, the procedure for analyzing the cellular beams and plates is

presented, followed by a comparison of the analytical results with the measured results.

4.3.1 Analysis of Cellular Sandwich Structures

The measured flexural stiffness (E1) of the cellular beams and pates is compared
to analytical results, which are obtained using the micromechanical equations of section
4.2 and conventional mechanics of materials (see Hibbler 1994).

The measured flexural stiffness is first extracted from the measured load-
displacement response. From the load-displacement response obtained in the flexural
testing of the cellular beams and plates, the stiffness (force per unit displacement) is

determined as
P
K = — , 4.3.1
5 ( )

where P and 6 are the load and displacement at the mid-span section. The measured
stiffness can be related to the flexural stiffness using equations that relate displacement to
force by the flexural stiffness and the loading geometry. The displacement of a beam
loaded in four-point bending is (Hibbler 1994):

Pa
48E]

6 = (3L2 — 4a2), (4.3.2)

 

125

where L is the span length, a is the side span between the reaction and point load, and P
is the sum of the two point loads. Solving Equation (4.3.2) for E1, and using Equation

(4.4.1), the measured flexural stiffness can be calculated as:
E] = trig-(u2 — 402), (4.3.3)

where K is measured from the load-displacement response, and L and a are known from
the test setup geometry.
The same approach is used to calculate the flexural stiffness of the cellular plates.

The displacement of a beam loaded in three-point bending is (Hibbler 1994):

_ PL3
48EI '

 

(4.3.4)

Solving Equation (4.4.4) for BI and using Equation (4.4.1), the measured flexural

stiffness of the cellular plates can be calculated as:

L3

E1 = K .
48EI

 

(4.3.5)

The flexural stiffness was obtained analytically by separately calculating the
modulus of elasticity of the material and the moment of inertia of the cross-section, and
combining the two terms. The modulus of the biocomposite material systems were
calculated using the Halpin-Pagano (1969) quasi-isotropic laminate approximation
together with the Halpin-Tsai (1969) equations to determine the orthotropic elastic
constants of the material (see Section 4.2).

The moment of inertia of the cellular beams and plates can be calculated using the
moment of inertia of the gross cross—section, and subtracting the moment of inertia of the

voids in the cellular structure using the parallel axis theorem:

126

1=i+Ad§ (4am

The measured dimensions of the manufactured cellular beams and plates were
used to calculate the moment of inertia of each test unit. The face sheet layers in the
hybrid cellular plates were transformed to an equivalent biocomposite material for the
moment of inertia calculations. The modulus of the face sheet materials was predicted
using an appropriate micromechanics material model (e.g., rule of mixtures for the
unidirectional carbon).

The analytical failure load for the cellular beams and plates was obtained using
conventional mechanics of materials (Hibbler 1994).

The flexural formula can be used to relate the failure stress in the extreme fiber of
the section to the failure moment at mid-span as follows:

44y.
6,: 2 , (41D

 

where a} is the failure stress, Mf is the moment capacity or failure moment, and y,,, is the

distance to the extreme fiber. Noting that 0: (E, the failure strain in the extreme fiber is:

M'y
f m
8 =——, 4.3.8
, El ( )

where E! can be calculated from the measured flexural test results, y,,. can be measured
from the test unit, and the failure strain can be obtained from the measured test results of
Chapter 2 (see Table 2.9).

The moment in Equation (4.4.8) can then be related to the applied load using
basic statics. For a simply supported beam loaded in three-point bending, the bending

moment at mid-span is:

127

M = —. (4.3.9)

Rearranging the terms in Equation (4.4.9) and assuming linear-elastic behavior up to

failure, the failure load is given by:

P = f . (4.3.10)

 

Combining Equations (4.4.8) and (4.4.10), the failure load for the cellular plates is

 

P; (4.3.11)

= f Lym .
For the cellular beams loaded in four-point bending, the bending moment at mid-span is

M =%, (4.3.9)

where a is the side span between the reaction and point load, and P is the sum of the two
applied point loads. Using the same approach followed to obtain Equation (4.4.11), the
failure load for the cellular beams is:

_8 2E1
’ fay...

(4.3.13)

 

4. 3.2 Comparison with Measured Values

The measured and analytically predicted flexural stiffness of the cellular beams
are provided in Table 4.4, while the measured and analytically predicted strengths are
provided in Table 4.5. The analytical predictions for the flexural stiffness are in good
agreement with the measured results. One source of error in the analysis in both the
cellular beams and plates could be in the calculation of the moment of inertia. In the
moment of inertia calculations it was assumed that the original cellular architecture

remained centered, while the excess material was evenly distributed to the top and bottom

128

 

of the section. This actually may not be the case, which would lead to errors in the
calculated moment of inertia of the section. Additional errors could be due to the
uncertainty in the analytical models used to predict the modulus of the biocomposite

material systems.

Table 4.4. Measured versus predicted flexural stiffness of the biocomposite cellular beams

 

Fiber Measured Predicted

 

Material Beam Wt. Fraction E! El Error
ID (%) (GPa-mm‘) (GPa-mm‘) (%)
G.Hemp/UPE CB1 10 188,886 179,805 5%
G.Hemp/UPE 0132 10 250,557 231,178 8%
G.Hemp/UPE 033 10 217,058 219,180 1%

 

Table 4.5. Measured versus predicted flexural strength of the biocomposite cellular beams

 

Fiber Measured Predicted

 

Material Beam Wt. Fraction PM Pm, Error
'0 (%) (N) (N) (%)

G.Hemp/UPE C81 10 670 532 21%
G.Hemp/UPE 082 10 813 659 19%
G.Hemp/U PE CBS 1 0 677 625 8%

 

The measured and analytically predicted flexural stiffness of the cellular plates are
provided in Table 4.6, while the measured and analytically predicted strengths are
provided in

Table 4.7. The results follow the same trend as in the cellular beam results, i.e.,
the flexural stiffness predictions are in agreement with the measured results but the
strength predictions underestimate the measured results. The same sources of error in the

cellular beams apply to the cellular plates.

129

Table 4.6. Measured versus predicted flexural stiffness of the cellular plates

 

Fiber Wt. Measured Predicted

 

Beam Fraction El El Error
Material ID (%) (GPa-mm‘) (GPa-mm‘) (%)
G.Hemp/UPE CP1 10 160,466 140,769 12%
G.Hemp/UPE - Top and Bot. Glass CP1-G 10 126,358 137,602 9%
G.Hemp/UPE - Bottom Carbon CP1-C 10 233,919 213,159 9%
G.Hemp/UPE - Top and Bot. Jute CP1-J 10 122,007 106,633 13%
Raw Hemp/UPE CP1 10 127,746 146,345 15%
Raw Hemp/UPE CP2 10 198,210 199,746 1%
Raw Hemp/UPE - T&B Carbon CP2-C 10 584,919 512,844 12%
FIax/UPE CP1 20 77,643 79,871 3%
Flax/UPE - T&B Jute CP1-J 20 102,523 97,205 5%

 

Table 4.7. Measured versus predicted flexural strength of the biocomposite cellular plates

 

Fiber Wt. Measured Predicted

 

Material Beam Fraction Pm, Pm, Error
ID (%) (fl) (Ni 4%)

G.Hemp/UPE CP1 10 1,483 1,026 31 %
G.Hemp/UPE - Top and Bot. Glass CP1-G 10 2,237 2,759 23%
G.Hemp/UPE - Bottom Carbon CP1-C 10 10,397 6,569 37%
G.Hemp/UPE - Top and Bot. Jute CP1-J 10 2,197 1,479 33%
Raw Hemp/UPE CP1 10 1,859 1,055 43%
Raw Hemp/UPE CP2 10 2,557 1,308 49%
Raw Hemp/UPE - T&B Carbon CP2-C 10 11,010 13,449 22%
Fiax/UPE CP1 20 1,172 686 41%
Flax/UPE - T&B Jute CP1-J 20 1,980 1,335 33%

 

4.4 Flexural Performance

Engineering design generally seeks to maximize some aspect of performance.
Here, performance will mean maximum flexural stiffness or strength for a given weight,
but in general performance can be any combination of properties. The performance for a
load-bearing component depends on three things: the mode of loading, i.e., tension,
bending, twisting, the properties of the material, and the shape of the section (Ashby

1991).

130

 

Many materials are available for engineering design, each with a wide range of
attributes. The performance of viable materials for a given design criteria can be
compared by combining material properties to form material indices. Material indices
are groupings of material properties which, when maximized, maximize some aspect of
performance. The shape of a beam will affect the performance of the beam in bending.
That is, hollow box or tube sections are more efficient than solid sections. The efficiency
of different shapes can be quantified by defining a dimensionless shape factor. The value
of the shape factor is not dependent on size, only on the shape of the section. Indices for
selecting the performance-maximizing combination of material and shape are defined
next using both material indices and shape factors.

In the following sections, material indices, shape factors (material indices that
include shape), and micro-structural shape factors for bending stiffness and strength are
derived. The derived factors are followed by the measured flexural performance of the

cellular beams and tested in the experimental program.

4. 4.1 Material Indices

Material properties can be combined to form material indices, which are
groupings of material properties that, when maximized, maximize some aspect of
performance. The indices are derived from the design requirements for a component
based on the function, objective, and constraint.

Consider first the material index for a light, stiff beam of stiffness KB, length l,
and of minimum mass m. For the purposes of this derivation the section of the beam will
be square (Figure 4.19). However, the results are applicable to any cross-section. The

mass of a beam of length l and cross-sectional area A is

131

m = Alp , (4.4.1)
where p is the material density.

The stiffness constraint requires that the bending stiffness of the beam be

_£_qm
”6‘13

 

, «43

where C) is a constant that depends on the loading configuration, E is the elastic
modulus, and I is the moment of inertia about the axis of bending. The moment of inertia
of a square section is given by:

_ 2 _ b/2 2 . _b‘ _A2
I —.[y (liq—[buy b d)’ —-1-5—-l—i-. (4..43)

The constraints are the specified stiffness K3 and length I, while the section area A is free

 

 

 

 

 

 

 

 

 

 

 

to change.
lame, F )(V
/ b /
r” ‘ ”WW/4 4 .
f \ ///
L/ J— Deflection, a 1A,“, A
if? L =J—1 Moment of inertia, lxx

Figure 4.19. A beam of square section, loading in bending about the x-x axis

The mass of the beam can be reduced by reducing A, but only as long as the
stiffness constraint is met. Using Equations (4.4.2) and (4.4.3) to eliminate A in Equation

(4.4.1) gives:

 

rrz

12KB 3( .0 J
= -————- l . 4.4.4
m [ C11 ] Em ( )

In Equation (4.4.4) the first bracketed term is a functional requirement. The next term is

the specified geometry (the length l of the beam). The last bracketed term defines the

132

material properties. The most efficient materials for a light, stiff beam are those with

large values of the material index (Ashby 1999):

1/2
114;, = E , (4.4.5)

r0

 

where the superscript e denotes elastic and the subscript B denotes bending.
Now, consider the material index for a light, strong beam of length l and
minimum mass m to support a specified load F without failing (Figure 4.19). The

maximum bending stress is given by
0 = ———"'— , (4.4.6)

where M is the bending moment, I is the moment of inertia of the section about the axis
of bending, and y,,, is the distance from the neutral axis to the extreme fiber of the beam.
The beam will fail when the stress in the extreme fiber exceeds the strength of the
material 0;. Using Equation (4.4.6), the bending moment when the failure stress is
reached is given by:

10,
. (4.4.7)

ym

f

 

Noting that the bending moment is equal to the applied load times the length of the beam

divided by a constant that accounts for the loading configuration,

 

Ffl
Mf = 'C—’ ,
2
the failure load is:
I0,
F, = C2 , (4.4.8)
M

133

 

where C2 is a constant that depends on the loading configuration. For the case shown in
Figure 4.19, C2 = 4, and y,,, = b/Z. Using Equations (4.4.8) and (4.4.3) to eliminate the

area A from Equation (4.4.1) gives:

6 F 2/3
m = [Fr—2f) r3[—%,7]. (4.4.9)
2 0']

This is the mass of the beam required to support the failure load. Again, the first
bracketed term contains the functional requirements (F f and l), the middle term represents
the specified geometry, and the third bracketed term contains the material properties.
The most efficient materials for a light, strong beam are those with large values of the

material index (Ashby 1999):

0.213
Mg: ’ , (4.4.10)
,0

 

where the superscript f denotes failure and the subscript B denotes bending.

A comparison of the performance of common materials in bending is given in
Table 4.8. Other material design criteria, such as cost, fatigue, and energy absorption,
lead to many other indices (Ashby 1999). The material indices work well for comparing
materials, but they ignore the shape of the component. Thus, for materials with different

cross-sectional shapes other factors must be used to quantify efficiency.

Table 4.8. Material properties and material indices for minimum weight

 

 

Density Modulus Strength E1/2/p 63%
Material ,0 (Mg/m3) E (GPa) 0 1 (MP3)
Wood (spruce) 0.49 15 45 7.9 25.8
Concrete 2.30 25 30 2.2 4.2
Carbon FRP 1 .60 100 450 6.3 36.7
Glass F HP 1 .78 28 300 3.0 25.2
Steel 7.85 210 1100 1.8 13.6
Aluminum 2.70 69 130 3.1 9.5

 

134

4.4.2 Shape Factors

Sections of shaped material carry load more efficiently than solid sections do.
Here, shaped means that the cross-section is formed in a tube, a box, an I-section, or the
like. Efficient means that for a given loading condition the section uses as little material
as possible, and is therefore as light as possible. To measure the shape and efficiency of
a section, for a given mode of loading, shape factors have been developed by Ashby
(1999). A shape factor is a dimensionless number that characterizes the efficiency of the
shape of a section, regardless of scale, for a given mode of loading.

Consider first the shape factor for the elastic bending of a beam. For a beam of

length l and elastic modulus E, the bending stiffness is:

CEI
KB: 23

 

, (4.4.11)

where C) is a constant that depends on the loading configuration. The shape of the
sections influences Equation (4.4.11) through the second moment of area, I, about the
axis of bending, or more commonly know as the moment of inertia.

Ashby (1999) defines the shape factor for elastic bending as the ratio of the
stiffness KB of the shaped beam to that of a solid circular section K g (with moment of
inertia 1") with the same cross-section A, and thus the same mass per unit length. Using
Equation (4.4.11) the shape factor is given by:

_K,, 1
K; 10'

 

e;
For a solid circular section of area A the moment of inertia is:

4 2
“a; A
4

= —. 4.4.12
4% ( )

135

 

rm—

 

Thus, using Equation (4.4. 12), the shape factor for elastic bending is:

, 4n!
¢B = A2 ’

 

(4.4.13)

where the superscript e denotes elastic and the subscript B denotes bending.

Now, consider the shape factor for failure in bending. Failure of a material can
have many definitions. For a ductile material, failure may occur at the onset of plasticity,
when the material first reaches its yield strength 0,. For a brittle material failure occurs
when the stress exceeds the fracture strength 0),. When fatigue is an issue, failure occurs
when the endurance limit a; is exceeded. The shape factor for failure in bending covers
all three modes of failure. Thus, the symbol 0, will be used to denote the failure stress
for yielding, fracture, or fatigue failure.

In bending, the stress is largest in the extreme fiber of the beam, ym, which is the
surface of the beam that lies furthest from the neutral axis. The maximum bending stress

is given by

_ My.

, 4.4.14
I ( )

0'

 

[‘1
s

where M is the bending moment and I is the moment of inertia of the section about the
axis of bending. The shape enters Equation (4.4.14) through the section modulus, S = I/
ym. If the maximum stress in the beam exceeds the failure stress of the material, the beam
will fail. The bending moment when the failure stress is reached is given by:

M, = 20,. (4.4.15)

Ashby (1999) defines the shape factor for failure in bending as the ratio of the
failure moment My of the shaped section to that of a solid circular section with the same

cross sectional area A. Thus, the shape factor is

136

 

l

 

M: ’= . m4m)

 

 

For a solid circular section of area A the section modulus is

 

72. A3/ 2
S”=—¢3= . «Aim
4 4J7?
Thus, using Equation (4.4.16) the shape factor for failure in bending is:

where the superscript f denotes failure and the subscript B denotes bending. The two
expressions for q); , and e; have the same form, and are often similar in value.

Both shape factors are dimensionless, and depend only on shape. That is, large
and small beams have the same shape factor value if their sections have the same shape,
and are proportional. Solid, symmetric sections (circles, squares, etc.) will have shape
factors close to l, but elongated or I-shapes can have shape factor values of 15 or more
(Table 4.9). These shaped sections are more efficient than the solids shapes in that less

material is required to achieve the same bending stiffness and strength.

137

 

Table 4.9. Section shape factor (sections of equal area)

 

 

Shape <0; <0],
Y
/%x
1.0 1.0
. 17.4 6.2
r/
/ 1.1 1.2
%
g.) 13.8 6.3
2M”? 18 0 6
/ d
17.0 7.8

 

- Hollow sections: thickness (t) is constant
- Asymmetrical sections: long axis is 1.5x the short axis

4. 4.3 Performance I ndices with Shape Factors

The performance-maximizing combination of material and shape for a given
mode of loading involves both the material indices and shape factors. Ashby (1999)
derived indices for comparing the performance of beams of different materials and
shapes.

Consider first the selection of a material for a light stiff beam of stiffness K3,

length l, and minimum mass m, allowing the section-shape to be variable. Using the
bending stiffness given in Equation (4.4.11) and replacing I by the shape factor (I); in
Equation (4.4.13) gives:

C A2
K =-——‘—E ‘—. 4.4.19
3 4n ¢B l3 ( )

Using Equation (4.4.19) to replace A in Equation (4.4.1) gives the mass of the beam as:

138

 

 

 

 

472K 1/2 2 1/2

m = —i I3 '0 . (4.4.20)
C,l (03E

Thus, the best material-shape combination for a light stiff beam is that with the greatest

value of the index

, /2
M. = LL53“? , (4.4.21)

p
Notice that, for sections of the same shape (i.e. constant (2);) the best choice for the
lightest beam is the material with the largest value of Em/p, as shown in section 4.4.1.
Consider now the selection of a material for a light strong beam of length l and
minimum mass m, to support a specified load F without failing, allowing the section-

shape to be variable. Using the failure moment in Equation (4.4.15) and replacing S by

the shape factor e); in Equation (4.4.18) gives:

0
M, = —f¢g.«12'3. (4.4.22)

4J7:

Using Equation (4.4.22) to replace A in Equation (4.4.1) gives the mass of the beam as:

M 2r3 3,2 2/3
m=[4JZ—IT’-] If” J . (4.4.23)

rt 0,

 

Then, the best material-shape combination for a light, strong beam is that with the

greatest value of the index:

f 2/3
M2 = (31%—L. (4.4.24)

,0

For sections of the same shape (i.e. constant ¢Bf ) the best choice for the lightest beam is

the material with the largest value of (pm/p, as shown in section 4.4.1.

139

cl .0 “new

 

 

4.4.4 Micro-structural Shape Factors

The efficiency of a macroscopic shape can be improved by introducing shape at a
small scale, microscopic shape. For example, a rectangular section made of a cellular
material (microscopic shape) will have higher efficiency than a solid rectangular section
(macroscopic shape). The additional efficiency of the micro-structural shape is
characterized by micro-structural shape factors (Ashby 1999).

Microscopic shape can be achieved through cellular structures, which are solids
made up of an interconnected network of solid struts. The most important characteristic
of a cellular structure affecting its mechanical properties is the relative density, p79,, i.e.

the density of the cellular material, p', divided by the density of the solid from which the

cell walls are made, ,0, (Gibson and Ashby 1988). The higher the relative density for a
cellular structure, the more solid material in the structure, resulting in increased
mechanical properties at an increased weight.

Consider first the micro-structural shape factor for bending stiffness. Ashby
(1999) defines the micro-structural shape factor 1,1/similar to the macroscopic shape factor
(0. If the beam in Figure 4.19 is composed of a micro-structured material, the second

moment of the area is:

e 0 4
1 =E—jm yzb-d =11, (4.4.25)
p5 —-b12 p5 12

where a differential element of area dA contains an area of solid materials (,0. /)05 )bdy.

The total shape factor that includes both the micro-structural and macro-structural shape

of the beam is thus:

140

J LII

 

 

 

(¢;)TOT_4M:(£)[ 1 J (4426)

A2 4 272.
where the area of the micro-structured material is A = (,0' /,05 )b2. The term in the first

brackets in Equation (4.4.26) is simply the shape factor (’3 for the macro-structure of the

beam. The second term in brackets is the micro-structural shape factor for bending

stiffness:

 

w; = p! . (4.4.27)

In other words, the micro-structural shape factor here is the inverse of the relative
density. Note that, in the limit, for a solid micro-structure (p' = p,), w; takes the value

of l as it clearly should. Also, notice the total shape factor is the product of the macro-

structural and micro-structural shape factors:

)TOT

(18‘. = r); r;- (4.4.28)
Consider now the micro-structural shape factor for bending strength. If the beam

in Figure 4.19 is composed of a micro-structured material, the section modulus is:

t

r 4
S =.L:£._2__ 134,3. (4.429)

where the moment of inertia I is given in Equation (4.4.25). The total shape factor that

includes both the micro-structural and macro—structural shape of the beam is thus:

 

1r2
(¢;)TOT : HES :EJELIZS] ’ (4.430)

where the area of the micro—structured material is A = (p'/ps )bz. The term in the first

brackets in Equation (4.4.30) is simply the shape factor 0); for the macro-structure of the

141

.. 1.er

 

 

 

beam. The second term in brackets is the micro-structural shape factor for bending

strength:

 

1/2
(a; =[p,] . (4.4.31)

The micro-structural shape factor here is proportional to the inverse of the relative
density. In the limit, for a solid micro-structure (p' = pg), ill; takes the value of l, as it

should. As with the case of bending stiffness, the total shape factor is the product of the

macro-structural and micro-structural shape factors:

)TOT

(11>; = (ii if; . (4.4.32)

Both stiffness and strength micro-structural shape factors show that by
introducing micro-structure to a solid shape the performance is increased, resulting in a
lighter more efficient structure. This idea of micro-structuring could be extended further
by introducing cell walls that are themselves micro-structured, thus creating a structural
hierarchy. Such efficient structural hierarchy can be found in nature (Gibson and Ashby
1988, Gunderson and Thorp 1993), but there are limits on the level of hierarchy that can
be achieved in man-made structures due to manufacturing difficulties and the
accompanying manufacturing costs. Nevertheless, the use of hierarchical structures can
further improve the performance of a structural member resulting in light efficient
structures.

The effect of micro-structure can also be shown by including the microscopic
shape factor into the material performance indices. Considering a cellular structure and

using the micro—structural shape factor, the performance index for maximum stiffness of

a given weight for a cellular beam is given by Equation (4.4.5) to give:

142

 

 

 

 

1/2 * 1/2 1/2
M; =5 ____(E ) =5: [1.0;], (4.4.33)

where the modulus of the cellular material is equal to the modulus of the cellular material

times the relative density (Huang and Gibson 1995).

4.4.5 Measured Flexural Performance

The relative performance index of a micro-structured beam using Equation
(4.4.33) divided by the performance of a beam of solid material is shown in Figure 4.20.
The figure shows the theoretical curve for the biocomposite material system, and is
accompanied by the experimental data from the biocomposite cellular beam tests
(Chapter 3). Due to manufacturing constraints in the experimental studies, the relative
densities studied were very similar. Thus, the experimental data points are only in a
limited region of the analytical curve. Nonetheless, the data is consistent with the
predicted performance, and shows that it is possible to produce cellular beams with
higher values of the performance index than beams of solid material. Further
improvements in the mechanical efficiency of cellular beams (and plates) can be achieved

at lower relative densities (Figure 4.20). For example, at a relative density of 0.20, the

performance index of a cellular beam to a fully dense beam is (E " 2 / p) /(E g’ 2 / p5 ) = 2.2.

143

 

 

 

 

 

 

 

 

 

 

 

 

5 1
1 Analytical
,. . 0 CellularBeami
§ 4 A CellularBeemz
u 4 " 0 Cellular Beams
5 A I
o l .
0 V .
C
a 3'33’
§g. ;
t In .
o \ 2i.
Q Q 4
\ .
gs .
E “13» "‘
E 1
0-2-
0.0 0.2 0.4 0.6 0.8 1.0

Relative Density, p/p. (-)

Figure 4.20. Effect of relative density on the relative performance

The measured material performance index, shape factor, and performance index
with shape factor for minimum weight, for a given flexural stiffness is shown in Table
4.10. The measured values show the effect of the hierarchical distribution on the
performance of the cellular beams. Recall that, cellular beam 1 (CB1) was manufactured
with a periodic cellular architecture, while cellular beams 2 and 3 (CB2, and CB3) were
manufactured with hierarchical cellular architectures (see Figure 3.4). Again, the
manufacturing constraints limited the design of the hierarchical architectures. Thus, the
flexural performance is similar for the three cellular beams. Nonetheless, the data shows
an improvement in the flexural performance achieved by using efficient cellular
structures.

The most efficient cellular beam tested (in terms of flexural stiffness) was CB3,

which had a material index of 2.67, shape factor of 0.67, and a material index with shape

144

 

 

 

factor of 2.18. The other hierarchical cellular beam tested, CB2, showed the lowest
flexural performance. The results indicate that the hierarchical architecture of CB3 was
more efficient than CB2 in concentrating material near the top and bottom of the beam.
Thus, the use of hierarchical structures can improve the flexural performance of a cellular

beam, provided the hierarchical architecture is correctly designed.

Table 4.10. Measured flexural performance of the biocomposite cellular beams

 

 

Fibers
by Wt. Em/p Shape Factor ¢ (¢ E) ”27p
Material ID (%) (spam/149m?) (j (G Pam/Mg/ma)
G.Hemp/UPE CB1 10 2.64 0.65 2.14
G.Hemp/UPE CB2 10 2.33 0.51 1.66
G.Hemp/UPE CB3 10 2.67 0.67 2.18

 

4.5 Hierarchical Cellular Structures

An analysis procedure for the analysis of hierarchical cellular structures was
developed through discretization of the material continuum present in the cellular
structure. As show in Figure 4.21, a biocomposite hierarchical cellular plate can be
thought of as being composed of a series of bonded layers, or lamina, each of them
featuring a particular material cellular architecture. Thus, the properties of the entire
hierarchical cellular plate, or laminate, can be determined by integrating the properties of
each one of the representative cellular material lamina. Each of the cellular lamina layers
is considered to feature a governing cellular architecture such that its properties can be
determined. The properties of a two-dimensional cellular material can be determined
form the cellular microstructure, which includes the material properties of the solid
material making the cellular solid and the architecture of the cells. Finally, the properties

of the solid biocomposite material can be determined using the properties of randomly

145

 

 

oriented short fiber composites. The properties of the biocomposite will depend on the
individual properties of the fibers and resin, and the percentage of fibers to resin in the
composite.

The structural performance of biocomposite hierarchical cellular plates can thus
be determined by integrating the above mentioned levels of analysis (Figure 4.21). This
analysis may be used in lieu of the methods presented in section 4.3. However, due to
manufacturing constrains that limited the level of hierarchical structure that could be
achieved in the cellular beams and plates (Chapter 3), the proposed analysis method
could not be verified.

In the subsequent sections a summary of each analysis level is presented
beginning with the analysis biocomposite material systems, followed by mechanics of
cellular solids, micro-mechanics of a lamina, and lamination theory. The limitations of

the analysis procedure are also discussed.

Biocomposite Materials

 

+ ” %
Natural Cross-Linked
Fibers Polymer Cellular Structures

Hierarchical Cellular Plates P mes omanm, 0mm

 

 

 

 

 

Fiber Commgites

Mechanics 01 Cellular Solids

 

Lamination my)! Cellular Lamina

    
 

(01.201010310103101?

Micro-Mechanical Behavior ot a Lamina

Figure 4.21. Overview of the analysis process for hierarchical hybrid cellular biocomposite beams
and plates

146

YV-fi‘r alive—$1

.‘f 11's

 

 

4.5.1 Analysis

The proposed analysis procedure begins at the material level with the analysis of
the biocomposite material system. The analysis of randomly oriented short fiber
composites presented in Section 4.2 can be used to model the biocomposite material
system. Based on the comparison with measured results, the quasi-isotropic laminate
approximation by Halpin and Pagano (1969) can be used with the Halpin-Tsai (1969)
equations to determine the elastic constants of the biocomposite materials.

The next level of analysis is at the microstructural level, where mechanics of
cellular solids are used to model the properties of the cellular material layers within the
hierarchical structure. While there are several approaches to the mechanics of cellular
solids, Gibson and Ashby (1988) have effectively shown that the mechanical properties
of cellular solids can be effectively described using structural mechanics to determine
their properties such as stiffness and strength.

The properties of a 2-D honeycomb can be described by analyzing the response of
a unit hexagonal cell (Figure 4.22). The mechanical response of the cellular solid can
then be studied by consideration of its interconnected array of one-dimensional members
(Figure 4.22). By loading the unit cell and calculating the resulting deformations in the
cell walls Gibson and Ashby (1988) derived equations for the elastic constants of a
cellular material. The results are provided in the following equations, where the
superscript * refers to the cellular material and the subscript S refers to the solid material

from which it is made:

E," =ES£—==ES% (4.5.1)

147

 

int-w...

 

 

E“ -E (5)3 6°50 (452)
2 SI (h/l+sin6l)sin26 "

 

3 .
._ t (h/l+s1nt9)
vf, = vs (4.5.4)
. . E;
V = V —* 4.5.5
21 12 E1 ( )
0,", = 0, 0.5779 (4.5.6)

The most important characteristic of a cellular structure affecting its mechanical
properties is its relative density, p205, i.e., the density of the cellular material, p‘, divided
by the density of the solid material from which the cell walls are made, ps. The higher
the relative density for a cellular structure, the more solid material in the structure,

resulting in increased mechanical properties, albeit with increased mass.

 

 

Figure 4.22. Unit cell model for the mechanical behavior of cellular solids

Beams and plates structures with repeating layers of cellular material can be

treated as laminated structures (Huang and Gibson 1995) where each layer maybe defined

148

 

 

by a characteristic cell architecture. This idealization applies to cellular beams and plates
with periodic cells that are oriented in different directions about the longitudinal axis of
the element and/or hierarchical cellular structures that feature different levels of cellular
architectures along the member depth.

Assuming a plane stress condition for each characteristic layer of cellular
material, a lamina of cellular material can be described using the equivalent properties
obtained from the mechanics of cellular solids equations proposed by Gibson and Ashby
(1988), as previously discussed. The stress-strain relation for the single cellular material
layer (lamina) with respect to the material 1—2-3 coordinate system, assuming that the

response is at most orthotropic is given by (Jones 1999):

* ’ $
0'1 Q11 Q12 Q16 51

02 = 052 QE. 62 . (4.5.7)
712 Symm. Q66 712
where
. E” . E"
Q11 = ———.l—.—‘ Q22 =_'.2—.—‘
(1‘ V12V21) (1 ’ V12V21)
. v' E‘ . .
Q12 = 2' ' Q... = 6,2. (4.5.8)

(1 " V1261)

The structural properties of a hierarchical cellular beam or plate can then be found
by adding the individual stiffnesses of the different characteristic cellular material layers
over the depth of the structure. Each layer may have different material properties due to
their constituent material, or different cellular architecture or cell orientation. In order to

be able to add the stiffnesses of each layer, their properties with respect to a common

149

 

 

structural coordinate system (x-y-z) must be determined. This is done by rotating the

base material properties given by Equation (4.5.8) through a transformation matrix.

Thus, the material response for the layer about any coordinate system [5‘] within the

plane is defined by:
an —Q—1.1 Q12 :Q:l; EL!
0y, : Q22 Q26 5,, 1 (4.5.9)
7,, Symm. Q6; 7,,
where
15'1= [TillQ’sz'l (4.5.10)
and
00326 sinzd —200868in0
[72}: sinzd 003261 ZCOSdsinB . (4.5.11)

cosdsind —cos€sin0 (coszd-sinzd)

Addition of all the layer stiffnesses over the section depth through classical

lamination theory (Jones 1999) leads to the section properties of the laminated plate, or

the so called [A,B,D] matrix:
= "(Qij )k (Zk— 21(4) (4.512)
k=1
1"
B,- =-2- (01,-).(21— —zi-.) (4.5.13)
kzl
1" _
0,, =1 (Q ),(z,3—z,§_,). (4.5.14)

where i = l, 2, 3 and k = 1, 2, 3...n, for n number of layers. The geometry of the

laminate structure is shown in Figure 4.23. In the [A,B,D] matrix [A] is the extensional

150

stiffness, [B] is the bending-extension coupling stiffness, and [D] is the bending stiffness.
Note that the terms are plate stiffnesses, so that in order to obtain the flexural stiffness for

a beam the stiffness terms must be multiplied by (1 - 1}).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1
NZ 2
20
Z1
22 Middle Surface 1
zk -1 t
zrt
k V z tk 2" 1
i z,
n
L Layer Number

Figure 4.23. Section geometry of a general laminate

The section properties for hierarchical cellular structures can thus be obtained by
considering the different scales of the material continuum, form the microstructural

mechanics of the randomly oriented short fiber reinforced biocomposite to the cellular

 

material arrangement, to the arrangement in the hierarchical structure. The analysis
method was integrated into a custom program and can be used in optimization strategies
to determine efficient hierarchical structures for improved efficiency of biocomposite
cellular structures. However, this approach could not be verified with the experimental
data from the cellar beams and plates (Chapter 3) because of limitations in the procedure,

as discussed next.

151

4.5.2 Limitations

The proposed analysis procedure for hierarchical cellular structures is limited by
assumptions made in the cellular material equations. The mechanics of cellular solids
equations for the elastic constants of a cellular material are based on analyzing a unit cell
and assuming average properties. This assumption is reasonable when there are many
layers of cellular lamina in the hierarchical structure. However, when there is a small
number of cellular lamina in the structure the averaged cellular material does not account
for the actual distribution of material through the section. This issue leads to errors in the
moment of inertia of the lamina, and hence the flexural stiffness of the hierarchical
structure. This concept is illustrated in (Figure 4.24).

In Figure 4.24 the flexural stiffness of cellular structures with varying numbers of
cellular lamina was calculated using the proposed analysis procedure (Elmysis), and using
the conventional mechanics of materials approach of Section 4.3 (Elmhamcs). The
flexural stiffness using mechanics of materials divided by the flexural stiffness using the
proposed analysis procedure is plotted against the number of layers of cellular lamina in
the hierarchical structure (Figure 4.24). The results show that for hierarchical structures
with four or more layers the proposed analysis procedure agrees with the results from
mechanics of materials, but for less that four layers the proposed procedure is inaccurate.
Thus, the proposed analysis procedure is limited to cases of hierarchical cellular
structures with four or more layers. Because this was not the case in the cellular beams
and plates tested in the structural experiment study (Chapter 3) the proposed analysis

procedure could not be used. Nevertheless, the proposed analysis procedure is still a

152

 

 

viable option for performing parametric, and optimization studies, where hierarchical

cellular structures may be discretized with many different layers of cellular material.

 

 

    
  
    

 
 
   
 
 

 

 
 

 

  

 

 

 

 

 

 

 

 

 

 

1.50 4
Cellula Stmcture
1.40 :5 7’47 Icy/1017'! / III/I/roI/l Ir” a
4% 120 3. 7"'// "’¢””//””/ :W/e/ "
e, . e/o/o/o/o/o/o/o nit... [
g 1.10 -
m 1,00 : .- ii.-.-_....--_-----..-
0.90 3- 1
. l
0,30‘...4-..-...2......4_.....+ ........................
0 1 2 3 4 s e 7 8 9 10

Number of Layers

Figure 4.24. Accuracy of the proposed analysis procedure for varying number of cellular layers

4.6 Prestressed Hybrid Sandwich Structures

Thermal induced deformations can occur in composite materials where two or
more materials are co-cured and then cooled down to service conditions. The different
materials bond together at a high temperature. After curing, the composite is cooled to
service temperatures. The temperature change induces residual deformations (strains) in
the structure as a consequence of the different “effective” thermal coefficients of
expansion between the materials constituting the composite. The difference in
“effective” thermal coefficients of expansion can be due to material composition or the
material relative orientation. These temperature induced deformations create stresses and

strains in the section with no application of external load, thus creating a prestress state.

153

 

Temperature induced strains (and stresses) are typically a negative effect and their effect
is minimized through design, i.e. careful arrangement of the material and its orientation.
However, this effect can be beneficial if strategic material arrangement can lead to an
advantageous prestressed state, i.e. opposite to the stress state induced by mechanical
loads. Thus, when mechanical (external) loads are applied to the structure, the materials
must release the temperature strains and then undergo deformation in the opposite stress-
strain regime. This increases the strain range the material in the structure must deform
through before failure, thus increasing the capacity of the structure.

The biocomposite hybrid cellular plates manufactured with carbon fiber face
sheets are susceptible to temperature induced deformations due to the small coefficient of
thermal expansion of the carbon fibers compared to the biocomposite material system and
the integral bond between the two fiber-reinforced composites. In the following sections
analytical models are provided, which can be used to calculate the temperature induced
deformations, stresses, and strains. The prediction models are compared with the
measured deflection in the hybrid cellular plate with a bottom carbon face sheet, and used
to investigate the prestressing effects in the hybrid cellular plates with carbon face sheets
tested in the structural experimental program (Chapter 3). The temperature prestress
effects are then investigated analytically for a full-scale hybrid cellular plate to evaluate

its potential benefit to structural performance.

4.6.1 Thermal Effects in Hybrid Cellular Plates

When laminates of two different materials with significantly different coefficients

of thermal expansion experience a change in temperature the laminate will deflect and

154

 

 

 

bend at a given radius of curvature as shown in Figure 4.25. This was the case for the
biocompostie hybrid cellular plates that were manufactured with carbon fiber face sheets.

To illustrate this concept, consider two cases: (i) two unbonded materials and (ii)
two bonded materials with different coefficients of thermal expansion as shown in Figure
4.25 (where 02 > an). For the first case, at curing temperature To the two materials have
the same length. When the two materials cool to room temperature T, both materials will
shrink different amounts. For the second case, at curing temperature To the two materials
have the same length. When the bonded materials are cooled to room temperature T the
top material wants to shrink more than the bottom material. However, the two materials
are bonded, thus introducing bending. The thermal deformations will cause compression
in the bottom material that wants to expand, but is restrained by the bond with the top
material. Conversely, the thermal deformations will cause tension in the top material that
wants to shrink, but is restrained by the bond with the bottom material. Note that the
temperature induced, or residual, stresses and strains exist in the section with no

application of external load, thus creating a state of prestress.

155

 

 

 

[ Unbonded layers
at curing temperature, To

  

                 
  
   

P.0......OO'QQO'OOOOOQ9.......9'............Q.'.......'...'.0......C
»:oz.:o:o:o:o:o:o:o:o:o:o:o:ozozo:ozozozozozozozozozozo:ozozozozozozoas:«eases:o:ozozozozezozo333333333334 '
a . .9...9.0.0.9.9.9...9.9.9.9.0.93.9.9...0.9.9.....0...9.0.0.0.0.0.... 9.9.08.9...93.9.9.0....a0.9.0.9.0...93...‘3.9.0.89.03...

.0

      

al

 

Unbonded layers
at room temperature, T

 

    
   

 

 
    
  

... ..."'..........."............'....'...."..'..........'....
0afipfififiQQ§§fiQQ§QfifiQ§§fifififi$QQfiQQ§fififififiQfifiQfififig§%fififififi&fi&fififififififi§fifififi

                       

 

Bonded layers
at curing temperature, To

 

 

 

 

 

al
Bonded layers
at room temperature, T
.. 0‘0
I O 0 0 O 0
§@@% ,QQQP
0000 .000
gfififi. ‘g501
%%%% ,dVVV d
000 .0000
ggg. lgfifi;
0000 4%00I

00. .00”!

$90. €0V‘9

00000 ,0000’

~0000 .0000

‘Qfififiwo #999”

%%%%%3 €0009v
‘%%%%%h00 od@b§9?
‘%&%~%§W30 gaafifip0!
‘0900000 ..... ...0 00000000‘
0.0. .0.0.0 0 0 0.0.0.0.0.0.0.0.0.0.0.0 0 0 0.0.0, »
“9&&fiflfifififififififi§ 0'
0.0.0.0330 0

Figure 4.25. Temperature deformations in unbonded and bonded materials with a; > a,

The case described above is similar to the case of the biocomposite hybrid cellular
plates with carbon fiber face sheets. The carbon fiber face sheets have a much lower
coefficient of thermal expansion than the biocomposite material system, which leads to
temperature induced stresses and strains in the section. In the following section,
analytical models that can be used to calculate the temperature deformations are

presented.

156

J: . Limits-m.“

. 1'“

 

 

4.6.1.1 Analytical Models for thermal efiects in Hybrid Cellular Plates

Two analytical methods were used to calculate the curvature and the deflection of
the biocomposite cellular plate with a bottom layer of carbon fiber fabric. The first
method uses discrete equations (Ghali and Neville 1997, Barker and Puckett 1997) to
calculate the axial strain and curvature in the plate due to temperature effects. The
stresses in the plate are then calculated using the axial strain and curvature profile. The
second method is based on a solution using the theory of elasticity as presented by
Timoshenko (1925). The deflection of the plate was then calculated using the curvatures
calculated using both methods and simple geometric relations.

The temperature induced deformations due to nonlinear temperature gradients can
be determined by integrating the gradient effect over the cross-section of the member.
This analysis approach is presented by Ghali and Neville (1997) and is commonly used
for bridge design where nonlinear temperature gradients are encountered. However, if
the temperature gradient is linear and the cross-section is made of distinct materials with
different coefficients of thermal expansion, the integration can be modeled by piece-wise
linear functions (Barker and Puckett 1997).

As shown by Ghali and Neville (1997), the axial strain due to a nonlinear

temperature gradient is given by:
a a
Eo—Xjndy—ZJTdA. (4.6.1)

If the temperature gradient varies at most linearly through the section depth, then the
cross-section can be discretized into layered segments and Equation (4.6.1) can be
simplified to a discrete summation (Barker and Puckett 1997). Consider the segment

shown in Figure 4.26 subjected to a linearly varying temperature gradient. Although the

157

segment shown is rectangular, the segment shape may be arbitrary. The segment elastic

centroidal axis is located at a distance y, from the section neutral axis, and y is the

location of the differential segment area dA. The area of the segment and its second

moment of area are denoted by A,- and 1,, respectively. Also, note that y, = y — 5,. Thus,

the temperature at location y is:

A . AT.
T=T(y)=Ta, +—d—’y, =70, +—d—'(y—y,), (4.6.2)

where Ta,- is the temperature at the segment centroid, AT,- is the temperature difference

from the bottom of the element to the top (AT, =Tb",, —T’ ), and d,- is the depth of the

top

 

 

 

 

 

 

 

 

segment.
I ‘— b —.-1
T'OP

[71/1/[f/f/Iff/ ’

Y r d
Cellular n d a I ’
can I A’ 7r i

The! '

N A

 

 

 

 

 

 

L—b—e ’

Figure 4.26. Example cross-section with discrete element

Substitution of Equation (4.6.2) into Equation (4.6.1) yields:

0 AT, _
£0 =XZI[Tai +E—(y — yi )}A,‘ 9 (4.63)

158

l 2.7.: 326:!» _F,

 

 

where the summation is over all the segments in the cross-section and the integration is
over the domain of the discrete segment. Integration over each term in Equation (4.6.3)

leads to:
0 AT, AT,
80 =ZZ|:T0,IdA., +-d—l-Iyd14,——d'—IdA,]. (4.64)

Recognizing that A, =IdA, and y,A, =J‘ ydA, , and substituting into Equation (4.6.4)

results in:
0 AT. AT.
8 =— T. .+—‘—'.A.———‘-.A. . 4.6.5
. AZ[.A. din. ax] ( )
Noting that the second and third terms sum to zero, Equation (4.6.5) simplifies to
a
£0 = 321T..- A, ]. (4.6.6)

Because the coefficient of thermal expansion may vary for the elements in the hybrid
cellular plate, ais moved into the summation:

e, =%Za,[Ta,A,], (4.6.7)

which is the discrete form for Equation (4.6.1).
The temperature induced curvature is now considered. The curvature, (0, due to

the temperature gradient is (Ghali and Neville 1997):
a a
¢=711bydy=71TydA, (4.6.8)

where I is the second moment of area of the entire cross-section about the elastic

centroidal axis. Substituting Equation (4.6.2) into Equation (4.6.8) yields:

AT.
¢=%Z[TaijyidAi +TJ'0’2 —yiy)dAi]’ (4-6~9)

159

 

 

where again, the summation is over all the segments in the cross-section and the
integration is over the domain of the discrete segment. Evaluating the integration in

Equation (4.6.9) leads to:

a _ AT.- 2
¢=TZ[T..-y.-A.- +7.4), dA -—y.-IydA.-] (4.6.10)

AT. AT.
¢=%Z[T.J.A.-+-;,—'-I."iii/1,]. (4.6.11)

Recalling the parallel axis theorem to relate the cross-section properties in Equation

(4.6.11):

= i, + A. y}, (4.6.12)

and solving for 7, yields:

1, =1, —A.y,. (4.6.13)

Using Equation (4.6.13) with Equation (4.6.11) results in:

Again, because the coefficient of thermal expansion may vary for the elements in the
hybrid cellular plate, ais moved into the summation:
AT. -
¢=%Za,[Ta,y,A, +—d—'1,], (4.6.15)
which is the discrete form of the integral given in Equation (4.6.8).
The analysis of the deflection of a bimetal strip was first developed by
Timoshenko (1925), who presented a simple derivation using the theory of elasticity.

This analysis was derived for bimetal strips, but is equally applicable to bimaterial strips.

160

 

 

The general equation for the curvature radius of a bimaterial strip under uniform

temperature change from To to Tin the absence of any external forces is given by:

1:4,: 6(1+m)2(a2—a,)(T—To)

2 2 , (4.6.16)
R t[3(a + m) + (1+ mn)(m + 1/ mn)]

 

where: a] = coefficient of thermal expansion of the low expansive material
a2 = coefficient of thermal expansion of the high expansive material
12 = E1/E2 , with E1 and E2 the elastic moduli of each material 1
m = t1/tz, with I] and 22 are the thickness of each layer

t = I] + t; the thickness of the strip,

 

and the width of the strip is taken as unity.

Once the curvature of the plate is calculated, using either of the above methods,
the deflection at the center of the beam can be calculated using geometric relations.
Using Pythagoras’ Theorem the following relation can be made from a right triangle
between the end of the plate, the middle of the undeformed plate, and the center of the

curvature of the plate:
(R—t,)2 =(R—d-t,)2+(L/2)2. (4.6.17)
Expanding Equation (4.6.17) and solving for the curvature of the plate we get:

= 8d (4.6.18)

1
R L2 +4d2 +8dt,'

 

Making the assumption that the deflection, d, and the thickness, t, are less than 10% of
the length of the plate the terms 8dt1 and 46'2 may be neglected and the expression

reduces to:

d = —. (4.6.19)

161

 

The axial shrinkage is also assumed to be small and thus neglected in the calculation of

the deflection of the center of the beam.

4.6.1.2 Results

The hybrid cellular plate with green hemp/polyester and a bottom layer of carbon
fabric was analyzed using both analytical models detailed in the previous section. The
elastic modulus of the unidirectional carbon layer was taken to be 77.3 GPa, assuming a
fiber volume fraction of 30%, and the coefficient of thermal expansion was taken as 7.24
x10'6/°C. For the green hemp biocomposite material system, the elastic modulus was
taken as 5.0 GPa, and the coefficient of thermal expansion was taken as 50x1045 /°C. The
assumed values are in agreement with the values measured in the material experimental
program (Chapter 2). The change in temperature was taken as —120°C, which is the
change in temperature from final curing conditions (150°C) to room temperature (30°C).
The mid-span deflection of the plate was measured using calipers accurate to 0.01 mm.
Because the deflection was relatively small compared to the width of the plate (101.6
m) it was not possible to measure the deflection at the center of the plate. Thus, the
deflection was measured at the edges of the plate and averaged. The measured deflection
at the mid-span of the plate was 4.46 mm.

The discrete equations (Equations 4.6.7, 4.6.15, and 4.6.19) were used to calculate
the axial strain, curvature, and mid-span deflection, d, of the hybrid cellular plate as
shown in Table 4.11. The mid-span deflection of the plate was calculated as 3.50 mm,
which is within 22% of the measured deflection (Table 4.11). The assumed properties of

the biocomposite section and the carbon layer may have lead to errors in the calculated

162

 

 

deflection. The measurement method used to determine the mid-span deflection of the

plate may have also lead to errors in the measured deflection.

Table 4.1]. Temperature deflection calculations of the biocomposite cellular plate with bottom
carbon layer (CP1-C) using discrete equations

 

 

 

 

 

SECTION 1 SECTION 2
PROPERTY Bio Carbon’ TOTAL.
E, (GPa) 5.0 77.3 --
E, (ksi) 725 11209 --
a, (We) 50015-05 7.24E-06 --

6. (mm) 15.84 0.42 16.26
A. (mmz) 1039 6583 7621.9
1,... (mm‘) 2.86E+04 9.61 E+OO 1,045+05
y...” (mm) 43.39 0.54 15.5
Ta.(°C) -120 ~120 --
AT. (°C) 0 0 --

 

 

 

*Carbon transformed to properties of the biocomposite

L: 304.8 mm
60: -1.57E-03 mm/mm

¢o= 3.10E-04 1/mm
d: 3.50 mm

dmemm= 4.46 mm

Based on the hypothetical free strains in the section if the layers were unbonded,
and the calculated temperature-induced axial strain and bending curvature of the bonded
section, the strain and stress distributions in the section were calculated (Figure 4.27).
The resulting stress distribution shows the top layer of the biocomposite material is in
tension due to temperature effects, while the carbon layer is in compression.

The stresses due to thermal effects (Figure 4.27) create a state of prestress in the
section that is opposite to the stress state in positive bending to be imposed by the

mechanical loads (3-point bending, see Section 3.3.4). In other words, the bottom of the

163

 

beam, which is to be subjected to tensile stress due to bending, is initially in a
compressed state, while the top of the beam, which is to be subjected to compressive
stresses due to bending, is initially in a state of tension. Thus, the bending stresses must
overcome the prestressed state in the section prior to creating tensile stresses at the
bottom of the section, or compressive stresses at the top of the section.

The strain in the bottom carbon layer due to thermal effects is 0.47e-3 mm/mm in
compression (Figure 4.27), which, based on the measured ultimate tensile strain of 1.67e-
2 mm/mm for the carbon material (see section 2.4.4.4, page 51), is 0.7% of the failure
strain. In order to fail, the carbon layer must deform through the compressive strain of
0476-3 mm/mm to reach zero strain in the layer, and then deform to its ultimate tensile
strain of 1676-2 mm/mm. Thus, the range of tensile strain the carbon layer must go

through before failure is increased by 0.47e-3 mm/mm, or 0.7%.

164

.. “my,

 

1!

 

Biocomposite
CellularPlate

16.213 mm \b <:Carbon

-6.00e-3

 

i
0.42 ml

 

 

 

 

////

 

-0.87e-3

Free Strain
(Hypothetical)

Cross Section

-1.57e-3 4.816-3

 

0230-3

Axial Strain Bending Strains

 

'54“ 22.15

-36.33
Principal Stress

Distribution
(MP8)

 

 

 

-t.57e-3 «1819-3
0236-3
Axial Strain Bending Strains
-6.00e-3 0386-3
-0.70e.3 4436-3
-O.87e-3 -O.47e-3
Free Strain Principal Strain
(Hypothetical) Distribution
-0.276
'7'” 3.213
-5.268
Principal Stress
Distribution

(KSI)

Figure 4.27. Calculated stress and strain distributions due to thermal effects in the biocomposite
cellular plate with bottom carbon layer (CP1-C)

The elasticity solution by Timoshenko (1925) was used to calculate the curvature,

and mid-span deflection, d, of the cellular plate as show in Table 4.12. The calculated

deflection of 3.51 mm is within 21% of the measured deflection of 4.46 m (Table 4.12).

Again, the assumed properties of the biocomposite section and the carbon layer may have

lead to errors in the calculated deflection.

165

 

 

Table 4.12. Temperature deflection calculations of the biocomposite cellular plate with bottom
carbon layer (CP1-C) using the elasticity solution

 

 

 

 

 

 

PROPERTY PROPERTY
01, (1/°C) 7.24E-06 t1 (mm) 0.42
oz (1/°C) 5.00E-05 t2 (mm) 15.84
E. (GPa) 77.3 t(mm) 16.26
E, (ksi) 11209 11 15.46
E; (GPa) 5.0 rn 0.0265
iflksi) 725

 

 

L = 304.8 mm

(to: 3.02E-04 1/mm
d: 3.51 mm

dmeasurad = 4.46 mm

Both analytical models used to calculate the temperature deformations in the
hybrid biocomposite cellular plate gave similar results and agreed well with the measured
deflection of the beam. The curvatures calculated using both methods also agree well
with each other (see Table 4.11 and Table 4.12). The results of the analytical analyses
may be improved with more accurate values for the elastic modulus and coefficients of
thermal expansion for the biocomposite and carbon fiber materials.

Based on the agreement of results for both methods for calculating deflections due
to thermal effects, the discrete equations were used to analyze the hybrid cellular plate
with raw hemp/polyester and top and bottom carbon layers. The measured dimensions of
the hybrid cellular plate were used, and the properties for the biocomposite material
system and carbon layers in the previous analyses were taken equal to the previous
analysis.

The properties used for the analysis are shown in Table 4.13, and the resulting

stress and strain distributions are shown in Figure 4.28. Notice that because the section is

166

 

1” n' ‘1. 'rsu I.‘ ' 110 .1:-.
I

 

symmetric there are no bending deformations in the section (Figure 4.28). However, the
top layer (carbon layer) is subjected to compression, which is undesirable for a section

that is to be subjected to positive bending.

Table 4.13. Temperature deflection calculations of the biocomposite cellular plate with top and

bottom carbon layers (CP2-C) using the discrete equations

 

 

 

 

 

 

 

SECTION 1 SECTION 2 SECTION 3
PROPERTY Carbon' Bio Carbon' TOTAL'
E. (GPa) 77.3 5.0 77.3 --

Ei (ksi) 1 1209 725 11209 --
or. (1/°C) 7.20E-06 2.50E-05 7.20E-06 --

d, (mm) 0.42 17.46 0.42 18.30
A. (mmz) 658.3 1039.0 658.3 1697.3
1,... (mm‘) 9.61E+00 4.12E+04 9.61E+00 9.39E+04
y...” (mm) -8.94 0.00 8.94 15.50
Ta. (°C) -120 -120 -120 --
A11(0C) 0 0 0 .-

 

*Carbon transformed to properties of the biocomposite

167

 

Biocomposite

 

 

 

 

 

 

 

 

 

 

   

 

jOAme Cellular Plate 41879-3 '2.516-3
’ 1, ////({[ 1
' <> 8 Carbon
, ////‘7 .
0 42mm_i -O.87e-3 0.0
Cross Section Free Strain Axial Strain Bending Strains
(Hypothetical)
-2.516-3 -0.87e-3 -1 649-3
+ _ -3.00e-3 0,499.3
0-0 o0.87e-3 -1.64e-3
Axial Strain Bending Strains Free Strain Principal Strain
(HypOthetical) Distribution'
426.77 -18.37
-1.87 0.27
-126.77 -18.37
Principal Strata P0005?” Stress
Distribution 018mm
(MPa) (KSI)

Figure 4.28. Calculated stress and strain distributions due to thermal effects in the biocomposite
cellular plate with top and bottom carbon layers (CP2-C)

4. 6.2 Temperature Prestress Efiects

The temperature prestress effects on the two hybrid cellular plates with carbon

fiber face sheets tested in the structural experimental program (Chapter 3) are illustrated

in Figure 4.29 and Figure 4.30. The load-displacement response of the two cellular plates

was used to calculate the force-strain response in the bottom carbon face sheet. The

168

force-strain response was then corrected to account for the thermal strains in the section,
which created a state of prestress prior to testing. For both hybrid plates, the thermal
strains in the bottom face sheet were compressive, which is opposite to the tensile strains
from the applied loading in the flexural test. Thus, the thermal strains were subtracted
from the measured strains. The resulting force-strain response for the hybrid plate with
bottom carbon only is shown in Figure 4.29, while the response for the hybrid plate with
top and bottom carbon is shown in Figure 4.30. The discontinuity in the hybrid plate
with top and bottom carbon (Figure 4.30) is due to the change in the section properties
(namely the moment of inertia) that occurs when the top face sheet buckled (Section
3.5.1). The section properties change because the buckling of the top carbon layer causes
the neutral axis to shift towards the bottom carbon layer.

The corrected force-strain results provided in Figure 4.29 and Figure 4.30 show
the corrected failure strains in the carbon face sheets are in agreement with the failure
stains measured in the material experimental program (see Table 2.9). Notice that for
both hybrid plates the measured failure strain in the carbon face sheet is larger than the
actual failure strain of the material. This apparent increase in the failure strain is due to
the temperature prestressing effect, which allows the material to go through a larger range
of strain before failure. This increase in capacity due to temperature prestress can thus be
taken advantage in the design of hybrid structures.

The beneficial effects of prestressing were small for the tested laboratory-scale
hybrid cellular beams and plates. This is due to the small dimensions of the cross-
section. The effects on full—scale components were found to be more attractive as

described in the next section.

169

12000

 

 

    
 
  
 
   

 

 

 

 

 

 

 

 

 

 

; —M°““'°d 5“" s, =0.0182 trim/mm
4 -—-Corrected Strain
10000 :5
. Plat; Data:
] CP1-C
8000 f b=102mm, h=13mm, L=267mm
2 , Tat:
v ] 3-Pt Bending
8 6000 w Location:
3 1 Mid-Span
u. : Straln - Bottom Fiber
4000 :r
1 Material Data:
j_ Green Hemp - 10 wt.%
200° « Bottom Carbon Layer
, Unsaturated Polyester
o........4:-.4.1....4....
-0.005 0.000 0.005 0.010 0.015 0.020

Strain (mm/mm)

Figure 4.29 Force-strain response of the biocomposite hybrid cellular plate with a bottom carbon
face sheet, with correction for temperature prestress effects

 

 

 

 

 

 

 

 

 

 

 

 

12000
j —Meaeured Strain 8'30.0166 mmlmm
< —Corrected Strain
10000 ~-
2 Plate Data:
j CP2-C
8000 .. b=102mm, h=13mm, L=267mm
5 3 L53;
v : 3-Pt Bending
8 6000 1L Location:
3 i Mid-Span
u. : Strain-Bottom Fiber
4000 f
: Matgial gm:
1, Raw Hemp ~10 wt.%
2°00 , T&B Carbon Layers
, Unsaturated Poiyeehr
o..fcr....:....i....1f....
-0.005 0.000 0.005 0.010 0.015 0.020

Strain (mm/mm)

Figure 4.30. Force-strain response of the biocomposite hybrid cellular plate with top and bottom
carbon face sheets, with correction for temperature prestress effects

170

4. 6.3 Prestressing Effects in F all-Scale Panels

The temperature prestressing effects on a full-scale cellular plate with integrated
carbon fiber fabric were determined and compared with the effects on the small
laboratory-scale cellular plate. The full-size cellular plate was taken as a 152.4 mm deep
by 1219.2 mm wide cellular plate with two layers of carbon fabric on the bottom, which
have a total thickness of 0.84 mm. The relative density of the full-scale plate was taken
to be the same as the laboratory-scale plate, p70, = 0.588. The same material properties
were used for both plates. Calculations and results for the full-scale cellular plate are
shown in Figure 4.31.

The results show that the curvature for the full-scale hybrid cellular plate, 7.9e-8
l/mm (Figure 4.31), is very small compared to the curvature in the laboratory-scale plate,
3.16-4 llmm (Figure 4.27). Thus, thermal effects in the full-scale cellular plate may not
cause the large curvatures and deflections seen in the laboratory-scale plates. Despite the
small curvature in the full-scale hybrid cellular plate, the axial strain, 5.08e-3 min/mm
(Figure 4.31), is large compared to the lab scale plate, 1.57e-3 mm/mm (Figure 4.27).
Because of the large axial strain in the full-scale plate, the stresses are much higher. In
the full-scale plate the residual strain in the carbon layer is 41% of its failure strain, while
in the laboratory-scale plate the strain in the carbon layer was only 0.7% of its failure
strain. Thus, the prestressing thermal effects in the full-scale plate considerably increase
the range of strain the carbon layer must go through before failure, and consequently

increase the capacity of the section.

171

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

   

 

SECTION 1 SECTION 2 Biocomposite
PROPERTY Bio Carbon' TOTAL' Xce'm'a'P'a'e ‘6-00‘3‘3 5.0353 ~0-008e-3
Er (GPa) 5.0 77.3 -- Kb
2
E. (ksi) 725 1 1209 - ‘52 5”“ ,Caroon
a.- (1/°C) 5.00E-05 7.24E-06 .. i “(K
<1. (mm) 151.61 0.84 152.45 cum—l 0876-3 Goose-3
A. (mmz) 71977 3 15799 3 87776 1 Cross Section Free Strain Axial Strain Bending Strains
' ' ' (Hypothetical)
lo.” (mm‘) 2.62E+08 9.24E+02 3.45E+08
Yb.” (mm) _17.19 6250 8953 '5088'3 '00089’3 '6009-3 0.9318'3
Tar (°C) -120 -120 -
ATI (0 C) 0 O -- + - =
'Carbon transformed to properties of the biocomposite '4'209'3 09296-3
00066-3 -0.87e-3 4206-3

Axial Strain Bending Free Strain Principal Strain

L = 1219.2 mm Strains (Hypothetical) Distribution

3.00 0.68
5,: -0.00508 mm/mm
(4,: 7.87E—08 1/mm 209.31 2.99 '47-” 0.67

 

-209.31 47.10
d: 0.01778 mm Principal Stress Principal Stress
Distribution Distribution
(MPa) (KSI)

 

Figure 4.31. Strain and stress calculations and distributions due to temperature effects in a full-scale
biocomposite cellular plate

4.7 Performance and Feasibility Studies

The results from the flexural tests of the biocomposite cellular beams (Chapter 3)
were compared with beams of equal size made of conventional construction materials
(Figure 4.32). The conventional materials used for the comparison were wood (Douglas
Fur), and reinforced concrete. The reinforced concrete beam (3% steel reinforcement
ratio) was analyzed both cracked and uncracked. However, reinforced concrete beams
typically crack under service loads, reducing the stiffness of the component.

The results in Figure 4.32 show that the bio-beams compared very favorably with

the beams from conventional construction materials. The specific stiffness of the

172

hierarchical bio-beam outperforms both the cracked and uncracked concrete beams.
Their performance, however, falls short from that of the wood beam. The specific
stiffness of the wood beam is very large because of the low density, and the moderate
elastic modulus of Douglas Fur. The cellular bio-beam with a periodic cell arrangement
shows a lower specific stiffness than the hierarchical bio-beam but still outperforms the

cracked concrete member.

 

1.4

4

O O .5 d

I I I

or on o in
AAJALLIII LLALLAAJALJ

T I I

0
&

to '
lAlJLllll
—I

l
1

Specific Stiffness (kN-m’IN)

 

 

 

P
o

 

Concrete Concrete Wood Cellqu Hierarchical
(uncracked) (cracked) (Douglas Fur) BioBeam BioBeam

Figure 4.32. Performance biocomposite cellular beams with beams of conventional construction
materials
The results from the experimental studies on the biocomposite cellular plates were
extrapolated to evaluate the performance of full-scale structural components and compare
them with conventional flooring systems used in the construction industry. The
allowable pressure versus the simply supported span length for panel systems is shown in
Figure 4.33. The allowable pressure is the loading the component can withstand while

maintaining an industry-standard allowable deflection (U180). For this application, the

173

deflection is the critical design criteria and the strength requirements are met for all span

 

 

 

 

 

 

 

 

 

 

 

lengths shown.
15-0 j t=76mm, b=0.61m e t=203mm, b=1.29m
1 + Duraehield -e— PCIPs Hc Slab
'3 _-,1 + Compoeollte .. (. -e— PGIPS Solid Slab
g 125 . + Cellular Bio-Panel ' + OSB/EPS ISP
T; I 9 0 + Cellular Bio-Panel
10 o —
g I \1 “
g : \\‘ \
g 7.5 ‘- \
a- : ‘ ‘ e
8 5.0 -: , e
5 : ‘ .- r e e
< 2.5 "- \‘ L‘ ‘ “ j‘- ‘.
- \ ~ .1 ‘J
0.0 — - = =
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 0.5 7.0 7.5 8.0 8.5 0.0
Simply Supported Span (m)
Figure 4.33. Performance of full scale biocomposite cellular panels versus conventional flooring
systems

Figure 4.33 compares panels grouped into two categories. The first set of curves
is for a panel size of 76 mm by 0.61 m, which is a size typically used in commercial
building construction. For this group, the Bio-Panel is compared with commercially
available fiberglass sandwich panel systems (Durashield and Composolite [Strongwell
2003], Table 4.14). The second set of curves is for a panel size of 203 mm by 1.29 m,
which is a size typically used for conventional materials in commercial construction. For
this group, the bio-panel is compared with a precast prestressed hollow core concrete slab
(PC/PS HC slab, Table 4.15), a precast prestressed solid concrete slab (PC/PS solid slab,
Table 4.15), and an oriented wood strand board ISP (insulated structural panel [Insulspan
2003], Table 4.16). The PC/PS HC slab, and PC/PS solid slab are standard building

components found in the PCI Design Handbook (1999). A green hemp/polyester cellular

174

biocomposite plate (cellular bio-panel) with a fiber volume fraction of 30% and a relative

density of 0.56 was used for the comparison in both panel size groups.

Table 4.14. Pultruded fiberglass panel data

 

Flexural Flexural

Resin Depth Width Strength Modulus
Panel Type (mm) (mm) (MPa) (G Pa)
Composolite Vinyl Ester 76 0.61 168.9 6.10

Durashield Vinyl Ester 76 0.61 103.4 2.17

Table 4.15. Precast prestressed concrete panel data

 

 

Compressive Strand
Depth Width Strength Number Size
Panel (mrm (mm) (M Pa) of Strands (mm)
Precast Prestress Hollow Core Slab 203 1.29 34.5 6 9.5
Precast Prestressed Solid Slab 203 1.29 34.5 6 9.5

 

Table 4.16. Oriented strand board/expanded polystyrene core - insulated structural panel data

 

OSB OSB EPS Core
Depth Width Thickness Stiffness Depth

 

Panel 4mm) (mm) (mm) (G Pa) (mm)
OSB/Expanded Polystyrene core -
Insulated Structural Panel 203 1.29 11.1 6.5 180.8

 

The results in Figure 4.33 show that the cellular bio-panels compete well with all
of the conventional panels for both categories, and can even outperform some
conventional flooring systems. In the first category, the bio-panel competes well with the
Composolite panel and outperforms the Durashield panel. For example, for a span of 2.0
m the Composolite panel has an allowable pressure of 14 kPa and the Durashield panel

has an allowable pressure of 9 kPa, while the bio-panel has an allowable pressure of 12.5

175

kPa. In the second category, the bio-panel substantially outperforms the OSP/EPS ISP
panel for all span lengths, and outperforms the PCIPS solid slab for spans above 6.75 m.
The bio-panel compares well with the PCIPS HC slab for large spans (>8 m), but for
shorter spans the PCIPS HC slab outperforms the bio-panel. It should be noted that the
above comparison is for an all-biocomposite panel system, thus the performance of
hybrid and hierarchical panels is expected to be even better.

The results from the experimental studies on biocomposite cellular plates and the
analysis procedures presented in Section 4.3 were used to extrapolate the performance of
full-scale structural components for different applications. The full-scale performance
was used to select dimensions of biocomposite cellular panels for the following
applications: (i) highway bridge decks, (ii) commercial building floor panels, and (iii)
residential flooring systems. A three-lane bridge with a girder spacing of 2.13 m (7-ft)
was used for the highway bridge application. The American Association of State
Highway and Transportation Official-Load and Resistance Factor Design (AASHTO-
LRFD 1998) specifications were used for the bridge loading and allowable deflection
criteria. A one-way slab with a span of 6.10 m (20-ft) was used for the commercial
flooring system. The International Building Code (IBC 2000) specifications were used
for the slab loading and the allowable deflection criteria. For the residential flooring
system, a one-way floor system was used with a span of 0.61 m (2-ft) was considered.
The IBC (2000) specification was again used for the loading and allowable deflection
criteria.

For the cellular panels, a green hemp/polyester with a fiber volume fraction of

30% and a relative density of 0.56 was taken as the material system. A hybrid system

176

with top and bottom layers of carbon was also considered for the bridge and commercial
flooring applications. For the face sheets, a carbon/polyester material system with a fiber
volume fraction of 50% was used as the material system.

The dimensions of the biocomposite cellular panel are shown in Figure 4.34,
where the top and bottom flanges are solid material, and the core is assumed to have a
periodic cellular architecture. The selection of dimensions was based on the allowable
deflection limits for each application. Strength requirements were also considered and
found to be much larger that the demand, however no resistance factors exist for design
biocomposite materials. Thus, strength requirements for the bio-panels could not be
properly evaluated. For the analysis and dimensioning of the bio-panels a unit strip

method was used.

{Flange T Carbon Face Sheet f tF’aWO

)2.“ l v...

j ' Lamina-oases“ j

(a) Bio-panel (b) Hybrid bio-panel

 

 

 

 

 

 

 

 

 

 

 

 

Figure 4.34. Schematic of the dimensions selected for the bio-panel structural components

The dimensions of the bio-panels for structural applications are provided in Table
4.17 (SI units) and Table 4.18 (US units). For the highway bridge deck, the depth of the
bio-panel may be too thick for practical applications. However, the depth of the hybrid

bio-panel is more reasonable, and falls within the thickness of typical highway bridge

177

decks, which range from 180-230 mm (7-9 inches). In addition to providing increased
stiffness and strength to the hybrid bio-pane], the carbon face sheets also act as a barrier
to the biocomposite material providing protection from weather effects. For pedestrian
bridge applications, where the loads are not as high, the use of an all bio-panel may be a
viable solution. However, for highway bridge applications, hybrid bio-panels, or the use
of all-bio-panels on bridge systems with closer girder spacings is better suited.

For commercial building slabs, the depth of both the bio-panels and the hybrid
bio-panels are slightly larger than conventional reinforced concrete slabs, which have a
typical depth of 152 mm (6 inches). This suggests that hybrid bio-panel may be needed
for building slab applications. However, if the height between floors is not a major issue
in the building construction, as is the case in low rise building construction, it may be
possible to use all bio-panels for flooring systems.

For residential flooring systems the dimensions of bio-panels are practical and
compare well with the dimensions of commercial panels (Figure 4.33). Clearly, from a

structural standpoint, a hybrid bio-panel is not need for residential applications.

Table 4.17. Dimensions of bio-panels for structural applications (SI units)

 

 

Section dimensions: Section Dimensions:
System Bio-Panel Hybrid Bio-Panel
HIGHWAY BRIDGE DECK:
d = 305 mm d = 229 mm
tam. = 13 mm tum, = 25 mm
COMMERICAL BUILDING SLAB:
d: 229 mm d: 178 mm
tum, = 13 mm tame = 13 mm
RESIDENTAL FLOOR:
(1 = 51 mm d = Not mm

rm = 6 mm tm = Needed mm

178

Table 4.18. Dimensions of bio-panels for structural applications (US units)

 

 

Section dimensions: Section Dimensions:
System Bio-Panel Hybrid Bio-Panel
HIGHWAY BRIDGE DECK:
d = 12 in d = 9 in
tflange = 0.5 in tflange = 1.0 in
COMMERICAL BUILDING SLAB:
d = 9 in d = 7 in
tum = 0.5 in tflangg = 0.5 in
RES/DENTAL FLOOR:
(1 = 2 in d = Not in
hm = 0.25 in tm = Needed in

 

 

 

179

CHAPTER 5 RESULTS, FEASIBILITY AND RECOMMENDATIONS

5.1 Overview

The findings from the material and structural experimental programs, and the
analytical studies are summarized and used to assess the feasibility of using
biocomposites for load-bearing structural components. In addition to the findings in this
study, issues of processing and implementation in design procedures are also discussed in
assessing the feasibility of load-bearing biocomposite components. Based on the
feasibility assessment and the findings in this study, future work is recommended for the

use and development of biocomposite material systems and structural forms.

5.2 Experimental Studies Summary

The results from the material and structural experimental studies are summarized
in the following sections. The material experimental study included testing of all-
biocomposite material systems and hybrid material systems with natural fiber core and
fiber mat face sheets for tensile properties, impact strength, coefficient of thermal
expansion and moisture absorption. The structural experimental study included testing of
laboratory-scale biocomposite cellular beams and plates with varying cellular topologies
and hybrid material systems. Focus is given to results that give insight to the use of

biocomposite materials for load-bearing structural components.

5.2.1 Material Systems

The material experimental studies verified that the tensile properties of

biocomposite materials could compete with E-glass fiber reinforced composites. In

180

addition, the use of hybrid material systems with synthetic or aligned natural fiber face
sheets and a chopped natural fiber core dramatically improved the mechanical properties
over the all-natural fiber and all-E-glass composites. The hybrid material systems with
E-glass face sheets also improved the moisture absorption response, which was
comparable to the all-E-glass material system tested.

The experimental studies also showed that biocomposite material systems have a
much higher coefficient of thermal expansion than conventional structural materials
(approximately 3 times). This difference in thermal expansion must be considered if
biocomposites are to be used with conventional structural materials. The moisture
absorption of the tested biocomposites was also much higher than of the tested E—glass
composite (approximately 4 times). The absorption of moisture into the biocomposite
material may lead to swelling and durability issues. These properties were not studied in
this research but are important characteristics for structural applications.

The material experimental studies also found the biocomposite material systems
to be very brittle. This is due to the short natural fibers and the brittleness of the
polyester resin used. The brittle behavior of the biocomposite material systems is
undesirable for structural applications. Thus, improvements on the ductility of

biocomposites should be investigated.

5.2.2 Structural F arms

The results from the structural experimental study verfied that the use of
hierarchical cellular structures can improve the performance of biocomposite beams and
plates. This suggests that hierarchical cellular structures should be studied further to

develop efficient structural forms for biocomposite materials. The structural

181

 

experimental study also shows that using jute, glass strand mat, and carbon face sheets
can improve the stiffness and strength of biocomposite structures. These improvements
can widen the range of structural applications in which biocomposite cellular structures

can be used.

5.3 Analytical Studies Summary

The findings from the analytical studies are summarized for material systems and
structural forms in the following sections. The analytical studies on material systems
included performing a survey of available models to predict the stiffness and strength of
biocomposite material systems and evaluating their performance with measured results.
The studies on structural forms included the use of models to analyze the cellular beams
and plates of the structural experimental study. Performance factors that measure the
efficiency of structures in bending for minimum weight were also considered to study the

effect of cellular microstructure on the performance of structural components.

5.3.1 Material Systems

The stiffness constants of biocomposite material systems can be predicted using
the Halpin-Pagano laminate method, which approximates a randomly oriented fiber
composite material as a quasi-isotropic laminate. The unidirectional properties of the
composite material should be modeled using micromechanics models such as the Halpin-
Tsai equations, which account for the reinforcement fiber aspect ratio. Prediction of the
elastic modulus using this combination of models agreed well with the material

experimental results.

182

The strength of biocomposite material systems can be predicted using the
geometric averaging method, where the properties of a randomly oriented fiber composite
are modeled by averaging the strength of an off-axis lamina over all fiber orientations.
The unidirectional strength of the composite material can be modeled using Piggott’s
model, which accounts for the fiber aspect ratio. Strength predictions using these models
overestimated the strength of the biocomposite material systems. The overestimation
may be due to the assumed value used for the interfacial shear strength of the fiber and
matrix. The interfacial shear strength should be measured for improved performance of

the prediction models.

5.3.2 Structural F arms

Cellular, hierarchical, and hybrid cellular structures can be modeled using the
material prediction models previously described in combination with fundamental
mechanics concepts. The flexural stiffness and strength predicted using these tools
agreed well with the experimental results.

A more robust analysis procedure was proposed for analyzing hierarchical and
hybrid cellular structures by discretizing the structure into cellular lamina and using
lamination theory to analyze the discretized structure. This analysis procedure can be
used to analyze hierarchical cellular structures, hybrid structures, and thermal effects.
However, the proposed procedure could not be verified with the experimental studies
since the modeling is limited to hierarchical cellular structures with four or more layers of
distinct cellular material.

Thermal effects in structures with hybrid material systems can create a prestressed

state in the section. When a hybrid structural member is properly designed, thermal-

183

induced strains can increase the section capacity. For example, it was found that using
only a bottom layer of carbon resulted in pre-compression of the bottom layer of the
hybrid cellular plate and pre-tension in the top layer, while using top and bottom carbon
layers resulted in pre-compression of both the top and bottom layer of the hybrid cellular

plate. Pre-compression of the top layer is, however, undesirable for positive bending.

5.4 Feasibility Assessment

The results from the reported studies have shown that hybrid cellular bio
composite beams and plates not only have the potential to serve as primary load bearing
structural components, but that they can compete with conventional structural materials.
In Civil construction, biocomposite structural components can be used as flooring systems
and pre-fabricated components in residential and commercial construction, deck systems
in highway bridges, and complete superstructures of small pedestrian bridges.

A main obstacle facing the use of biocomposite materials (and advanced
composite materials in general) in civil structures is the lack of design procedures. One
way in which designers account for the variability and uncertainty in material properties
is through safety factors. In the load and resistance factor design (LRFD) methods
(AASTHO LRFD 1998), strength reduction and load factors are used to account for the
variability in the capacity of the structural material and the load demands, respectively.
Karbhari and Seible (2000) have proposed strength reduction factors for composite
materials that account for variations due to the material system, processing method, level
of cure, performance level in processing, and degradation in material properties over time

due to environmental effects. Thus, variations in all of these design characteristics must

184

 

be studied and understood in order to develop appropriate strength reductions factors for
design.

Another problem facing biocomposite materials for use in civil applications is the
lack standard material systems and structural components that can be easily selected by
designers. At this stage in the development of biocomposite materials and structural
components, it is not feasible to develop standard materials and sections. However, once
these materials have been developed, the development of standard sections and materials
will facilitate and promote the use of biocomposite components.

Before biocomposite structural components can be implemented in the
construction industry, automated manufacturing methods must be developed. In the
following section various manufacturing processes that have been used for natural fiber

composites are discussed (Brouwer 2000).

5.4.1 Automated Manufacturing

In general, the processing techniques for biocomposites are similar to those for
glass fiber composites. Several processes are available such as compression molding,
sheet—molding compound, and resin transfer molding (Brouwer 2000, Drzal et al. 2002).
However, vacuum assisted resin transfer molding VARTM has be shown to be an
effective process for manufacturing cellular structures using both glass fibers, and natural
fibers (Shenton et al. 2002, Stoll and Banerjee 2001, Berenberg 2003).

VARTM is a Clean, closed mold process that can be used to manufacture
biocomposites (Brouwer 2000). The process involves placing the dry fibers in a mold,
closing the mold with a thin bagging film, and injecting resin. The resin is forced

through the sample using a vacuum pump. Thus, the VARTM process allows

185

manufacturing of very large parts with high mechanical properties and complex shapes,
which is ideal for manufacturing large structural components. A common problem with
natural fibers is the springy character of the fibers. One solution is to pre-form the fibers
into pressed mats for a more compact and manageable shape.

VARTM manufacturing was studied experimentally by manufacturing and testing
cellular plates similar to those used in the structural experimental program. The results of

the manufacturing study are provided in the following sections.

5.4.2 VARTM Cellular plates

The automated manufacturing of cellular plates using VARTM method was
studied experimentally with the aim of showing the viability of using this method for the
manufacturing of biocomposite cellular components. The study included manufacturing
and testing of laboratory-scale cellular plates with varying material systems. A diagram
of the overall VARTM manufacturing process used in the experimental study is shown in
Figure 5.1. Selective material systems from the material experimental program (Chapter
2) and the cellular plate molds from the structural experimental program (Chapter 3) were
used. The workability of the natural fiber material systems was improved for VARTM
manufacturing by developing novel hybrid bio-fiber pre-form fabrics, which consisted of
a chopped bio-fiber core with synthetic or natural fabric face sheets. The hybrid material

system was bound together by compression.

186

 

 

 

Resin being drawn through and

    
 

 

 

 

 

Vacuum
Sealant Tape across reinforcement /— Bag
1 ® a}
Fiber Reinforcement
Pressure Gauge W
Vacuum Pump Resin Trap

 

 

 
 

 

 

 

Figure 5.1. Schematic of the VARTM manufacturing process

It should be noted that the scope of the experimental study was aimed towards
showing the feasibility of this manufacturing method for automated manufacturing of
cellular plates. Thus, multiple tests of equal samples were not considered and

optimization of the manufacturing procedure was not explored.

5.4.2.1 Materials and Constituents

The VARTM cellular plates were manufactured with hybrid fabric pre-forms
consisting of a chopped raw hemp fiber core and face sheets of either woven jute,
randomly orientated glass, or unidirectional carbon fabrics. Two types of pre-forms were
used for the VARTM program: (i) raw hemp core with jute face sheets (Figure 5.2), and
(ii) raw hemp fiber core with hybrid chopped glass strain mat and unidirectional carbon
face sheets (Figure 5.3). The pre-forrn with a raw hemp core and glass/carbon face sheets
was designed such that when placed in the cellular plate mold the carbon layers would
only be on the top and bottom of the plate (Figure 5.3). A polyester resin system similar
to that of the experimental program in Chapter 2 was used, with the exception that 10%

styrene monomer was added to reduce the viscosity for infusion. Thus, the fibers,

187

fabrics, and resin system used for the VARTM cellular plates were essentially the same

as those used in the material experimental study described in Chapter 2.

 

Figure 5.3. Hybrid pre-form fabric: raw hemp core with glass face sheets and top and bottom carbon

188

The springy nature of the raw hemp fibers (and natural fibers in general) made the
hybrid pre-form fabric thick (Figure 5.4a) and difficult to handle and fold when placing
the fabric in the cellular plate mold. Thus, to improve workability, the pre-form fabrics
were compressed prior to manufacturing under 550 kPa of pressure at 100°C for 5
minutes (Figure 5.4b). The pre-compression of the pro-form fabric decreased the
thickness (Figure 5.46), bound the materials together, and made handing, folding, and

placement of the fabric more convenient. All the hybrid pre-form fabrics used for the

VARTM experimental program were pre-compressed, as previously described.

 

(a) Uncompressed dry (b) Dry hybrid pre-form (c) Thickness comparison of dry hybrid
hybrid pro-form, with hemp compressed at 100°C under pre-forms
core and jute fabric face 80 psi for 5 minutes
sheets

Figure 5.4. Pre-compression of the hybrid fabric pre-forms

5.4.2.2 Manufacturing of VARTM Cellular Plates

The VARTM manufacturing process of the cellular plates is divided into three
main steps: (i) placement of the vacuum bagging materials and dry pre-form fabric, (ii)
vacuum bagging of the setup, and (iii) resin infusion. Each of these steps is detailed in
the following paragraphs with descriptions and photographic summaries.

The VARTM setup was prepared on a movable cart with a vacuum pump on the

bottom of the cart, and the mold and sample on the top surface. The vacuum bagging

189

materials were placed in the same cellular plate mold that was used in the structural
experimental program (Chapter 3). First, a non porous Teflon release ply was placed in
the bottom of the mold to prevent the sample from sticking to the mold after curing
(Figure 5.5a). A breather ply was then placed on top of the Teflon ply to absorb excess
resin for the sample (Figure 5.5b). Resin transfer media (high density polyethylene,
green house shade cloth) was placed on top of the breather ply to promote uniform flow
of the resin over the sample (Figure 5.5c). A porous bleeder ply (polyester cloth) was
placed on top of the resin transfer media to achieve a smooth surface after curing and
absorb excess resin from the sample (Figure 5.5d). A layer of porous peel ply was then
placed on top of the bleeder ply to allow excess resin to be squeezed out of the sample
(Figure 5.5). The vacuum bagging materials were placed so that they can be wrapped
around the entire samples once the fibers are in place. With the vacuum bagging

materials in place, the pre-form bio-fabric was then placed in the mold (Figure 5.50

190

 

(e) Porous release ply (1) Dry pro-form fabric

Figure 5.5. Preparation vacuum bagging materials

Two methods were used for placing the pre-form fabric in the cellular plate mold.
The first method used three discontinuous fabrics to comprise the top and bottom face
sheets, and the core cells were achieved by a sinusoidal layout of the fiber fabric around
the mold rods (Figure 5.6a). The second method used one continuous fabric to comprise
the entire unit continuing the top and bottom face sheets into the sinusoidal core layout

(Figure 5.6b). The first procedure for placing the pre-form fabric lead to open side walls

191

 

 

 

and general difficulty in placing the fabrics, especially at the ends of the sinusoidal core.
The second method was developed to improve the previous procedure for placing the pre-
form fabric. The typical fabric placement process is summarized in the photograph
sequence shown in Figure 5.7. A continuous pre-form fabric was used to first create the
bottom face sheet (Figure 5.73). The fabric was then wrapped around the first rod in the
plate mold to create the side wall (Figure 5.7b), and wrapped through the remaining rods
to create the sinusoidal core (Figure 5.7c-d). The fabric was then wrapped over the final
rod to create the side wall and top face sheet (Figure 5.7e). Both procedures used the

same vacuum bagging material layout.

[III/IIIIIIIII/II/I/IIIIIII/II/I/I/IIIII/III/IIII/[III]III/III/I/I/I/I/III/II/Illllllllllla

 

(b) Wrapping procedure using one continuous pro-form fabric

Figure 5.6. Placement of pro-form bio-fabric in the cellular plate mold

192

 

    
    

    

(b) Wrapping of fabric core

(c) Wrapping of fabric core ((1) Wrapping of fabric core

(6) Dry fiber sample: top layer (1) Vacuum bagging materials wrapped over the dry
fabric pro-form

(a) Dry fiber sample: bottom layer

 

 

Figure 5.7. Preparation of dry pre-form fabric and vacuum bagging materials

Once the dry fabric pre-form and vacuum bagging materials were set in the
cellular plate mold, a strip of sealant tape was placed around the edge of the sample, over
the faceplate, and under the rods (Figure 5.8a). Another strip of sealant tape was placed
between the faceplate and the body of the mold to create an airtight seal between them.
Breather cloth was placed over the ends of the rods to prevent the inner steel rods from

puncturing the vacuum bag when under pressure (Figure 5.8a). The resin ports (Figure

193

5.8b) and vacuum ports (Figure 5.8c) were then connected to the sealant tape by splicing
the tape cover and attaching the tubes with additional sealant tape. The resin and vacuum
ports were covered with additional resin transfer media and breather cloth to improve
resin transfer and prevent the vacuum bag from blocking the ports (Figure 5.8a). With
the vacuum bagging materials and ports in place, the sealant tape cover was removed and
the vacuum bag was attached to the mold (Figure 5.8a). The Polyester resin system was
added to the resin reservoir (Figure 5.8d) and the resin trap was connected to a pressure
gauge, which was connected to a cryogenic trap (Figure 5.8c). A valve was used between
the resin reservoir and the sample to control the flow of the resin during the resin infusion
process (Figure 5.8d). The cryogenic trap was filled with dry ice and acetone to
condense and trap any vapors (styrene) from traveling into the pump. The typical
vacuum bagging procedure is summarized in the photograph sequence shown in Figure

5.8.

194

   

   
    

(a) Vacuum bagging

7M

 

.a ., b-s -

(e) Resin trap and cryogenic trap (0 Completed VARTM setup

Figure 5.8. Sample with vacuum bag, resin ports, and vacuum ports

The resin infusion process began by compressing the sample under vacuum
pressure (Figure 5.9a) before resin was injected into the system. Once the sample was
under vacuum pressure, the valve between the resin reservoir and the sample was opened
to allow the resin to be pulled into the mold (Figure 5%). The initial resin flow front

followed a fan-like pattern from the resin ports towards the vacuum ports (Figure 5.9b-f).

195

Once the sample was completely impregnated with resin, the valve between the
resin reservoir and the sample was closed so that no additional resin would be pulled into
the system and any excess resin could be removed. The system was kept under vacuum
pressure for approximately one hour to allow the resin to gel and avoid air voids that

could be created from the “springy” short natural fibers.

 

(e) t = 5 minutes (i) t = 12 minutes

Figure 5.9. Resin flow front movement (total time to complete impregnation = 12 minutes)

196

Once the sample had gelled under vacuum pressure, the pressure pump was
turned off and the ports were detached from the mold. The sample was then even cured
with the mold and vacuum bagging materials. The sample was cured for 6 hours at
100°C. A lower curing temperature was used to avoid melting the resin transfer media,
which is made of high density polyethylene. Steel plates were placed on the sample
during curing to improve the density and quality of the sample.

To evaluate if the cellular plates made by VARTM were properly wet out through
their entire depth, a cellular plate was cut open to inspect the core material (Figure 5.10).
Inspection of the sample showed that the entire depth of the sample had been properly

infused with resin.

   

(3) Cut through the sample (b) Wet-out through the entire depth

Figure 5.10. Cut through VARTM cellular plate showing fiber wet-out through the entire depth

5.4.3 Testing of VARTM Cellular Plates

The VARTM manufactured cellular plates were tested in three-point bending
using the same setup described in section 3.3.4. Three VARTM cellular plates were
tested, as shown in Table 5.1 and Figure 5.11 through Figure 5.13.

All test units were of cellular plate type 1 (Figure 3.4) with a relative density of

0.56. A polyester resin system with 10% styrene monomer was used as the matrix for all

197

test units. Notice that with VARTM manufacturing it was possible to achieve much
higher fiber weight fractions that by hand layup (Table 5.1). The increase in fiber weight
fraction by VARTM manufacturing is mainly due to the use of dense bio-fiber pre-form
fabrics, which avoid the manufacturing issues encountered with the loose springy natural

fibers in the experimental studies of Chapter 3.

Table 5.1. VARTM cellular plate test matrix

 

Placement Test Fiber Weight

Pre-Form Fabric Tm Method Unit ID Fraction (%)
Flaw hemp fiber core, jute fabric face sheets Non-wrapped CP1-01 40%
Raw hemp fiber core, jute fabric face sheets Wrapped CP1-02 36%
Raw hemp fiber core, glass strand mat face sheets, Wrapped CP1-C 47%

top and bottom carbon layers

 

(a) Overall view of the sample (b) View of the side wall

Figure 5.11. VARTM cellular plate with a non-wrapped jute fabric/raw hemp core pre-form
(40% fibers by weight)

 

(3) Overall view of the sample (b) View of the side wall

Figure 5.12. VARTM cellular plate with a wrapped jute fabric/raw hemp core pre-form

198

(36% fibers by weight)

   

(a) Overall view of the sample (b) View of the side wall

Figure 5.13. VARTM cellular plate with a warpped glass strand mat/raw hemp core pre-form and
top and carbon layers (47% fibers by weight)

5.4.3.1 F lexural Test Results and Observed Behavior

The results from the flexural tests performed on the VARTM cellular plates are
tabulated in Table 5.2 and shown in Figure 5.14 and Figure 5.15. The force—displacement
response of the hybrid VARTM plate with top and bottom carbon layers is compared
with the hybrid test units made with carbon layers (Chapter 3) in Figure 5.16. The test
results of the VAR'IM cellular plates are compared with the results from all the cellular
plates made by hand layup tested in the structural experimental program (Chapter 3) in
Table 5.3.

All three VARTM manufactured test units behaved linear-elastically up to failure
but failed in different modes. The test unit manufactured with three separate pro-form
fabrics failed by delarrrination of the top layer with a loud snapping noise and no visible
cracks on the exterior of the plate. The test unit manufactured with one continuous fabric
failed in tension on the bottom surface with fiber cracking noises near failure. The unit

maintained a nominal load, at approximately 10% of the peak load, after tension failure in

199

the bottom layer. The test unit manufactured with the top and bottom carbon layers failed
in buckling of the top carbon layer. After buckling of the top carbon layer the sample did
not retain significant stiffness from the bottom carbon layer and rapidly lost its load
carrying capacity (Figure 5.14).

All VARTM cellular plates had a higher stiffness compared to the samples made
by hand layup (Table 5.3). For example, the VARTM cellular plate with the woven pre-
form had a stiffness of 582 N/mm while the green hemp cellular plate with top and
bottom jute fabric layers had a stiffness 309 N/mm (Table 5.3), an improvement of 88%.
This improved stiffness is mainly due to the higher fiber content achieved using the
VARTM process (Table 5.1). The hybrid VARTM cellular plate with top and bottom
carbon fabric also showed improved stiffness over the hand layup cellular plate with top
and bottom carbon layers. The VARTM cellular plate with top and bottom carbon had a
stiffness of 1,605 N/mm while the hand layup cellular plate with top and bottom carbon
had a stiffness of 1,480 N/mm (Table 5.3), an improvement of 8%. Despite the
improvement in stiffness, the strength of the VARTM cellular plate with top and bottom
carbon (5,596 N) was much lower that the hand layup cellular plate (11,010 N), see
Figure 5.16. The reduction in strength in the VARTM cellular plate is due to
delamination of the top layer of the pre-form fabric after buckling of the top carbon layer.
Because the pre-form fabric is only bonded to the core in a lirrrited number of locations
(see Figure 5.6) the plate is more susceptible to delamination than the hand layup sample,

which has the carbon layer built integrally with the core.

200

 

PI.

 

 

 

 

 

   
       
 

 

 

 

 

 

 

 

6000 ‘
~. — CP1-01: VARTM WIUI‘Wll’appCd Pro-Form
. / ‘1 — CP1-02: VARTM wIWrapped Pro-Form
5000 1~ /' i CP1-C: VARTM VII/T 818 Carbon Layers
* L ,5, — CP1: Green Hemp
l // ”a CP1-02
400° 1"— f \\
f 1 '/ \\-\ CP1-C
v CP1-01 ‘7“ ---. -~—— ~~ -- 4. ---__ g
B 3000 “\* n..-
3 Plate Data: i.=267mm,
. b=102mm, h=13mm
200° 3 Test: 3-Pt Bending
Location: Mid-Span
1000 3 MM
‘ CP1-01, 02, C: Jute FabrchHemp
Pro-Form Fabric/UPE
o . +4 . . . P . . s . s . .
0 5 10 15 20 25
Displacement (mm)

Figure 5.14. Force versus dkplacement response at mid-span of VARTM cellular plates, compared
with a green hemp plate made by hand layup

 

 

 

 

 

 

 

 

 

 

 

 

 

-—-CP1 -01: VARTM VII/Unwrapped Pro-Form
6000 . —CP1-02: VARTM Irv/Wrapped Pro-Form
. —CP1-C: VARTM wires Carbon Layers
j ——CP1: Green Hemp
5000 :-
4000 «~
A .
5 . CP1-02
'o 3000 ..
fl .
3 f Plate Data: L=267mm,
1 b=102mm, h=13mm
2°00 1" Test: 3-Pt Bending
3 Location: Mid-Span
1000 3. Material Data:
j CP1-01, 02, C: Jute Fabric/Hemp
. Pro-Form FabricJUPE
OJ .................. i ......... ifi..i....+rr4.i..r.
0 1 2 3 4 5 6 7 e 9 10

Displacement (mm)

Figure 5.15. loomed-in view of the force displacement response of VARTM cellular plates

201

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12000 ~~ -—k* "—— - A *— — — -- A a
— CP1-VARTM-C: T&B Carbon Layers
: — CP1-C: Bottom Carbon Layer CP2":
10000 ._ — CP2-C: T&B Carbon Layers
: Plate Data: cp1_c
j b=102mm, h=13mm, L=267mm
aooo .. Test:
: 3-Pt Bending
A
z 1 Location:
a ..
3 :
4000 j CP1-VARTM-C
j X
4 Material DEE:
2000 1- CP1-VARTM-C: Glass Strand/Hemp
1 Pro-Form FabricIUPE
‘ CP1-C a CP2-C: Hemp/LIFE
of-..1....4s4..s....:...4
0 5 10 15 20 25

Displacement (mm)

Figure 5.16. Comparison of the force versus displacement response at mid-span of the VARTM plate
with top and bottom carbon versus the hand layup cellular plates with carbon face sheets

Table 5.2. Flexural test results summary of VARTM cellular plates

 

 

Fibers
by Wt. Max Load Max Displ. Stiffness
Material 4%) ID (N) (mm) (N/mm)
Jute/Chopped Hemp Fabric (non-wrapped) 40 CP1-01 3655 5.99 729
Jute/Chopped Hemp Fabric (wrapped) 36 CP1-02 3287 8.44 582
Glass/Chopped Hemp Fabric T&B Carbon 47 CP1-C 5596 23.25 1605

 

202

"Jane—— —H

 

Table 5.3. Flexural test results of the VARTM plates with all cellular plates tested in Chapter 3

 

 

Fibers Extensometer
by Wt. Max Load Max Displ. Stiffness
Material (%) IO (N) (mm) (N/mm)
G.HemplUPE 10 CP1 1483 3.65 406
G.HemplUPE - T&B Glass 10 CP1-G 2237 8.42 320
G.HemplUPE - Bottom Carbon 10 CP1-C 10397 19.70 592
G.HemplUPE - T&B Jute 10 CP1-J 2197 6.95 309
Raw Hemp/UPE 10 CP1 1859 5.67 323
Raw Hemp/UPE 10 CP2 2557 5.39 502
Raw Hemp/UPE - T&B Carbon 10 CP2-C 11010 17.56 1480
Flax/UPE 20 CP1 1172 6.16 196
FIax/UPE - T&B Jute 20 CP1-J 1980 7.77 259
Polyester 0 CP1 1764 9.45 186
Jute/Chopped Hemp Fabric (non-wrapped) 40 CP1-01 3655 5.99 729
Jute/Chopped Hemp Fabric (wrapped) 36 CP1-02 3287 8.44 582
Glass/Chopped Hemp Fabric T&B Carbon 47 CP1-C 5596 23.25 1605
Douglas Fur -- DF-P 3495 1 1.19 629
Orientated Strand Board - OSB-P 917 5.56 196
Particle Board - PB—P 694 6.15 123

 

5.5 Recommendations for Future Work

Recommendations for future work based on the findings of this study are
provided for both the biocomposite material systems and the structural forms in which
the biocomposite materials can be used. The future work for the material systems mainly
lies in the optimization of treatments for the fibers and resin, which can improve physical
and mechanical properties needed for structural applications Future work in structural
forms lies in optimization of the hierarchical structure and automated manufacturing of
full-scale components. Once the material system, structural form, and manufacturing
procedure have been optimized, probabilistic studies should be performed to define
strength reduction factors for design purposes. For design using the LRFD approach the
variations due to the material system, processing method, level of cure, performance level

in processing, and degradation in material properties over time and due to environmental

203

 

effects are needed in order to determine an overall safety factor (Karbhari and Seible
2000). Thus, the deviations in material properties and the processing effects must be

studied and measured.

5.5.1 Material Systems

Future work is needed in the characterization of biocomposite materials for
physical properties such as durability, weathering, creep, and susceptibility to fire, as
these properties are important for structural applications. In addition, the compression
behavior of biocomposite materials remains to be characterized. Once these properties
are assessed, future work is needed in the study of optimal fiber and resin treatments for
enhancement of the properties of biocomposite material systems. Possible treatments
include chemical treatments of the fibers for improved fiber-matrix adhesion, sizing of
fibers for the resin, improved processing of fibers, and the use of nano-reinforcement in
the resin (Vaia 2002). Treatments should be targeted for mechanical and physical
properties that may limit the use of biocomposites in load-bearing applications. The
treatments considered should also be economical to keep the low cost appeal of

biocomposites.

5.5.2 Structural F orms

Future work is needed in the optimization of hierarchical structures for
biocomposite cellular beams and plates. Optimization studies of hierarchical structures
can be performed using the analysis procedure proposed in Section 4.5. The studies
should not only be aimed at efficient structures in flexure, but other properties such as

thermal conductivity, blast, impact, and their combinations, to provide efficient structural

204

 

forms for biocomposite materials. In addition, three-dimensional cellular structures and
micro-sandwiching of the cell walls may be considered in the optimization of hierarchical
structures.

Full-scale manufacturing of biocomposite structural components must also be
studied in the future. Based on the results from the research work reported here,
VARTM is recommended for full-scale manufacturing of cellular biocomposite structural
components. The manufacturing procedure should be optimized for obtaining high fiber
volume fractions and fast production cycles for high-volume manufacturing in the
construction industry. Collaboration with industry may be helpful in developing the

manufacturing procedure.

205

 

I

 

CHAPTER 6 CONCLUSIONS

Results from the reported studies has shown that hybrid cellular biocomposite
beams and plates not only have the potential to serve as primary load bearing
components, but that they can compete with conventional structural materials. In
particular, the findings in this study have led to the following conclusions regarding the
biocomposite material systems:

0 The mechanical properties of biocomposite materials can compete with E-

glass fiber reinforced composites.

0 Biocomposite material systems have deficiencies with respect to impact
strength, and moisture absorption.

0 The use of hybrid material systems with synthetic or aligned fiber face sheets
and chopped natural fiber cores dramatically improve the mechanical
properties of biocomposites.

0 Hybrid material systems with fabric face sheets and chopped natural fiber
cores can serve as novel engineered pre-forms, or micro-sandwiches, for

VARTM manufacturing.

The findings in this study have also led to the following conclusions regarding the
biocomposite structural systems:
0 Cellular biocomposite structures can serve as load bearing structural

components, and compete with conventional solutions.

206

 

o The use of hierarchical cellular structures improves the structural performance
of biocomposite beams and plates.

0 The use of jute, glass strand mat, and carbon face sheets can dramatically
improve the stiffness and strength of cellular biocomposite structural
components.

0 Vacuum assisted resin transfer molding (VARTM) is a viable method for

automated manufacturing of biocomposite cellular structures.

In civil construction, biocomposite structural components can be used as flooring
systems and pre-fabricated components in residential and commercial construction, deck
systems in highway bridges, and complete superstructures for small pedestrian bridges.
Further research, development, and collaboration between researchers and industry are
needed in order for biocomposite structural components to become a reality in load-
bearing applications. However, the availably of low cost structural components based on

renewable resources will be a great asset for current and future structural applications.

207

 

 

REFERENCES

208

 

REFERENCES

AASHTO—Lm Bridge Desig Specifications. American Association of State Highway
and Transportation Officials, 1998, Washington, DC.

Allen, H.G. Analysis and Design of Structural Sandwich Panels, Pergarnon Press, 1969,
Oxford.

APA Engineered Wood Association. Company Homepage [online] Available:
<httD://www.2_1pgwood.org[> (Accessed July 2003).

Ashby, M.F. Materials Selection in Mechanical Desigr_i. Butterworth Heinemann, 1999,
Oxford, UK.

Ashby, M.F. (1991). “Materials and Shape.” Acta Metallugical et Materialia, 39, 1025-
1039.

Barbero, E. and GangaRao, H.V.S. (1991). “Structural Applications of composites in
Infrastructure: Part1.” SAMPE Journal, Vol. 27, No. 6, 9-16.

Barbero, E. and GangaRao, H.V.S. (1992). “Structural Applications of composites in
Infrastructure: Part II.” SAMPE Journal, Vol. 28, No. 1, 9-15.

Barker, RM and Puckett, J .A. Design of Highway Bridges: B_a_sed on AASMOW.
Bridge Design Specifications. John Wiley, 1997, New York, NY.

Baxter, W.J. (1992). “The Strength of Composites with Randomly Oriented Short
Fibers.” Proceedings of the ASC 7th Annual Technical Conference, University
Park, Pa, 615-623.

Berenberg, B. (2003). “Manufacturers Welcome New Reinforcement Forms.”
Composites Technology, April, 12-16.

Biswas, S., Mittal, A. and Srikanth, G. (2002). “Composites: A Vision for the Future.”
Proceedings of the International Conference and Exhibition on Reinforced

Plastics, Indian Institute of Technology, Madras, 26-36.

Biswas, S., Srikanth, G. and Nangia, S. (2001). “Development of Natural Fibre
Composites in India.” Proceedings of the Composites Fabricators Association ’8
Conference, Tampa, FL.

Brouwer, W.D. (2000). “Natural Fibre Composites: Where Can Flax Compete with
Glass?” SAMPE Journal, Vol. 36, No. 6, 18-23.

Christensen, RM. and Waals, FM. (1972). “Effective Stiffness of Randomly Oriented
Fibre Composites.” Journal of Composite Materials, 6, 518-532.

209

 

 

 

Ghali, A. and Nevill A.M. Structural Analysis: A Unified Classical and Matrix Approach.
4th edition, E&FM Spon, 1997, London, UK.

Gunderson, S.L. and Thorp, K.E.G. (1993). “Natural Cellular and Sandwich Structures
for Innovative Design Concepts.” Proceedings of the ASC 8th Annual Technical
Conference, Cleveland, OH, 431-440.

Chen, RE. (1971). “Strength Properties of Discontinuous Fiber Composites.” Polymer
Engineering and Science, Vol. 11, No. 1, 51-56.

Daimler Chrysler Corporation (2001). “Daimler Chrysler 2001 Environmental Report.”
Auburn Hills, MI.

Drzal, L.T., Mohanty, A.K., Mehta, G. and Misra M. (2002). “Low-Cost, Bio-Based
Materials for Housing Applications.” International Conference on Advances in
Building Technology, Hong Kong, China, 225-232.

Gibson, LI. and Ashby, M.F. Cellular Solids; Structure and Properties. lst edition,
Pergamon Press, 1988, Oxford, UK.

Gibson, R.F. Principles of Composite Material Mechanics. McGraw-Hill, 1994, New
York, NY.

Giurgiutiu, V. and Reifsnider, KL. (1994). “Development of Strength Theories for
Random Fiber Composites.” Journal of Composites technology and Research,
Vol. 16, No. 2, 103-114.

Gunderson, S.L. and Thorp, K.E.G. (1993). “Natural Cellular and Sandwich Structures
for Innovative Design Concepts.” Proceedings of the ASC 8th Annual Technical
Conference, Cleveland, OH, 431-440.

Halpin, J .C. (1969). “Stiffness and Expansion Estimates for Oriented Fiber Composites.”
Journal of Composite Materials, 3, 732-734.

Halpin, J.C. and Pagano, NJ. (1969). “The Laminate Approximation for Randomly
Oriented Fibrous Composites.” Journal of Composite Materials, 3, 720-724.

Halpin, J.C. and Tsai, S.W. (1969). Effects of Environmental Factors on Composite
Materials, Technical Report AFML-TR-67-423, June.

Hibbler,R.C. Mechanics of Materials. 3rd Edition, Prentice Hall, 1994, Upper Saddle
River, NJ.

Huang, IS. and Gibson, L]. (1995). “Microstructural Design of Cellular Materials:
Honeycomb Beams and Plates.” Acta Metallugical et Materialia, 43, 1643-1650.

Insulspan. Company Homepage. [online] Available: <http://www.insulspan.com/>
(Accessed July 2003).

210

 

l_:

IntemationzaBuildinLCode. International Code Council, 2000, Falls Church, VA.

Jones, RM. Mechanics of Composite Materizfi. 2nd edition, Taylor and Francis, 1999,
Philadelphia, PA.

Karbhari, V.M. and Seible, F. (2000). “Fiber Reinforced Composites - Advanced
Materials for the Renewal of Civil Infrastructure.” Applied Composite Materials,
7, 95-124.

Lees, J .K. (1968). “A Study of the Tensile Strength of Short Fiber Reinforced Plastics.”
Polymer Engineering and Science, Vol. 8, No. 3, 195-201.

Mathcad. Computer Software. Mathsoft, 2001 Professional, 2000. Cambridge, MA.
Mathematica. Computer Software. Wolfram Research, version 4.2, 2002. Champaign, IL.
MATLAB. Computer Software. The MathWorks, version 6.5, 2003. Natick, MA.

Mishra, S., Mohanty, A.K., Drazal, L.T., Misra, M., Parija, S., Nayak, SK. and Tripathy,
8.8. (2003). “Studies on Mechanical Performance of Biofiber/Glass Reinforced
Polyester Hybrid Composites.” Composites Science and Technology, 63, 1377-
1385.

Mishra, S., Misra, M., Tripathy, S.S., Nayak, SK. and Mohanty, AK. (2002). “The
Influence of Chemical Surface Modifications on the Performance of Sisal-
Polyester Biocomposites.” Polymer Composites, 23, 164-170.

Mohanty, A.K., Misra, M. and Drzal, LT. (2001). “Surface Modifications of Natural
Fibers and Performance of the Resulting Biocomposites: An Overview.”
Composite Interfaces, 8, 313-343.

Mohanty, A.K., Misra, M. and Hinrichsen, G. (2000). “Biofibers, Biodegradable
Polymers and Biocomposites: An Overview.” Macromolecular Materials and
Engineering, Vol. 276/277, 1-24.

Nickle, J. and Riedel, U. (2003). “Activities in Biocomposites.” Materials Today, Vol. 6,
No. 4, 44-48.

PCI Design Handboolg Prestressed and Precast Concrete. Precast Concrete Institute,
1999, Chicago, IL.

Phenix Biocomposites. Company Homepage [online] Available:
<http://www.phenixbiocomposites.com/> (Accessed July 2003).

Piggott, M.R. glad-Bearing Fibre Composites. Pergamon Press, 1980, Oxford, UK.

211

 

Rout, J ., Misra, M., Tripathy, S.S., Nayak, SK. and Mohanty, AK. (2001). “The
Influence of Fibre Treatment on the Performance of Coir-Polyester Composites.”
Composites Science and Technology, 61, 1303-1310.

Shenton III, H.W., Wool, R.P., Hu, B., O’Donnell, A.O., Bonnaillie, L., Can, E., Chapas,
R. and Hong, C. (2002). “An All-Natural Composite Material Roof System for
Residential Construction.” International Conference on Advances in Building
Technology, Hong Kong, China, 255-262.

Stoll, F. and Banerjee, R. (2001), “Measurement and Analysis of Fiber-Composite-
Reinforced-Foam Sandwich Material Properties.” Proceedings of the 46th
International SAMPE Symposium and Exhibition, Long Beach, CA, 1938-1946.

Strongwell. Company Homepage. [online] Available:<http://www.strongwell.com/>
(Accessed July 2003).

Tsai, S.W. and Pagano, NJ. (1968). “Invariant Properties of Composite Materials.”
Composite Materials Workshop, Lancaster, PA, 233-252.

Thompson, D.W. On Growtlgmd Form. Cambridge University Press, 1961, Cambridge,
MA.

Timoshenko, SP. (1925). “Analysis of Bi-Metal Thermostats.” Journal of the Optical
Society of America, 11, 233-255.

Vaia, RA (2002). “Polymer Nanocomposites Open a New Dimension for Plastics and
Composites.” The AMPTIAC Newsletter, Vol. 6, No. 1, 17-24.

212

-v- P. '7‘. Mun—5*

 

 

a!