Chad P. Kirlin for the degree of Master of Science in Forest Products and Mechanical Engineering presented on April 30, 1996. Title: Experimental and Finite-Element Analysis of Stress Distributions Near the End of Reinforcement in Partially Reinforced Glulaiu./ Signature redacted for privacy. -1'Robert J. 'Leichti Abstract approved: Abstract approved: N ABSTRACT OF THE THESIS OF Signature redacted for privacy. (Timoth C. yhedY Recently, fiber-reinforced plastics (FRP lamina) have been applied to glued-laminated (glulam) timber for the purpose of improving bending strength and stiffness. Initially, full length reinforcement using FRP lamina was developed. However, the cost of FR? lamina is a significant portion of the total cost of reinforced glulain. Therefore, it is advantageous to use reinforcement in high-moment areas of a beam. A glulain beam reinforced over less than the full length is referred to as "partially reinforced glulant." The understanding of in-service FRP lamina-wood interactions is limited. While stress distributions in full- length reinforced beams have been studied, there is a lack of information regarding stress distributions at the end of reinforcement in partially reinforced glulam beams. Interfaces and joints in composites are known to be areas of
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Chad P. Kirlin for the degree of Master of Science in ForestProducts and Mechanical Engineering presented on April 30,1996. Title: Experimental and Finite-Element Analysis ofStress Distributions Near the End of Reinforcement inPartially Reinforced Glulaiu./
Signature redacted for privacy.-1'Robert J. 'Leichti
Abstract approved:
Abstract approved:
N ABSTRACT OF THE THESIS OF
Signature redacted for privacy.(Timoth C. yhedY
Recently, fiber-reinforced plastics (FRP lamina) have
been applied to glued-laminated (glulam) timber for the
purpose of improving bending strength and stiffness.Initially, full length reinforcement using FRP lamina was
developed. However, the cost of FR? lamina is a significant
portion of the total cost of reinforced glulain. Therefore,
it is advantageous to use reinforcement in high-moment areas
of a beam. A glulain beam reinforced over less than the full
length is referred to as "partially reinforced glulant."The understanding of in-service FRP lamina-wood
interactions is limited. While stress distributions in full-length reinforced beams have been studied, there is a lack
of information regarding stress distributions at the end of
reinforcement in partially reinforced glulam beams.
Interfaces and joints in composites are known to be areas of
stress concentrations and failure initiation. The research
conducted in this study investigates the stress distributionat the end of tensile FRP reinforcement experimentally and
analytically.Experimental analysis of stress distributions was
performed on several partially reinforced glulam beams.
Strain gage analysis was used to measure axial strain (along
beam length) near the end of the FRP larnina. The analysis
indicated that strain (or stress) just past the end of theFRP laxnina is higher than elementary beam theory predicts.
Finite-element modeling was used to model partially
reinforced glulain to investigate potential effects on stresscomponents imposed by alternative geometries, loadings, and
materials. Specifically, the effects on stress distributiondue to FRP lamina thickness, FRP lamina stiffness, beam
width, percent length of reinforcement, span-to-depth ratio
and type of loading were investigated. A three-dimensional
structural solid element was used to model wood and the FRP
lamina in linear elastic analysis. Failure loads and
mechanisms were beyond the scope of this thesis.
Most stress distributions were found to be singular atthe end of reinforcement. In order to quantify themagnitude of each stress, average stress near the end of theFRP lamina was calculated.
The models suggest that FRP lamina thickness and stiffness
have significant effects on the magnitude of stress
components near the end of the FRP lainina.
®Copyright by Chad P. KirlinApril 30, 1996
All Rights Reserved
Experimental and Finite-Element Analysis of StressDistributions Near the End of Reinforcement in Partially
Reinforced Glulam
by
Chad P. Kirliri
A THESIS
submitted to
Oregon State University
in partial fulfillment ofthe requirements for the
degree of
Master of Science
Presented April 30, 1996Commencement June 1996
ACKNOWLEDG4ENTS
Funding for this project was provided by American
Larninators and the Center for Wood Utilization and Research.
Technical support was provided by the Wood Science and
Technology Institute.
I would also like to thank the Department of Forest
Products and College of Forestry for putting up with me and
providing me with scholarship and fellowship support over
the past six years. I would like to thank Bob Leichti for
his unrelenting patience with me over the past three years
and for his superb editorial skills.
ThBLE OF CONTENTS
Page
INTRODUCTION 1
Justification of Research 3
Objectives 6
LITERPTURE REVIEW 7
Composite Beam Theory 7
Reinforcement of Wood Products 14
Reinforcement Using Metal Plates 14
Reinforcement Using Steel Bars and Cables 15
Reinforcement Using Fiber Reinforced Plastics 17
Summary of Reinforcement 18
Stress Distributions 19
Partially Reinforced Glulam 19
Single and Double Lap-Joints 20
Saint-Venant Effects 22
EXPERIMENTAL ANALYSIS 23
Beam Manufacturing and Description 23
Strain Gage Placement 25
Strain Measurements 27
Experimental Results and Discussion 29
FINITE-ELEMENT 1NALYSIS 40
Material Orientation 40
Material Characterization for Finite-Element2nalysis 42
11
ThBLE OF CONTENTS (Continued)
Page
Wood 42
Adhesive 44
Fiber-Reinforced Plastics 45
Fiber and Matrix Materials 45
FRP Lamina Properties 46
Design of Analytical Investigation 48
Parameters of Study 48
Model Loading 50
Finite-Element Model Description 51
Element Mesh 54
Model Verification 60
Finite-Element Model Results 62
Methodology for Stress DistributionCharacterization 64
Glulam Without a Bumper 65
Maximum Stress Levels and StressDistribution 65
Effect of Reinforcement Thickness 75
Effect of Stiffness Ratio 84
Effect of Reinforcement Length 100
Effect of Beam Width 100Effect of Span-to-depth Ratio 109Effect of Loading Conditions 109
Glulam With a Bumper 114
Maximum Stress Levels and StressDistribution 119
Effect of Reinforcement Thickness 131
Effect of Stiffness Ratio 139Effect of Reinforcement Length 147
Effect of Beam Width 147
Effect of Span-to-depth Ratio 154
111
Summary of Analytical Results 154
TABLE OF CONTENTS (Continued)
Page
V. CONCLUSION 161
BIBLIOGRAPHY 162
APPENDICES 167
Appendix A Curve Fits for Stress Distributions 168
Appendix B ANSYS Finite-Element Model Input 172
iv
LIST OF FIGURES
Figure Page
Partially reinforced glulam beam with a cover(bumper) lamination 4
Typical cross section of a FRP lamina reinforcedglulam beam 7
Description of load and reinforcement locationsfor deflection calculation 12
Strain gages mounted near the end of the FRPlamina on a partially reinforced glulam beam 26
Block-diagram of the strain measurement setup 28
Stress distribution near the end of the FRP laminain beam 1 30
Stress distribution near the end of the FRP laminainbeam2 31
Stress distribution near the end of the FRP laminain beam 3 32
Stress distribution near the end of the FRP laminain beam 4 33
Stress distribution near the end of the FRP laminain beam 5 34
Stress distribution near the end of the FRP laminain beam 6 35
Stress distribution near the end of the FRP laIninain beam 7 36
Visual comparison of material directions used infinite-element model and material directionscommonly used for wood 41
The standard finite-element model for a partiallyreinforced glulam without a bumper having simplysupported end conditions 55
V
LIST OF FIGURES (Continued)
Figure Page
The standard finite-element model for a partiallyreinforced glulam with a bumper having simplysupported end conditions 56
Side view of finite-element mesh at end of the FRPlamina for a beam without a bumper 57
Side view of finite-element mesh at the end of theFRP lamina for a beam with a bumper 58
Finite-element mesh for beams with and without abumper 59
Distribution of near the end of the FRP laminaon the wood at the outside edge of the beam forglulam without a bumper 67
20 Distribution of a, near the end of the FRP laminaon the wood at the outside edge of the beam forglulam without a bumper 68
21 Distribution of near the end of the FRP larninaon the wood at the mid-with of the beam for glulamwithout a bumper 70
22 Distribution of near the end of the FRP laminaon the wood at the outside edge of the beam forglulam without a bumper 72
Distribution of near the end of the FRP laminaon the wood at the outside edge of the beam forglulam without a bumper 73
Distribution of near the end of the FR? laminaon the wood at the outside edge of the beam forglulam without a bumper 74
Typical distribution of c along the beam lengthand across beam width near the end of the FR?lainina for beams without a bumper 76
vi
LIST OF FIGURES (Continued)
Figure Page
Typical distribution of c along the beam lengthand across beam width near the end of the FRPlamina for beams without a bumper 77
Typical distribution of c along the beam lengthand across beam width near the end of the FRPlamina for beams without a bumper 78
Typical distribution of along the beam lengthand across beam width near the end of the FRPlamina for beams without a bumper 79
The distribution of is plotted for various FRPthickness on a beam without a bumper 80
The distribution of c is plotted for various FRPthickness on a beam without a bumper 81
The distribution of is plotted for various FRPthickness on a beam without a bumper 82
The distribution of is plotted for various FRPthickness on a beam without a bumper 83
The effect of the FRP lamina thickness on c nearthe end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and0.15 in. past the end of the FRP lamina 86
The effect of the FRP lamina thickness on nearthe end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and0.15 in. prior to the end of the FRP lamina 87
The effect of the FRP lamina thickness on c nearthe end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and0.15 in. past the end of the FRP lamina 88
vii
LIST OF FIGT3RS (Continued)
Figure Page
The effect of the FRP lamina thickness on nearthe end of the FR? lamina is plotted for averagestress from the end of the FR? lamina to 0.05 and0.15 in. prior to the end of the FRP lamina 89
The distribution of is plotted for variousFRP-to--wood stiffness ratios on a beam without abumper 91
The distribution of is plotted for variousFRP-to-wood stiffness ratios on a beam without abumper 92
The distribution of is plotted for variousFRP-to-wood stiffness ratios on a beam without abumper 93
The distribution of is plotted for variousFRP-to-wood stiffness ratios on a beam without abumper 94
The effect of the FRP-to-wood stiffness ratio on s,near the end of the FR? lamina is plotted foraverage stress from the end of the FR? lamina to0.05 and 0.15 in. past the end of the FR? lamina 96
The effect of the FRP-to-wood stiffness ratio onnear the end of the FR? lamina is plotted foraverage stress from the end of the FR? lamina to0.05 and 0.15 in. prior to the end of the FRPlamina 97
The effect of the FRP-to-wood stiffness ratio on onear the end of the FR? lamina is plotted foraverage stress from the end of the FRP lamina to0.05 and 0.15 in. past the end of the FRP lamina 98
The effect of the FR?-to-wood stiffness ratio onnear the end of the FRP lamina is plotted foraverage stress from the end of the FR? lamina to0.05 and 0.15 in. prior to the end of the FR?lamina 99
VIII
LIST OF FIGURES (Continued)
Figure Page
The distribution of c is plott?d for various FRPlamina lengths on a beam without a bumper 101
The distribution of c is plotted for various FRPlamina lengths on a beam without a bumper 102
The distribution of a is plotted for various FRPlamina lengths on a beam without a bumper 103
The distribution of is plotted for various FRPlamina lengths on a beam without a bumper 104
The distribution of is plotted for various beamwidths on a beam without a bumper 105
The distribution of c is plotted for various beamwidths on a beam without a bumper 106
The distribution of c is plotted for various beamwidths on a beam without a bumper 107
The distribution of is plotted for various beamwidths on a beam without a bumper 108
The distribution of c is plotted for variousspan-to-depth ratios on a beam without a bumper 110
The distribution of o is plotted for variousspan-to-depth ratios on a beam without a bumper 111
The distribution of c is plotted for variousspan-to-depth ratios on a beam without a bumper 112
The distribution of is plotted for variousspan-to-depth ratios on a beam without a bumper 113
The distribution of c is plotted for uniform andthird-point loading on a beam without a bumper 115
lx
LIST OF FIGURES (Continued)
Figure Page
The distribution of c is plotted for uniform andthird-point loading on a beam without a bumper 116
The distribution of a is plotted for uniform andthird-point loading on a beam without a bumper 117
The distribution of a is plotted for uniform andthird-point loading on a beam without a bumper 118
Distribution of c near the end of the FRP laminaon the wood at the outside edge of the beam forglulam without a bumper 120
Distribution of o, near the end of the FRP laminaon the wood at the outside edge of the beam forglulam with a bumper 121
Distribution of o near the end of the FRP laminaon the wood at mid-width of the beam for glulamwith a bumper 122
64 Distribution of o near the end of the FR? laminaon the wood at the outside edge of the beam forglulam with a bumper 123
65 Distribution of o near the end of the FR? laminaon the wood at the outside edge of the beam forglulam with a bumper 124
66 Distribution of near the end of the FR? laminaon the wood at the outside edge of the beam forglulam with a bumper 125
Typical distribution of c along the beam lengthand across beam width near the end of the FR?lamina for beams with a bumper 128
Typical distribution of a along the beam lengthand across beam width near the end of the FR?lamina for beams with a bumper 129
x
LIST OF FIGURES (Continued)
Figure Page
Typical distribution of a, along the beam lengthand across beam width near the end of the FRPlamina for beams with a bumper 130
The distribution of is plotted for various FRPthickness on a beam with a bumper 132
The distribution of a7 is plotted for various FRPthickness on a beam with a bumper 133
The distribution of a, is plotted for various FRPthickness on a beam with a bumper 134
The effect of the FRP lamina thickness on c nearthe end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and0.15 in. past the end of the FRP lamina 136
The effect of the FRP lamina thickness on o nearthe end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and0.15 in. past the end of the FRP lamina 137
The effect of the FRP lamina thickness on cy, nearthe end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and0.15 in. prior to the end of the FRP lamina 138
The distribution of a is plotted for variousFRP-to-wood stiffness ratios on a beam with abumper 140
The distribution of c is plotted for variousFRP-to-wood stiffness ratios on a beam with abumper 141
The distribution of a, is plotted for variousFRP-to-wood stiffness ratios on a beam with abumper 142
xi
LIST OF FIGURES (Continued)
Figure Page
The effect of the FRP-to-wood stiffness ratio on cnear the end of the FRP lainina is plotted foraverage stress from the end of the FRP lamina to0.05 and 0.15 in. past the end of the FRP larnina. . . .144
The effect of the FRP-to-wood stiffness ratio on cnear the end of the FRP lamina is plotted foraverage stress from the end of the FRP lamina to0.05 and 0.15 in. past the end of the FRP lamina. . . .145
The effect of the FRP-to-wood stiffness ratio onnear the end of the FRP lamina is plotted foraverage stress from the end of the FRP lamina to0.05 and 0.15 in. prior to the end of the FRPlamina 146
The distribution of c is plotted for two FRPlamina lengths on a beam with a bumper 148
The distribution of o is plotted for two FRPlamina lengths on a beam with a bumper 149
The distribution of c, is plotted for two FRPlamina lengths on a beam with a bumper 150
The distribution of is plotted for two beamwidths on a beam with a bumper 151
The distribution of c is plotted for two beamwidths on a beam with a bumper 152
The distribution of o, is plotted for two beamwidths on a beam with a bumper 153
The distribution of c is plotted for two span-to-depth ratios on a beam with a bumper 155
The distribution of o is plotted for two span-to-depth ratios on a beam with a bumper 156
xl'
LIST OF FIGURES (Continued)
Figure Page
90. The distribution of o is plotted for two span-to-depth ratios on a beam with a bumper 157
LIST OF TABLES
Table Page
Comparison of adding reinforcement to the tensionface of a beam 13
Beam size, lamination combination, and degree ofreinforcement. All beams manufactured as AITCcombination No. 5 - FiRP Glulam 24
Wood property calculations and properties used inthe finite-element models 43
FRP lamina elastic properties used in the finite-element models 48
Description of finite-element models for beamswithout a bumper 52
Description of finite-element models for beamswith a bumper 53
Closed-form and finite-element model values forcenterline deflection 61
Strength properties for small, clear, straight-grained samples interior west Douglas-fir WoodHandbook, 1987) 63
Stresses and stress ratios at the end of the FRPlamina for various FRP thicknesses for glulamwithout a bumper 85
Stresses and stress ratios at the end of the FRPlamina for various stiffness ratios for glulamwithout a bumper
Stresses and stress ratios at the end of the FRPlamina for various FRP thicknesses for glulamwith a bumper 135
Stresses and stress ratios at the end of the FRPlamina for various stiffness ratios for glulamwith a bumper 143
xiv
LIST OF APPENDIX TABLES
Table Page
Al. Coefficients for polynomial curve fits to finite-element stress distributions for glulam beamswithout a bumper, where thickness of reinforcementis varied 168
Coefficients for polynomial curve fits to finite-element stress distributions for glulam beamswithout a bumper, where stiffness of reinforcementis varied 169
Coefficients for polynomial curve fits to finite-element stress distributions for glulam beamswith a bumper, where thickness of reinforcement isvaried 170
Coefficients for polynomial curve fits to finite-element stress distributions for glulam beamswith a bumper, where stiffness of reinforcementis varied 171
xv
Experimental and Finite-Element Analysis of StressDistributions Near the End of Reinforcement in Partially
Reinforced Glulam
I. INTRODUCTION
The glue-laminated timber (glulam) industry produces
beams, columns and beam-columns for use in residential and
commercial structures. In 1993, the glulam industry used 260
million board feet of lumber (APA, 1995) . The Tacoma Dome,
Tacoma, Washington, spanning more than 500 ft, is one
example of the many structures glulam is be used for (FPL,
1987).
Glulam timbers are structural members composed of two
or more layers of structural lumber glued together. The
laminations are typically one or two-inch nominal thickness
and may be of various species and grades. Glulam beams are
engineered wood products designed primarily to resist
bending. In horizontally laminated glulam, high-quality
laminations are placed on the tensile and compressive sides
of the beam, providing stiffness and strength where it is
needed most.
The advantages of glulam timbers are numerous. The
main drive behind glulam is that solid timbers are more
variable, limited in size, and increasingly difficult to
obtain. Wood properties are naturally variable due to
2
differences in density, grain orientation, species, and
growth features such as knots. Glulam lay-up allows for
these properties to be redistributed as smaller and
discontinuous defects. For example, a large knot that spans
the cross-section of a timber is reduced to a knot that is
no larger than the cross section of a single lamination.
Glulam can be manufactured to virtually any depth or
length, limited only by manufacturing press size.
Laminations can be manufactured to any length using finger-
joints, which allow two pieces of lumber to be joined end-
to-end. Glulam cross-sections vary from 2-1/2 x 6 in. to
10-3/4 x 81 in, while timber cross-sections range from 5 x 5
in. to 24 x 24 in. (AF&PA, 1991) . The glulam manufacturing
process also allows for production of curved, tapered, or
cambered members.
The advantages of glulam are clear. Because of the
advantages glulam offers, and the growing export market,
glulam production is expected to increase dramatically over
the next several years (APA, 1995) . The 1995 APA report
indicates total glulam production for all market segments
will grow from 280 million board feet in 1995 to more than
400 million board feet by the year 2000.
Clearly, any innovation will be beneficial if it
decreases costs, reduces fiber demand, and minimizes grade
requirements. The objectives of glulam reinforcement
include reduction of variability, improved strength and
stiffness, and reduced cost.
Glulam reinforcement facilitates size reduction as
compared to conventional glulam. Production of a glulam
member that is smaller in size and lighter in weight
decreases shipping costs, member dead weight, preservative
treatment costs, and the burden on our valuable timber
resources. Furthermore, decreased cost and size allows
glulam to become more competitive with steel and concrete.
Justification of Research
This research project involves evaluation of stress
distributions near the ends of reinforcement in glulam that
are partially reinforced. Partially reinforced glularn
refers to a glulam beam that is reinforced using a fiber-
reinforced plastic (FRP lamina) over less than its total
length. The reinforcement can be used as tensile
reinforcement when placed on the tensile side of a member or
as compressive reinforcement when placed on the compressive
side of a member. Reinforcement is placed either on the
outer surface of the beam (top or bottom), or it is
protected by one lamination of wood. A diagram of glulam
reinforced partially with tensile and compressive
reinforcement and having cover laminations is shown in
Figure 1.
3
Compressive Reinforcement
Length of Reinforcement
Bumper >
IBumper Tensile Reinforcement
Figure 1. Partially reinforced glulam beam with a cover(bumper) lamination.
5
Partially reinforced glulam has been approved by ICBO
Evaluation Service, Inc. and is utilized in several
structures including the Lighthouse Bridge in Port Angeles,
Washington (Gilham, 1995).
In composite materials, the proper evaluation of edge
and end effects is crucial to reliability in service. It is
known that in composite materials intralaminar ends can be
the location of significant stress concentrations.
Furthermore, the degree of anisotropy contributes to the
extent of localized stress effects influencing the magnitude
and dissipation of stress.
The FRP reinforcement and the wood in partially
reinforced glulam have significantly different stiffness
properties in the longitudinal direction. The intralaminar
terminus of the FRP lamina represents a potential site for
initiation of delamination. The structural safety and
reliability of partially reinforced glulam is compromised if
the stress distributions in this critical area is not fully
understood.
Objectives
The objectives of the research include experimental and
analytical stress analysis to identify the effects of FRP
lamina on stress distributions at the end of the FRP lamina.
Strain gage analysis is used to measure the strain
distribution near the end of the FRP lamina in seven full-
size glulam beams. The strain distribution from gages will
be compared to beam theory.
Finite-element analysis will be used to predict the
stress distribution at the end of the FRP lamina. A valid
finite-element model (FEM) requires proper characterization
of the structural components used in the model. In this
case, solid wood, adhesive and FRP lamina are characterized
by their respective material properties for use in the
finite-element models.
These methods will be used in concert to study the
stress distributions at the intralaminar FRP terminus. The
stresses at this location in the bending member are thought
to be influenced by several critical material and geometric
features. The objective of this thesis is to examine the
impact of potentially influential parameters - reinforcement
solid wood beam analogous to the location of the end of the
FRP lamina on the reinforced beam.
Element Mesh
The "standard" model type used for beams without a
bumper is shown in Figure 14. The "standard" model type
used for beams with a bumper is shown in Figure 15. One-
quarter of the beam was modeled in both cases since load,
dimensions and boundary conditions were symmetric about mid-
length and mid-width of the beam. In both types of models,
the region of the beam near the end of the FRP lamina was
finely meshed to provide a clear picture of the stress
distributions in this region and to reduce element aspect
ratio. In order to reduce model size and analysis time, the
mesh was scaled so that mesh density increases from top to
bottom, mid-width to edge, and from 2 in. either side of the
end of the FRP lamina to the end of the FRP lamina. A side
view of mesh distribution at the end of the FRP lamina is
shown in Figures 16 and 17 for beams without and with
bumpers respectively. The Mesh distribution across the width
is shown in Figure 18 and is identical for both model types.
Note that a two-inch gap is located at the end of the
FRP lamina in the models having a bumper. In a manufactured
beam, the gap length may be greater and the gap shape may be
more wedge shaped beyond the end of the FRP lamina.
D an e by rnrne r y
Hney Meshed eion
Fig-ure 14. The standard finite-element model for apartially reinforced glulam without a bumper having simplysupported end conditions.
55
ERF
L
F1:n
i mc e
r y Meshed
Figure 15. The standard finite-element model for apartially reinforced glulam with a bumper havina simplysupported end conditions.
56
!fl
:
1$L.' '-
?
J;V
:t
Imw
.,_______
---__-=
==
=
:;
11
_____:r:,-tv
!1- 1;--. .4-.-.t-
FR F
Figure 16. Side view of finite-element mesh at end of theFRP lamina for a beam without a bumper.
57
IH
1
I1-
I__
--
*
I__
____
_II
I4
III
s..
II.
WV
V
:4<
I
:4'J rn e
d oF E?
Figure 17. Side view of finite-element mesh at the end ofthe FRP lamina for a beam with a bumper.
58
rr;
ltf'c!
qb/,
P'" :r "w
'iit
'" "
,;,'I(
I4{I c'u$$:;
ci
-:'"j
i:
_______F
prL
ilt
:;i2i
,
F(
4
'!
pF
jL ";!
p
-'i :41.Y't4'
(1t4E
J;.._L'.:I..,-.
__________''
..'
4-
:t'I21
'Y1!J,1 :''
!,
.i *
*1e
4r
'S
'.
;I
St
'
4i
i:1
gs, (IT
1t4
I
If('1
/jl. (J(rI
'I'
31
'Ly,
is)
:t'y
59
Figure 18. Finite-element mesh for beams with and without abumper; mesh density increases near the FRP lamina end andtoward the edges.
60
Although the model does not produce the same shape found in
a manufactured beam, the effect of the gap shape opposite
the end of the FRP lamina likely has little effect on
stresses closer to the end of the FRP lamina.
Model Verification
The final model for a beam with a bumper and a beam
without a bumper were verified in several ways.
Verification was performed using solid wood beam models
where the same material properties are used as in the
standard model type. The verification was used to verify
element performance, boundary conditions and meshing. All
models represented a 5-1/8 x 12 in. x 21 ft beam loaded at
third-points. These results should indicate that boundary
conditions, material properties and element selection are
adequate.
Theoretical beam deflection was compared to model
deflection for solid wood beams and partially reinforced
beams. Derivation of theoretical beam deflection for a
composite beam composed of isotropic materials was given in
Chapter 1.
Table 7 compares centerline deflection and maximum
bending stress predicted by theory to values from the
finite-element model for 1) the model without a bumper and
without FRP lamina (the FRP lamina was made nonfunctional by
*Centerljne deflection.
Aspect ratios at the end of the FRP lamina were held
around 2:1; yet for the SOLID45 element, PNSYS® solutions
exhibited little effect due to much higher aspect ratios
(30:1). To ensure that effects of high aspect ratio in
portions of the mesh did not influence stress concentrations
near the end of the FRP lamina, a closed-form solution for
stress was compared to model stress for a partially
61
reducing E values to 10 psi), 2) the model without a bumper
with 0.14 in. of 16.6 x 106 psi MOE along 60% of the length,
3) the model with a bumper without FRP lamina (FRP lainina
assigned properties for wood), and 4) the model with a
bumper with 0.14 in. of 16.6 x 106 psi MOE along 60% of the
length.
Table 7. Closed-form and finite-element model values forcenterline deflection.
Beam Description FRP MOE
(106 psi)
FR?Length
(%)
- Load
(lb)
closed-
form 5CL
(in.)
Model
(in.)
5-1/8 x 12 in. x 21 ftNo BumperNo FR?
N/A N/A 14643 2.923 2.935
5-1/8 x 12 in. x 21 ftNo Bumper0.14 in. FR?
16.6 60 14643 2.36 2.38
5-1/8 x 12.14 in. x 21 ft1.5 in. BumperN0FRP
N/A N/A 20535 3.90 3.97
5-1/8 x 12.14 in. x 21 ft1.5 in. Bumper0.14 in. FR?
16.6 60 20535 3.514 3.57
62
reinforced beam without a bumper with an FRP lamina of
negligible stiffness (100 psi) . The maximum difference
between FEM and closed-form bending stress at the FRP
lamina/wood interface level, over a 2-in, length, centered
at the end of the FRP lamina, and half-width of the beam was
0.35%.
The agreement between closed-form solutions and FEM
results verified that boundary conditions were correct and
element performance and aspect ratios were adequate.
Finite-Element Model Results
Before comparing effects of various parameters, it is
important to have an understanding of the distribution and
magnitude of the stress components near the end of the FRP
lamina. It is also important to compare stresses predicted
by the FEM to mechanical properties for wood used in the
beam. Since material properties used in the FEM were for
Douglas-fir, mechanical properties of Douglas-fir were used
as a guideline as to what value of stress is considered to
be significant.
Table 8 shows values of mechanical properties of
interior west Douglas-fir from the Wood Handbook (FPJ.,
1987) . Note that properties given in the Table 8 are for
small, clear, straight-grained specimens at 12% moisture
content.
Table 8. Strength properties for small, clear, straight-grained samples interior west Douglas-fir Wood Handbook,1987).
63
Since the effect of FRP lamina on stress distributions
at the end of the FRP lamina is localized, and the FRP
lamina should be manufactured such that no defects occur at
the end of the FRP lamina, properties from small clear
samples are appropriate for comparison purposes.
For models with a bumper and for models without a
bumper, the distribution of each stress. was plotted for the
scenario that caused the greatest stress at the end of the
FRP lamina. These stress distributions were compared to
mechanical properties in Table 8. The stresses found to be
Most significant were studied further for all scenarios
investigated.
Since models built for glulam with a bumper and without
a bumper, results from these models will be dealt with
separately.
MoistureCondition
MaximumTensileStress
Parallel toGrain(psi)
CompressionPerpendicular
to Grain.
(psi)
MaximumShear
Parallel toGrain
(psi)
Maximum TensileStrength
Perpendicularto Grain
(psi)
> 30% 7700 420 940 290
12% 12600 760 1290 350
Methodology for Stress Distribution Characterization
Innumerable options exist for characterizing the stress
distributions at the end of the FRP lamina. Some options
include comparison of maximum stresses, comparison of
average.stress over an arbitrary distance in a given
direction at the end of the FRP lamina, comparison of
stresses at a consistent location relative to the end of the
FRP larnina or a coiribination of these or other options.
The basis for comparison chosen for this study was to
compare the average stress over a given distance near the
end of the FRP lamina at the outer fiber, or wood surface at
the plane of the FRP lamina-wood interface.
In order to calculate average stress values, a
polynomial curve fit was used to describe the stress
distribution at or near the end of the FRP lamina. The
best-fit polynomial for each stress distribution was
integrated over an arbitrary distance to calculate the
average stress. This type of approach is common in analysis
of stress distributions near small notches and holes in
laminated composites (Whitney and Nuismer, 1974) and peeling
stresses near free edges of laminated composites (Kim and
Soni, 1984). The reasoning behind calculating, an average
stress near discontinuities lies within a material's
inherent ability to redistribute the large stress
64
concentrations predicted from theory (Whitney and Nuismer,
1974)
Glulam Without a Bumper
All stress levels presented include the stress
distribution for the load applied to the model, which would
produce a 3000 psi bending stress in a solid wood beam at
the same location as the location of the end of the FRP
lamina in the reinforced beam.
For summary plots and tables, the ratio of stress
predicted in the model was compared to the 3000 psi bending
stress since all stresses, with the exceptionof , are
zero at the outer surface of the beam. Thus,, the "stress
ratio" was calculated by equation (11).
Stress Ratio.U 3000
Maximum Stress Levels and Stress Distribution
The six stress components a were plotted for the model
with an FRP lamina thickness of 0.35 in. (see Table 5 for
model details). The stress components for this model were
investigated first to find out which stresses were
significant, because the combination of variables used in
this model produced the greatest levels of stress at the FRP
65
66
lamina terminus. The maximum stress levels near the end of
the FRP lamina occurred at the wood-FRP lamina interface
plane; therefore, all stress distributions for models
discussed in this section are distributions at the wood-FRP
lamina interface plane and occurring on the wood lamina.
Figure 19 shows the stress distribution in the wood for
c in the neighborhood of the end of the FRP lamina along
the outside edge of the beam. The distribution clearly
shows a dramatic increase in , at the end of the FRP
lamina. In fact, a appears to be singular at this
location. The distribution of y, before the end of the FRP
lainina is fairly uniform in magnitude and lower than the
3000 psi stress that would occur if this were an
unreinforced beam. The rise in stress quickly decays past
the end of the FRP lamina and nearly approximates the
predicted 3000 psi level at 2 in. past the end of the FRP
lamina. The magnitude of o, at the end of the FRP lamina is
clearly greater than the maximum tensile strength for
Douglas-fir (Table 8). Therefore, was investigated for
all models over 0.05 and 0.15 in. distance just beyond the
end of the FRP lamina at the outside edge of the beam.
Figure 20 shows the distribution for a, at the end of
the FRP lamina, at the outside edge of the beam on the wood
surface above the FRP lamina. The distribution for,
3
25000
20000
15000
10000
5000
0.
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Location Relative to End of FR? Reinforcement (in.)
Figure 19. Distribution of a,, near the end of the FRPlamina on the wood at the outside edge of the beam forglulain without a bumper. The stress distribution is for0.35-in, thick FRP lamina, the FRP combination resulting inthe highest level of all stresses.
67
1600
1400
1200
I
800
600
40
1
0.5 1 1.5 2-0.5
200
Location Relative to End of FR? Reinforcement (in.)
68
Figure 20. Distribution of c near the end of the FRPlamina on the wood at the outside edge of the beam forglulam without a bumper. The stress distribution is for0.35-in, thick FRP lamina, the FRP combination resulting inthe highest level of all stresses.
69
indicates that the solution is singular at the end of the
FRP lamina. The magnitude of cs rapidly decreases to zero
over a 0.75 in. distance beyond the end of the FRP lamina.
The magnitude of , at the end of the FRP lamina appears to
be greater than the maximum tensile strength for Douglas-fir
perpendicular to grain(Table 8). Therefore,,was
investigated in all models at the outside edge of the beam,
over a 0.05 and 0.15 in. distance up to the end of the FRP
lamina. Stress cy, was integrated prior to the end of the
FRP lamina for two reasons. First of all,there appears to
be a greater average stress in this region than beyond the
end of the FRP lamina. Secondly, may be an important
factor leading to the FRP lamina peeling off of the beam.
Figure 21 shows the distribution for at the end of
the FRP lamina at mid-width of the beam. The distribution
and magnitude of are very similar to that of,however
the stress appears to be distributed along a greater length
of the beam. Stress a approaches zero at approximately 1.5
in. beyond the end of the FRP lamina. Since the maximum
value of c may be greater than the maximum tensile strength
of Douglas-fir perpendicular to grain, a was also
investigated for all models. The a stress was be averaged
over a 0.05 and 0.15 in. distance beyond the end of the FRP
lamina at mid-width of the beam; the stress was distributed
1000
900
800
700
600
500
400
U'-1.5 -1 -0.5 p
200
Location Relative to End of FR? Reinforcement (in.)
Figure 21. Distribution of near the end of the FRPlamina on the wood at the mid-with of the beam for glulamwithout a bumper. The stress distribution is for O.35-in.thick FRP lamina, the FRP combination resulting in thehighest level of all stresses.
70
0.5 1 1.5
71
over a greater length of the beam beyond the end of the FRP
lamina, and it was highest at mid-width of the beam.
Figure 22 shows the distribution for o at the end of
the FRP lamina in the wood at the outside edge of the beam.
The distribution of .shows much greater levels of stress
over the 2 in. prior to the end of the FRP lamina than over
the 2 in. beyond the end of the FRP lamina. The
distribution also appears to be singular at the end of the
FR? lamina. The maximum c, is greater than the maximum
shear parallel to grain for Douglas-fir (Table 8). Hence,
, will be averaged dyer a 0.05 and 0.15 in. distance prior
to the end of the FR? lamina at the outside edge of the
beam.
Figure 23 shows the distribution for at the end of
the FRP lamina at the Outside edge of the beam. The
magnitude of o is much smaller than the maximum shear
strength parallel to grain (Table 8), which is much lower
than the maximum shear strength perpendicular to grain. For
these reasons, was not investigated further.
Figure 24 shows the distribution for at the end of
the FRP lamina at the outside edge of the beam. Although
the magnitude of o is close to the maximum shear parallel
to grain for Douglas-fir (Table 8), it is very localized,
and lower in magnitude than . Since,and are both
72
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Location Relative to End of FRP Reinforcement (in.)
Figure 22. Distribution of near the end of the FRPlamina on the wood at the outside edge of the beam forglulam without a bumper. The stress distribution is for0.35-in, thick FRP lamina, the FRP combination resulting inthe highest level of all stresses.
Location Relative to End of FRP Reinforcement (in.)
Figure 23. Distribution of a near the end of the FRPlamina on the wood at the outside edge of the beam forglulam without a bumper. The stress distribution is for0.35-in, thick FRP lamina, the FRP combination resulting inthe highest level of all stresses.
73
I
7
-1.5
-100
-200
-300
-400
-500
-600
-700
00
0.5 1 1.5 2
Location Relative to End of FR? Reinforcement (in.)
Figure 24. Distribution of near the end of the FRPlaiaina on the wood at the outside edge of the beam forglulam without a bumper. The stress distribution is for0.35-in, thick FRP lamina, the FRP combination resulting inthe highest level of all stresses.
74
75
shear parallel to grain and c, is much greater than, and
less isolated than , only a, will be investigated.
The distributions of y, cs,, and across the beam
width and along the length are shown in Figures 25, 26, 27
and 28 respectively. In these figures, the x and y-axes are
not to scale. Figure 25 shows that c is distributed
uniformly across the width. Figure 26 shows that is
uniformly distributed across the width, however, it
decreases slightly toward the outside edge. Figure 27 shows
that a decreases at the outside edge by about one-third the
level at mid-width. Figure 28 shows that , is distributed
uniformly across the width, while decreasing slightly at the
outside edge.
Effect of Reinforcement Thickness
All beams in these models were 5-1/8 x 12 in. x 21 ft
and had FRP laniina over 60% of the length on the tensile
side. The models were subjected to third-point loads. The
FRP lamina thicknesses are given in Table 5
The stress distributions for c, o, c and are
plotted in Figures 29 to 32.
Polynomial curve fits were used to average stress
levels over a 0.05 and 0.15 in. distance prior to the end of
Location Relative to End of FRPReinforcement (in.)
If) DistanceC . From Center
of BeamWidth (in.)
Figure 25. Typical distribution of along the beam lengthand across beam width near the end of the FRP lamina forbeams without a bumper. Location relative to the end of theFRP lamina is not to scale.
76
16000
14000
12000
10000
U)8000
U)w 0000
4000
2000
2.4456
1. 9152
900
800
700
600
500
400
300200100
0
-100-200
2.4456
C 1.9152C', C'
C C (N m. N U-
C C C0
Location Relative to End of FRP Distance
Reinforcement (in.) From Centerof Beam
Width (in.)
Figure 26. Typical distribution of,along the beam length
and across beam width near the end of the FRP lamina forbeams without a bumper. Location relative to the end of theFRP lamina is not to scale.
77
500
400
300
200
100C')
0
-100(N
(N
o
'?Location Relative to End FRP
Reinforcement (in.)
LI) '
o 0Distance From Centerof Beam Width (in.)
78
Figure 27. Typical distribution of a along the beam lengthand across beam width near the end of the FRP lamina forbeams without a bumper. Location relative to the end of theFRP lamina is not to scale.
IC)
'.D C\J
C\J
c
'.0
'-4Location Relative to End of FR? Distance FromReinforcement (in.) Center of Beam
Width (in.)
Figure 28. Typical distribution of along the beamlength and across beam width near the end of the FRP laminafor beams without a bumper. Location relative to the end ofthe FRP lamina is not to scale.
79
--.-1
1400
1200
1000
800
a)
600
4JCl) 400
200
0
30000
25000
20000 -
-IU,
, 15000 -CI,
4)Cl)
10000 -
0
10.07' Thick FR?
0.14" Thick FR?5000 - A0.21" Thick FR?
0.27" Thick FR?
-0.35" Thick FR?
Location Relative to End of FRP Reinforcement (in.)
Figure 29. The distribution of is plotted for variousFRP thickness on a beam without a bumper. Best fitpolynomials are plotted through the finite-element data.
80
0.1 0.2 0.3 0.40 0.5 06
0
Location Relative to End of FR? Reinforcement (in.)
Figure 30. The distribution of c is plotted for variousFRP thickness on a beam without a bumper. Best fitpolynomials are plotted through the finite-element data.
81
900
800 -
700
600
300 -
200 -
100 -
0
0 0.05 0.1 0.15 0.2
0.07" Thick FR?
0.14' Thick FR?
£0.21" Thick 'FR?
0.27" Thick FR?
-0.35" Thick FR?
82
0.25 0 3
Location Relative to End of FRP Reinforcement (in.)
Figure 31. The distribution of a is plotted for variousFRP thickness on a beam without a bumper. Best fitpolynomials are plotted through the finite-element data.
U)Cl)
a)
-1.)
U)
-0.35 in. FR?0.28 in. FR?A0.21 in. FR?£0.14 in. FR?10.07 in. FR?
500
500 -
0
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 32. The distribution of a, is plotted for variousFRP thickness on a beam without a bumper. Best fitpolynomials are plotted through the finite-element data.
83
84
the FRP lamina ( and c) or beyond the end of the FRP
lamina ( and az). Coefficients for the polynomial fits
.are given in the Appendix, Table Al. The best fit
polynomials are plotted through the FEM data in Figures 29
to 32.
The effect of FRP lamina thickness on , c and
are given in Table 9 and shown graphically in Figures 33 to
36. It is apparent that FRP lamina thickness does have an
effect on stress distributions at the end of the FRP lamina.
The stresses all increase with increasing FRP lamina
thickness, but approach a common maximum value with greater
FRP lamina thickness. The distance between the 0.05 and
0.15 in. averaged stress levels is an indication of stress
localization. For example, the magnitude of c,, Figure 34,
decreases by half when averaged over 0.15 in. compared to
0.05 in.; this indicates that is very localized. The
magnitude of o (Figure 36), on the other hand, is very
close for the 0.05 and 0.15 in. averages; this indicates
that the a is not very localized.
Effect of Stiffness Ratio
All beams represented in these models were 5-1/8 x 12
in. x 21 ft and had a 0.14-in, thick FRP lamina over 60% of
Table 9. Stresses and stress ratios at the end of the FRPlamina for various FRP thicknesses for glulam without abumper.
Figure 33. The effect of the FRP lamina thickness on cYnear the end of the FRP lamina is plotted for average stressfrom the end of the FRP lamina to 0.05 and 0.15 in. past theend of the FRP lamina.
Figure 34. The effect of the FRP lamina thickness onnear the end of the FRP lamina is plotted for average stressfrom the end of the FRP lamina to 0.05 and 0.15 in. prior tothe end of the FRP lamina.
UO.05 in. Avg--O.15 in. Avg
700
- 600
_____. 500
-HU)
- 400
U)U)a,
- 300Cl)
- 200
- 100
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Reinforcement Thickness (in.)
Figure 35. The effect of the FRP lamina thickness On cYz
near the end of the FRP lamina is plotted for average stressfrom the end of the FRP lamjna to 0.05 and 0.15 in. past theend of the FRP lamina.
88
0.25
0.20 -
0.15-
U)U)a)
- 0.10-CI)
0.05-
0.00
-U--O.05 in. Avg-+-O.15 in. Avg
2000
I 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Reinforcement Thickness (in.)
Figure 36. The effect of the FRP lamina thickness onnear the end of the FRP lamina is plotted for average stressfrom the end of the FRP lamina to 0.05 and 0.15 in. prior tothe end of the FRP lamina.
- 1500
- 500
89
0.7
0.6
0.5
0--I
0.4-
U)U)a)
4-IC/D
0.2-
0.1 -
0.0
90
the length. The models were subjected to third-point loads.
Stiffness ratios were given in Table 5.
The stress distributions at the end of the FRP lamina
for various stiffness ratios are shown in Figures 37 to 40
for cy, cy, a and a, respectively. It is apparent that
stiffness ratio has a significant effect on these stress
distributions.
Polynomial curve fits were used to average stress
levels over a 0.05 and 0.15 in. distance prior to the end of
the FR? lamina ( and a) or beyond the end of the FRP
lamina (and cj. Coefficients for the polynomial fits
are given in the Appendix, Table A2. The best fit
polynomials are plotted through the FEM data in Figures 37
to 40.
The effects of stiffness ratio on c, c,, and a are
given in Table 10 and shown graphically in Figures 41 to 44.
The stresses all increase with increasing stiffness ratio.
While , a and cy, increase significantly with
increased stiffness ratio, y, increases much less; this may
be a result of having E as a constant for all FR?
applications.
25000
20000
:: 15000 -C,)
C!)
Cl)
ci)
- 10000 -(ID
5000 - -Stiffness Ratio = 4.0
AStiffness Ratio = 5.88
Stiffness Ratio = 8.3
Stiffness Ratio 10
0
91
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FRP Reinforcement (in.)
Figure 37. The distribution of is plotted for variousFRP-to-wood stiffness ratios on a beam without a bumper.Best fit polynomials are plotted through the finite-elementdata.
-o
-Stiffness Ratio = 4.0
AStiffness Ratio 5.88
Stiffness Ratio 8.3
Stiffness Ratio = 10
25 -0.2 -0.15 -0.1 -0.05
Location Relative to End of FRP Reinforcement (in.)
Figure 38. The distribution of a, is plotted for variousFRP-to-wood stiffness ratios on a beam without a bumper.Best fit polynomials are plotted through the finite-elementdata.
900
80
00 -
500 -
400 -
300 -
200 -
100 -
92
700
600
500
--I
a 400 -
-
-
-
A
-Stiffness Ratio = 4.0
AStiffness Ratio 5.88
Stiffness Ratio = 8.3
RStiffness Ratio = 10.0
0.25
Location Relative to End of FRP Reinforcement (in.)
Figure 39. The distribution of is plotted for variousFRP-to-wood stiffness ratios on a beam without a bumper.Best fit polynomials are plotted through the finite-elementdata.
03
93
Cl)
300-lJCr
200
100
0
0 0.05 0.1 0.15 0.2
-Stiffness Ratio =
£Stiffness Ratio
Stiffness Ratio
UStiffness Ratio =
2000 -
1800
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 40. The distribution of is plotted for variousFRP-to-wood stiffness ratios on a beam without a bumper.Best fit polynomials are plotted through the finite-elementdata.
94
800
600
4.0 4005.88
8.3
10200
95
Table 10. Stresses and stress ratios at the end of the FRPlamina for various stiffness ratios for glulam without abumper.
0.05 in.Average
0.15 in.Average
0.05 in.Average
0.15 in.Average
StiffnessRatio
Stress atApplied Load
(psi)
Stress atApplied Load
(psi)
Stress Ratio(Equation (11))
Stress Ratio(Equation (11))
4.0 14120 11090 4.71 3.70
5.8 16780 13070 5.59 4.36
8.3 19640 15190 6.55 5.0610 21230 16370 7.08 5.46
4.0 496 237 0.165 0.0795.8 518 256 0.173 0.085
8.3 537 279 0.179 0.093
10 537 283 0.179 0.095
4.0 369 253 0.123 0.084
5.8 432 304 0.144 0.1018.3 497 357 0.166 0.119
10 530 386 0.177 0.129
ocy
4.0 885 801 0.295 0.267
5.8 1155 1074 0.385 0.358
8.3 1420 1280 0.473 0.427
10 1560 1376 0.520 0.459
8
7-
- 20000
- 15000
96
U)4-
U)U) U)U)
100004JCl) 3- 4-,
U)
2-- 5000
1- UO.05 in. Avg----O.15 in. Avg
0 0
0 2 4 6 8 10
Stiffness Ratio, FRP Reinforcinent-to-Wood
Figure 41. The effect of the FRP-to--wood stiffness ratio on
near the end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and 0.15 in.past the end of the FRP lamina.
0.200 -
0.180 -
0.160 -
0.140-
Stiffness Ratio, FRP Reinforcement-to-Wood
Figure 42. The effect of the FRP-to-wood stiffness ratio onnear the end of the FRP lamina is plotted for average
stress from the end of the FRP lamina to 0.05 and 0.15 in.prior to the end of the FRP lamina.
- 400.2 0.120-
0100
- 0.080-crJ 0
97
0.060 - - 200
0.040 -- 100
0.020 UO.05 in. Avg4--O.].5 in. Avg
600
S
- 500U-
0.000 - 00 2 4 6 8 10
0.200
0.180 -
0.160-
0.140 -
- 0.120-(U
0.100-U)a)
0.080-U)
0.060 -
0.040 -
0.020 -
0.000
O.05 in. Avg--O.15 in. Avg
600
- 500
- 400
U)
- 300U)a)
U)
- 200
- 100
0
98
0 2 4 6 8 10
Stiffness Ratio, FRP Reinforcement-to-Wood
Figure 43. The effect of the FRP-to-wood stiffness ratio ono near the end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and 0.15 in.past the end of the FRP lamina.
0.6
0.5
0.4 -
04J
0.3-
a,
U)
0.2-
0.1---O.05 in. Avg---O.15 in. Avg
0 2 4 6 8 10
Stiffness Ratio, FRP Reinforcement-to-Wood
Figure 44. The effect of the FRP-to-wood stiffness ratio onnear the end of the FRP lamina is plotted for average
stress from the end of the FRP lamina to 0.05 and 0.15 in.prior to the end of the FRP lamina.
1800
- 600
400
- 200
99
0.0 0
Effect of Reinforcement Length
All beams represented in these models were 21 ft long,
5-1/8 in. wide, 12 in. high and had 0.14 in. thick FRP
lamina. The models were subjected to third-point loads.
The FRP laniina lengths are given in Table 5.
Figures 45 to 48 show the stress distributions cY, cY,
and o for various lengths of FRP lamina (percent of
total span length). It is apparent from these graphs that
the length of the FRP lamina does not have a major impact on
stress distributions. However, it appears that for the 80%
reinforced beam length, stress levels are increased
slightly. Sine the other FRP lamina lengths are
approximately the same and different from the 80% reinforced
length, it is reasonable to justify that the effect of the
support 2 ft away was influential.
Effect of Beam Width
All beams represented in these models were Width x12
in. x 21 ft and had a 0.14-in, thick FRP lamina over 60% of
the length. The models were subjected to third-point loads.
The beam widths are given in Table 5.
Figures 49 to 52 show the stress distributions , c,
a and a, for beam widths ranging from 3.125 in. to 10.75
100
-
-
-
-
101
25000
20000
15000U)
U)U)
U
10000Cl)
5000
RFRP Length 40%
FRP Length = 50%AFR? Length = 60%FR? Length = 80%
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FRP Reinforcement (in.)
Figure 45. The distribution of is plotted for variousFRP lamina lengths on a beam without a bumper.
Location Relative to End of FR? Reinforcement (in.)
Figure 46. The distribution of is plotted for variousFRP lamina lengths on a beam without a bumper.
400 -
200 -
00
102
.2 -0.15 -0 . 1 -0.05
700
200 -
100 -
0
RFR? Length 40%
--FR? Length = 50%
A--FR? Length = 60%
---FR? Length = 80%
103
0 0.05 0.1 0.15 0.2 0.25 0 3
Location Relative to End of FRP Reinforcement (in.)
Figure 4.7. The distribution of c is plQtted for variousFRP lamina lengths on a beam without a bumper.
1800
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 48. The distribution of c, is plotted for variousFRP lamina lengths on a beam without a bumper.
104
1200
1000
800
600
--FRP+FRP
Length = 40% 400Length 50%
+FRP Length = 60%.--FRP Length = 80% 200
0
00 -_---- 2'
-
-U--3.125 in. Wide--5.125 in. WideA--6.75 in. Wide10.75 in. Wide
105
25000
20000
15000U)
U)U)CI,
10000CI)
5000
0
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FR? Reinforcement (in.)
Figure 49. The distribution of o is plotted for variousbeam widths on a beam without a bumper.
-o
-3.125 in. Wide-5.125 in. WideA-6.75 in. Wide10.75 in. Wide
-0.15 -0.1 -0.05
Location Relative to End of FRP Reinforcement (in.)
Figure 50. The distribution of c is plotted for variousbeam widths on a beam without a bumper.
oo
0
0
106
25 -0.2
700
500
200 -
0
Location Relative to End of FRP Reinforcement (in.)
Figure 51. The distribution of cY is plotted for variousbeam widths on a beam without a bumper.
107
U-3.125 in. Wide-----5.125 in. Wide
*---6.75 in. Wide
10.75 in. Wide100
0 0.05 0.1 0.15 0.2 0.25 0 3
-3.125 in. Wide---5.125 in. Wide
A---6.75 in. Wide10.75 in. Wide
160
00 -
1200 -
1000 -
800 -
600 -
400 -
200 -
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 52. The distribution of a, is plotted for variousbeam widths on a beam without a bumper.
108
109
in. It is apparent from these graphs that beam width has
very little influence on stress distributions at the end of
the FRP lamina.
Effect of Span-to-depth Ratio
All beams represented in these models had a 5-1/8 x 12
in. cross section and had a 0.14-in, thick FRP lamina over
60% of the length. The models were subjected to third-point
loads. The span-to-depth ratios are given in Table 5.
Figures 53 to 56 show the stress distributions , cY,,
a and a for span-to-depth ratios ranging from 10 to 27.
It is apparent from these figures that span-to-depth ratio
has very little influence on stress distribution at the end
the of FRP lamina. However, the stress levels tend to be
higher for the span-to-depth ratio of 10. It is likely that
the higher value for this ratio is due to the proximity of
the end support, which was 2 ft from the end of the FRP
lamina.
Effect of Loading Conditions
All beams represented in these models were 5-1/8 x 12
in. x 21 ft and had a 0.14-in, thick FRP lamina over 60% of
the length.
110
25000
USpan:Depth = 10:1
4--Span:Depth = 15.5:1
20000ASpan:Depth = 21:1
Span:Depth = 27:1
- 15000 -
4J 10000 -U)
5000 -
0
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FR? Reinforcement (in.)
Figure 53. The distribution of a, is plotted for variousspan-to-depth ratios on a beam without a bumper.
USpan:Depth 10:1
$--Span:Depth = 15.5:1
A----Span:Depth = 21:1
Span:Depth = 27:1
1000
600 -
400 -
200 -
111
-0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 54. The distribution of o, is plotted for'varjousspan-to-depth ratios on a beam without a bumper.
700
200
100 -
0
-Span:Depth 10:1
$.--Span:Depth = 15.5:1
ASpan:Depth = 21:1
Span:Depth = 27:1
0 0.05 0.1 0.15 0.2 0.25 0 3
Location Relative to End of FRP Reinforcement (in.)
Figure 55. The distribution of is plotted for variousspan-to-depth ratios on a beam without a bumper.
112
RSpan:Depth = 10:1
4.--Span:Depth = 15.5:1
Span:Depth 21:1
.Span:Depth = 27:1
2000
1800U
00 -
U 1200 -
1000 -
800 -
600 -
400 -
200 -
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FR? Reinforcement (in.,)
Figure 56. The distribution of a is plotted for variousspan-to-depth ratios on a beam without a bumper.
113
114
Figures 57 to 60 show the stress distributions cJ, ,
a and o for third-point and uniform loading conditions.
These figures show that uniform loading results in virtually
identical stress levels as third-point loading.
Glulam With a Bumper
All stress levels presented include the stress
distribution for the load applied to the model. This load
would produce a 3000 psi bending stress in a solid wood beam
at the same location as the location of the end of the FR?
lamina, at the upper surface of the FR? lamina, in the
reinforced beam.
For summary plots and tables, the ratio of stress
predicted in the model were compared to the 3000 psi bending
stress since all stresses, with the exception of a, are
zero at the outer surface of the beam. The stress ratio is
calculated by equation (10).
Since FR? lamina thickness and stiffness ratio were the
only variables found to be influential on stress levels for
glulam without a bumper, it is reasonable to assume the same
holds true for glulam with a bumper. However, to ensure
that this is the case, two models will be compared for the
effects of width and span-to-depth ratio.
25000
20000
15000 -0)
U)U,a,
10000 -
5000 -
0
UUniform LoadingThird-Point Loads
115
0.1 0.2 0.3 0.4 0.5 0.6
Location Relative to End of FRP Reinforcement (in.)
Figure 57. The distribution of is plotted for uniformand third-point loading on a beam without a bumper.
-o
-Uniform LoadsThird-Point Loads
25 -0.2 -0.15 -0.1 -0.05
Location Relative to End of FRP Reinforcement (in.)
Figure 58. The distribution of cy, is plotted for uniformand third-point loading on a beam without a bumper.
0o
116
700
600 A
500 -
400 -
(I)U)
300 -4)Cl)
200 -
100 -
0
0
AUniform Loading
--- Third-Point Loads
0.05 0.1 0.15 0.2
Location Relative to End of FRP Reinforcement (in.)
Figure 59. The distribution of a7 is plotted for uniformand third-point loading on a beam without a bumper.
0.25 0 3
117
-Uniform LoadingThird-Point Loads
1200 -
1000 -
800 -
600 -
400 -
200 -
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 60. The distribution of o is plotted for uniformand third-point loading on a beam without a bumper.
118
Maximum Stress Levels and Stress Distribution
Figures 61 to 66 show stress distributions for a.j at
the end of the FRP lamina at the level of the upper wood/FRP
lamina interface. All stress distributions were taken at
the outside edge of the beam, with the exception of ,
which was taken at mid-width of the beam.
Figure 61 shows the distribution of near the end of
the FRP lamina. This distribution is very similar to the
distribution for the analogous beam without a bumper (Figure
19); it is singular at the end of the FRP lamina and levels
off to the predicted bending stress in a solid wood section
at about 2 in. beyond the end of the FRP lamina. The rise
in stress at the node located at about 1.7 in. is due to the
gap at 2 in. past the end of the FRP lamina (Figure 17).
The magnitude is greater than the maximum tensile strength
for Douglas-fir (Table 8); therefore a will be investigated
over a range beyond the endof the FRP lamina at the outside
edge of the beam.
Figure 62 shows the distribution of,near the end of
the iRP lamina. a is singular and compressive just before
the end of the FRP lamina and is tensile beyond the end of
the FR? lamina. Since compressive stress is unlikely to
cause failure and the tensile stress is less than the
119
3
30000
25000
20000
15000
10000
5000
Location Relative to End of FRP Reinforcement (in.)
Figure 61. Distribution of near the end of the FRPlamina on the wood at the outside edge of the beam forglulam without a bumper. The stress distribution is for0.35-in, thick FRP lamina, the FRP combination resulting inthe highest level of all stresses.
120
2-2 -1.5 -1 -0.5 0 0.5 1 1.5
a
200
100
I 0
-1.5 -1 -0.5 0.5 1 1
Location Relative to End of FRP Reinforcement (in.)
Figure 62. Distribution of a, near the end of the FRPlairiina on the wood at the outside edge of the beam forglulam with a bumper. The stress distribution is for 0.35-in. thick FRP lamina, the FRP combination resulting in thehighest level of all stresses.
121
Location Relative to End of FRP Reinforcement (in.)
Figure 63. Distribution of near the end of the FRPlamina on the wood at mid-width of the beam for glulam witha bumper. The stress distribution is for 0.35-in, thickFRP lamina, the FRP combination resulting in the highestlevel of all stresses.
122
500
400
300
200
100
a
2 -1.5 -1 -e. 0.5 1 1.5
-0
-20
-300
123
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Location Relative to End of FR? Reinforcement (in.)
Figure 64. Distribution of c. near the end of the FRPlamina on the wood at the outside edge of the beam forglulam with a bumper. The stress distribution is for 0.35-in. thick FRP lamina, the FRP combination resulting in thehighest level of all stresses.
Location Relative to End of FRP Reinforcement (in.)
124
Figure 65. Distribution of o near the end of the FRPlaraina on the wood at the outside edge of the beam forglulam with a bumper. The stress distribution is for 0.35-in. thick FRP lamina, the FRP combination resulting in thehighest level of all stresses.
200
-1.5 -1 -0.5
-200
-400
-600
-800
-1000
1200
0.5 1 1.5 2
Location Relative to End of FRP Reinforcement (in.)
Figure 66. Distribution of near the end of the FRPlamina on the wood at the outside edge of the beam forglularn with a bumper. The stress distribution is for 0.35-in. thick FRP lamina, the FRP combination resulting in thehighest level of all stresses.
125
126
maximum tensile strength perpendicular to grain for Douglas-
fir (Table 8), o. will not be investigated further.
Figure 63 shows the distribution of a near the end of -
the FRP lamina at mid-width of the beam. The stress a is
compressive prior to the end of the FRP lamina and tensile
beyond the end of the FRP lamina. The tensile stress beyond
FRP lamina is greater than the maximum tensile strength of
Douglas-fir perpendicular to grain. Thus, will be
investigated for all models and averaged over a 0.05 and
0.15 in. distance beyond the end of the FRP lamina.
Figure 64 shows the distribution of near the end of
the FRP lamina at mid-width of the beam. The distribution
of is very similar to that in the analogous beam without
a bumper (Figure 20). Maximum values for are greater
than values for shear strength parallel to grain for
Douglas-fir (Table 8) . Thus will be investigated for
all models and averaged over a 0.05 and 0.15 in. distance
prior to the end of the FRP lamina.
Figure 65 shows the distribution of a near the end of
the FRP lamina at the outside edge of the beam. The
magnitude of is much less than the shear strength of
Douglas-fir parallel to grain (Table 8), which is much less
127
than shear strength perpendicular to grain, therefore c
will not be considered further.
Figure 66 shows the distribution of near the end of
the FRP lamina at the outside edge of the beam. The
magnitude and area covered by xz is much less than a,,
therefore, will not be considered further.
Figures 67 to 69 show three-dimensional plots of ,
and along the length of the beam near the end of the FRP
lamina and across the half-width of the beam. Distributions
are shown for the surface above the FRP lamina and the
surface below FRP lamina.
It is apparent in Figure 67 that a< is greater along
the surface above the FRP lamina than below the FRP lamina.
This is due to the bumper bending about the gap, reducing
the stress on the upper portion of the bumper near the end
of the FRP lamina. Figure 67 also shows that the maximum
occurs at the outside edge of the beam.
Figure 68 shows that a is tensile and greater in
magnitude on the upper surface than on the lower surface.
Figure 69 shows that is nearly two times greater on
the surface above the FRP lamina than on the surface below
the FRP lamina.
0
16000
14000
12000
10000
8000
6000
4000
2000
0
I Distance FromLocation Relative to Center of
End of FRP Beam WidthReinforcement (in.) (in.)
Figure 67. Typical distribution of along the beam lengthand across beam width near the end of the FRP lamina forbeams with a bumper.
128
300
250
200
150
Cl) 100
50Ii)Cl)
U)-50
-100
-150
-200
-250
Location Relative to End ofFR? Reinforcement (in.)
N
U)
Distance FromCenter of BeamWidth (in.)
129
Figure 68. Typical distribution of o along the beam lengthand across beam width near the end of the FRP lamina forbeams with a bumper.
FR?1200
1000
800
600
400
200
0
-200
-400
-600
Location Relative to Endof FRP Reinforcement (in.)
CC
('.1
C
'-4
C
DistanceFrom Center
of BeamWidth (in.)
130
Fig-ure 69. Typical distribution of, along the beam
length and across beam width near the end of the FRP laminafor beams with a bumper.
Effect of Reinforcement Thickness
All beams represented in these models were 5-1/8 x 12
(+ FRP lamina thickness) in. x 21 ft and had FRP lamina over
60% of the length. The models were subjected to third-point
loads. The FRP lamina thicknesses are given in Table 6.
Polynomial curve fits were used to average stress
levels over a 0.05 and 0.15 in. distance prior to the end of
the FRP lamina for cy,, and beyond the end of the FRP lamina
for and . Coefficients for the polynomial fits are
given in the Appendix, Table A3.
Figures 70 through 72 show the effect of FRP lamina
thickness on c, c and c,. It is apparent from these
figures that thickness does influence the stress
distribution at the end of the FRP lamina.
Polynomial curve fits were used to average stress
levels over a 0.05 and 0.15 in. distance prior to the end of
the FRP lamina for and beyond the end of the FRP lamina
for and . Coefficients for the polynomial fits are
given in the Appendix, Table A3. The best fit polynomials
are plotted over the data points on Figures 70 through 72.
The effect of FRP lamina thickness on stress
distribution are given in Table 11 and shown on Figures 73
to 75 for a, c and a,. The magnitudes of all stress
131
35000
30000 -
25000 -
a 20000 -
15000 -4J
10000 -
5000
0
0 0.1 0.2 0.3 0.4
0.07 in. FR?-0.14 in. FRPA0.21 in. FR?0.28 in. FR?0.35 in. FR?
0.5 0 6
Location Relative to End of FR? Reinforcement (in.)
Figure 70. The distribution of a, is plotted for variousFR? thickness on a beam with a bumper. Best fit polynomialsare plotted through the finite-element data.
132
600
500 -
400 -
200 -
100 -
0
0.07 in. FR?-0.14 in. FR?£0.21 in. FR?.0.28 in. FR?NO.35 in. FR?
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 4
Location Relative to End of FRP Reinforcement (in.)
Figure 71. The distribution of o is plotted for variousFRP thickness on a beam with a bumper. Best fit polynomialsare plotted through the finite-element data.
133
0.07 in. FR?-0.14 in. FR?
£0.21 in. FR?0.28 in. FR?10.35 in. FR?
3-0-00
500 -
0
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 72. The distribution of is plotted for variousFRP thickness on a beam with a bumper. Best fit polynomialsare plotted through the finite-element data.
134
Table 11. Stresses and stress ratios at the end of the FR?lamina for various FR? thicknesses for glulam with a bumper.
Figure 73. The effect of the FRP lamina thickness on ynear the end of the FRP lainina is plotted for average stressfrom the end of the FRP lamina to 0.05 and 0.15 in. past theend of the FRP lamina.
136
10 30000
9--
---U
-U-- 25000
8-
7-
5- -__----_-
__.-_____-15000
,
a)
-4- ----- 4-)Cl)
- 100003--
2-- 5000
1 - -U-O.05 in. Avg-4-0.15 in. Avg
0 0I
0.20
0.18 -
0.16-
0.14 -
- 0.12-(j
Cl)0.10-
O3
0.08-Cl)
0.06-
0.04 -- 100
0.02 -- -O.05 in. Avg+--O.15 in. Avg
0.00 t 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Reinforcement Thickness (in.)
Figure 74. The effect of the FRP lamina thickness on cnear the end of the FRP lamina is plotted for average stressfrom the end of the FRP lamina to 0.05 and 0.15 in. past theend of the FRP lamina.
600
-500
137
- 400
- 300Cl)
U)IV
.lJCl)
- 200
0.9
.25000.8-
- -0.7-
- -2000
0.6-
- n c_ -1500 a
138
Figure 75. The effect of the FRP laminá thickness onnear the end of the FRP lamina is plotted for average stressfrom the end of the FRP lamina to 0.05 and 0.15 in. prior tothe end of the FRP lamina.
0.2-
0.1
0.0
- SO.05 in. Avg---O.15 in. Avg
i
- 500
0:
0 0.05 0.1 0.15 0.2 0.25
Reinforcement Thickness (in.)
0.3 0.35
distributions increase significantly with FRP lamina
thickness.
Effect of Stiffness Ratio
All beams represented in these models were 5-1/8 x
12.14 in. x 21 ft and had FRP lamina over 60% of the length.
The models were subjected to third-point loads. The
stiffness ratios are given in Table 6.
Figures 76 to 78 show the stress distributions for ,
a and near the end of the FRP lamina for various
stiffness ratios. It is apparent from these plots that
stiffness ratio does have a significant effect on stress
distributions at the end of the FRP lamina.
Polynomial curve fits were used to average stress
levels over a 0.05 and 0.15 in. distance prior to the end of
the FRP lamina (, and y,) or 0.05 and 0.15 in. beyond the
end of the FRP lamina ( and cy). Coefficients for the
polynomial fits are given in the Appendix, Table A4. The
best fit polynomials are plotted over the FEM data on
Figures 76 through 78.
The effects of stiffness ratio on stress distribution
are given in Table 12 and shown on Figures 79 to 81 for ,
139
-
-
140
I
25000
20000
liD
15000
U)0)a)
- 10000 -C/)
5000 -
0
AStiffness Ratio = 4.0
Stiffness Ratio 5.8
-Stiffness Ratio = 8.3
Stiffness Ratio = 10.0
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FRP Reinforcement (in.)
Figure 76. The distribution of c is plotted for variousFRP-to-wood stiffness ratios on a beam with a bumper. Bestfit polynomials are plotted through the finite-element data.
400
350 -
300
250--'-4
200--
4.4
150-
100 -
50 -
0
-Stiffness Ratio 4.0
AStiffness Ratio = 5.8
Stiffness Ratio 8.3
UStiffness Ratio = 10.0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 4
Location Relative to End of FRP Reinforcement (in.)
Figure 77. The distribution of a is plotted for variousFRP-to-wood stiffness ratios on a beam with a bumper. Bestfit polynomials are plotted through the finite-element data.
141
-Stiffness Ratio = 4.0
AStiffness Ratio = 5.8
Stiffness Ratio 8.3
Stiffness Ratio = 10.0
Location Relative to End of FRP Reinforcement (in.)
Figure 78. The distribution of a is plotted for variousFRP-to-wood stiffness ratios on a beam with a bumper. Bestfit polynomials are plotted through the finite-element data.
1000 -
800 -
600 -
400 -
200
142
0-0.3 -0.25 -0.2 -0.15 -0.1 -0.05
Table 12. Stresses and stress ratios at the end of the FR?lamina for various stiffness ratios for glulam with abumper.
Figure 79. The effect of the FRP-to-wood stiffness ratio onc near the end of the FRP lamina is plotted for averagestress from the end of the FR? lamina to 0.05 and 0.15 in.past the end of the FR? lamina.
0.14
0.12 -
0.10 -
0
0.08 -
0.06-Cl)
0.04 -
U-
- 400
---U
-350
. 300
- 250
- 200
- 150
- 100
145
0.02U--O.05 in. Avg - 50
O.15 in. Avg0.00 : 0
0 2 4 6 8 10
Stiffness Ratio, FR? Reinforcement-to-Wood
Figure 80. The effect of the FRP-to-wood stiffness ratio onc near the end of the FRP lamina is plotted for averagestress from the end of the FRP lamina to 0.05 and 0.15 in.past the end of the FRP lamina.
0.4 -
146
0.5- -2500
0.2- -1000
0.1- -500
UO.05 in. Avg--O.15 in. Avg
0.0 0
0 2 4 6 8 10
Stiffness Ratio, FRP Reinforcement-to-Wood
Figure 81. The effect of the FRP-to-wood stiffness ratio onnear the end of the FRP lamina is plotted for average
stress from the end of the FRP lamina to 0.05 and 0.15 in.prior to the end of the FRP lamina.
0.6 3000
and a,. The magnitudes of all stress distributions
increase significantly with stiffness ratio.
Effect of Reinforcement Length
All beams represented in these models were 5-1/8 x
12.14 in. x Length ft and had a 0.14-in, thick 16.6 x 106
psi FRP lamina. The models were subjected to third-point
loads. The FRP lamina lengths are given in Table 6.
Figures 82 to 84 show the stress distributions, c, o
and a, respectively, for 60% length FRP lamina and for 40%
length FRP lamina. It is apparent that length of FRP lamina
does not contribute significantly to stress distributions.
Effect of Beam Width
All beams represented in these models were Width x
12.14 in. x 21 ft and had FRP lamina over 60% of the length.
The models were subjected to third-point loads. The beam
widths are given in Table 6.
Figures 85 to 87 show the stress distributions, ax, a
and a respectively for 5-1/8 in. and 10.75 in. beam
widths. It is apparent that width does not significantly
effect stress distributions at the end of the FRP lamina.
147
25000
20000 -
:;: 15000 -
4- 10000 -UD
5000 -
0
-60% Length Reinforced-40% Length Reinforced
-
Location Relative to End of FRP Reinforcement (in.)
Figure 82. The distribution of is plotted for two FRPlamina lengths on a beam with a bumper.
148
0 0.1 0.2 0.3 0.4 0.5 0 6
400
350 -
300 -
250-
,200-
(I)a)
150-
100 -
50 -
0
0
N
-60% Length Reinforced-40% Length Reinforced
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 4
Location Relative to End of FRP Reinforcement (in.)
Figure 83. The distribution of o is plotted for two FRPlamina lengths on a beam with a bumper.
149
-60% Length Reinforced.-40% Length Reinforced
-
14
1200 -
1000 -
800 -
600 -
400 --
200 -
0
150
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
Figure 84. The distribution of c is plotted for two FRFlamina lengths on a beam with a bumper.
Cl)
25000
20000
15000
U)
10000
5000 -
0
-5.125 in. Wide-10.75 in. Wide
151
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FRP Reinforcement (in.)
Figure 85. The distribution of a is plotted for two beamwidths on a beam with a bumper.
400
350 -N
300 -
250 --U)
200-U)ci)
150-
0
0
100 --
50 - -10.75 in. Wide-10.75 in. Wide
Figure 86. The distribution ofwidths on a beam with a bumper.
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 4Location Relative to End of FRP Reinforcement (in.)
152
is plotted for two beam
Figure 87. The distribution ofwidths on a beam with a bumper.
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FRP Reinforcement (in.)
153
is plotted for two beam
0
1400
1200
1000 -
800 -
600 -
400 -
-5.125 in. Wide 200 --10.75 in. Wide
Effect of Span-to-depth Ratio
All beams represented in these models were 5-1/8 x
12.14 in. x Length ft and had FRP lamina over 60% of the
length. The models were subjected to third-point loads.
The span-to-depth ratios are given in Table 6.
Figures 88 to 90 show the stress distributions, a, a
and a respectively, for 21:1 and 27:1 span-to-depth
ratios. It is apparent from these figures that span-to-
depth ratio does not contribute significantly to stress
distributions at the end of the FRP lamina.
Summary of Analytical Results
Magnitudes of stresses found at the end of the FRP
lamina show stress levels in wood much greater than
published strength values for Douglas-fir. While these
stress levels may appear to be impossible or too high, it is
important to consider these factors: 1) this is a linear
model based on linear material properties, 2) the length of
wood affected by the stress rise is less than half an inch,
and 3) there is little understanding of how individual wood
cells behave due to isolated shear and tensile stresses at a
microscopic level.
154
25000
20000 -
15000
.4-) 10000 -U) -
5000 -
-Span:Depth 21:1-Span:Depth = 27:1
0
0 0.1 0.2 0.3 0.4 0.5 0 6
Location Relative to End of FRP Reinforcement (in.)
Figure 88. The distribution of cT is plotted for two span-to-depth ratios on a beam with a bumper.
155
400
350 -
300 -
250
100 -
50
0
.--Span:Depth = 21:].--Span:Depth = 27:1
156
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 04
Location Relative to End of FRP Reinforcement (in.)
Figure 89. The distribution of a is plotted for two span-to-depth ratios on a beam with a bumper.
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
Location Relative to End of FR? Reinforcement (in.)
Figure 90. The distribution of is plotted for twospan-to-depth ratios on a beam with a bumper.
140
157
1200
1000
800
600
400
.Span:Depth = 21:1 200=Span:Depth 27:1
158
As a model based on linear material properties, it does
not incorporate nonlinear material behavior..
The second factor to consider is the localized nature
of the stress riser. Tensile and other strength properties
for wood are based on large specimens when compared to the
region of the stress rise. The Wood Handbook (FPL, 1987)
gives values between 12,400 and 12,600 psi for coast and
interior west Douglas-fir at 12% moisture content.
Individual fiber (0.1 to 0.25 in. in length) strengths for
Douglas-fir range from 49,850 psi for earlywood to 138,000
psi for latewood (Bodig and Jayne, 1982).
It is understandable that FRP lamina thickness and
stiffness are related to the stress distributions at the end
of the FRP lamina. FRP lamina thickness and stiffness
directly relate to the amount of force carried by FRP lamina
that must be transferred to wood at the end of the FRP
laniina.
Although beam width was thought to influence stress
distributions at the edge, the model found no significant
influence on stress distributions.
Length of FRP lamina was thought to be important since
the distance over which stress is transferred to the beam
was unknown; i.e. how stress is transferred through the
glueline by shear.
159
Span-to-depth ratio was found to have no influence on
stress distributions when beam depth was held constant.
Uniform loading and third-point loading provided
essentially the same stress levels.
The lack of influence due to FRP lamina length, span-
to-depth ratio and third-point versus uniform loading
conditions implied that the moment distribution near the end
of the FRP lamina was more important than how the beam was
loaded some distance from the end of the FRP lamina.
However, a small increase in stress was observed for the 80%
length reinforced beam and the beam with a span-to-depth
ratio of 10 where the end of the FRP lamina was only 2 ft
from the support. This suggests that the effect of the
support caused the change in stress levels and not the
length of FRP lamina or stress levels.
While there are many similarities in stress component
distribution and parameters of influence between beams with
and without a bumper, there are differences. A significant
difference is due to the gap at the end of the FRP lamina in
beams with a bumper. When the beam with a bumper is placed
in bending, the gap causes the end of the FRP lamina to be
somewhat compressed through its depth. This compression
causes the stress (peel stress) to be compressive instead
of tensile. This effect is beneficial since a compressive
peel stress is unlikely to cause the lamina to "peel" off.
160
There are also notable differences in the magnitudes of
stress components. The stress is similar for small FRP
lamina thicknesses in both beam types, while the stress is
greater for larger FRP lamina thickness in beams with a
bumper. The larger c in beams with a bumper is also likely
caused by the bumper bending more at the gap. The shear
stress , is also larger for beams with a bumper.
V. CONCLUSION
Experimental and analytical studies of stress
distributions at the end of the FRP lamina for partially
reinforced glulam were performed. Based on these analyses,
stress levels are higher near the end of the FRP lamina than
levels expected at the same location in a solid wood member.
The most important stresses to consider include a, ay, z
and for glulam without a bumper and c, and for
glulam with a bumper. FRP lamina thickness and the ratio of
FRP lamina stiffness to wood stiffness have the greatest
effect on stress levels of the variables studied.
While this study contributes to the understanding of
stress distributions at the end of the FRP lamina in
partially reinforced glulam, more research needs to be
performed. In order to effectively apply the results of
this study, stress levels predicted by the model need to be
compared to load levels that cause failure in actual tests.
Further research should be aimed toward reducing stress
levels at the end of the FRP lamina by comparing various
shapes of the FRP lamina tail.
161
BIBLIOGRAPHY
Adams, D. A., and W.C. Wake. 1984. Structural adhesivejoints in engineering. Elsevier Applied SciencePublishers. New York, NY. 309 p.
American Forest & Paper Association (AF & PA). 1991.
National Design Specification®, NDS® Supplement.AF & PA, Washington, DC. 125 p.
American Plywood Association (APA) - The Engineered WoodAssociation. 1995. Regional Production andDistribution Trends for Structural Panels andEngineered Wood Products: E58. APA, Tacoma, WA. 59 p.
American Society for Testing and Materials (ASTM). 1995.Standard method of static tests of lumber in structuralsizes, D 198. p. 57-75. In 1995 Annual Book ofStandards, Vol. 04.10. ASTM, Philadelphia, PA.
Biblis, E.J. 1965. Analysis of wood fiberglass compositebeams within and beyond the elastic region. ForestProducts Journal. 25(24):81-88.
Bodig, J., and B.A. Jayne. 1993. Mechanics of Wood andWood Composites. Krieger Publishing Company. Malabar,FL. 712 p.
Bohannan, B. 1962. Prestressing wood members. ForestProducts Journal. 12(12):596-602.
Braun, M.O., and R.C. Moody. 1977. Bending strength ofsmall glulam beams with a laminated-veneer tensionlamination. Forest Products Journal. 27(11) :47-51.
Chawla, K.K. 1987. Composite Materials Science andEngineering. Springer-Verlag, New York, NY. 292 p.
Cheng, S., D. Chen, and Y. Shi. 1991. Analysis ofadhesive-bonded joints with nonidentical adherends.Journal of Engineering Mechanics. 117 (3) :605-623.
Choo, V.K.S. 1990. Fundamentals of Composite Materials.Knowen Academic Press, Inc. Dover, DE. 313 p.
162
Coleman, G.E., and H.T. Hurst. 1974. Timber structuresreinforced with light gage steel. Forest ProductsJournal. 24(7) :45-53.
Crocoinbe, A.D., and R.D. Adams. 1981. Peel analysis usingthe finite-element method. Journal of Adhesion.12(2) :127-139.
Davalos, J.F., and E.J. Barbero. 1991. Modeling of glass-fiber reinforced glulam beams. p. 3.234-3.241. InProceedings of the 199]. International TimberEngineering Conference, vol. 3. September 2-5, 1991.London, England. Timber Research and DevelopmentAssociation, High Wycomb, UK.
Dziuba, T. 1985. The ultimate strength of wooden beamswith tension reinforcement. Holzforschung undHolzverwertung. 37(6) :115-119.
Enquist, B., P.J. Gustafsson, and H.J. Larsen. 1991.Glass-fibre reinforcement perpendicular to the grain.p. 3.242-3.250. In Proceedings of the 1991International Timber Engineering Conference, vol. 3.London, England. September 2-5, 1991. Timber Researchand Development Association, High Wycomb, UK.
Forest Products Laboratory (FPL). 1987. Wood Handbook -Wood as an Engineering Material. Agriculture HandbookNo. 72. U.S. Department of Agriculture. Washington,D.C. 466 p.
Gardner, G.P. 1991. A reinforced glued laminated timbersystem. p. 3.295-3.300. In Proceedings of the 1991International Timber Engineering Conference, vol. 3.London, England. September 2-5, 1991. Timber Researchand Development Association, High Wycomb, UK.
Gere, J.M., and S.P. Timoshenko. 1990. Mechanics ofMaterials. PWS-KENT Publishing Company, Boston, M7½.807 p.
Gilham, P.C. 1995. Design of timber bridge stringers withhigh strength fiber reinforcing. p. 13-27. InProceedings of Structures Congress XIII, vol. 1.Boston, M7.. April, 1995. 1merican Society of CivilEngineers, New York, NY.
Gupta, G.D. 1973. A Layered Composite with a BrokenLaminate. International Journal of Solids Structures.9:1141-1154.
163
164
Horgan, C.O. 1982. Saint-Venant end effects in composites.Journal of Composite Materials. 16:411-422.
International Conference of Building Officials (ICBO). 1995.Fiber-reinforcred Plastic (FiRP) Reinforced Glued-laminated Wood Beams, NER No. 5100. ICBO. Whittier,CA. 20 p.
Jones, R.M. 1975. Mechanics of Composite Materials.Hemisphere Publishing Corporation. New York, NY.355 p.
Kim, R.Y., and S.R. Soni. 1984. Experimental andanalytical studies on the onset of delamination inlaminated composites. Journal of Composite Materials.18(1) :70-80.
Krueger, G.P. 1973. Ultimate strength design of reinforcedtimber: state of the art. Wood Science. 6(2):175-186.
Krueger, G.P., and L.B.Sandberg. 1974. Ultimate-strengthdesign of reinforced timber beams: Evaluation of designof reinforced timber beams: evaluation of the designparameter. Wood Science. 6(5):316-330.
Lantos, G. 1970. The flexural behavior of steel reinforcedlaminated timber beams. Wood Science. 2(3):136-143.
Lee, Stuart, M. Handbook of Composite Reinforcements. VCHPublishers, Inc. New York, NY. 715 p.
Mark, R. 1961. Wood-aluminum beams within and beyond theelastic range. Part I: Rectangular sections. ForestProducts Journal. 11(10):477-484.
Mark, R. 1963. Wood-aluminum beams within and beyond theelastic range. Part II: Trapezoidal sections. ForestProducts Journal. 13(11):508-516.
165
Moulin, G.P., and P. Jodin. 1990. FGRG: Fibreglassreinforced gluelam - a new composite. Wood Science andTechnology. 24:289-294.
Rowlands, R.E., R.P. Van Deweghe, T.L. Laufenberg, and G.P.Krueger. 1986. Fiber-reinforced wood composites.Wood and Fiber Science. 18(1):39-57.
Sliker, A. 1962. Reinforced wood laminated beams. ForestProducts Journal. 12(12):91-96.
Spaun, F.D. 1981. Reinforcement of wood with fiberglass.Forest Products Journal. 31(4):26-33.
Swanson Analysis System, Inc. (ANSYS). 1995. ANSYS,version 5.1. Swanson Analysis System, Inc. Houston,PA.
Tingley, D.A. 1994. Glued-laminated beams having a high-strength fiber reinforcement: The bimaterialinterface. p. 665-675. In Proceedings of the 1994Pacific Tinther Engineering Conference, vol. 1. July11-15, 1994. Gold Coast, Australia.
Tingley, D.A. 1990 Predicting strength criteria for Keviarand fiberglass reinforced plastic (KRP and FRP) gluedlaminated beams. p. 42-45. In Proceedings of the 1990International Timber Engineering Conference, vol 1.October 23-25, 1990. Tokyo, Japan. Steering Committeefor the 1990 International Timber EngineeringConference, Tokyo, Japan.
Tingley, D.A., and R.J. Leichti. 1992. Reinforced glulam:improved wood utilization and product performance. p.221-223. In Proceedings of the Globalization of Wood:Supply, Processes, Products, and Markets. November 2-3, 1992. Portland, OR. Forest Products Society,Madison, WI.
Triantafillou, T.C., and Nikola Deskovic. 1992.Prestressed FRP sheets as external reinforcement ofwood members. Journal of Structural Engineering.118(5) :1270-1284.
166
Van de Kuilen, J.W.G. 1991. Theoretical and experimentalresearch on glass fibre reinforced laminated timberbeams. p. 3.226-3.233. In Proceedings of the 1991International Timber Engineering Conference, vol. 3.London, England. September 2-5, 1991. Timber Researchand Development Association, High Wycoinb, UK.
Whitney, J.M., and R.J. Nuismer. 1974. Stress fracturecriteria for laminated composites containing stressconcentrations. Journal of Composite Materials.8(8) :253-265.
Wood Science and Technology Institute (WS&TI) (N.S.), Ltd.1995. Personal Communication.
167
APPENDIX
APPENDIX A
Table Al. Coefficients for polynomial curve fits to finite-element stress distributions for glulam beams without abumper, where thickness of reinforcement is varied.
Table A2. Coefficients for polynomial curve fits to finite-element stress distributions for glulam beams without abumper, where stiffness of reinforcement is varied.
Table A3. Coefficients for polynomial curve fits to finite-element stress distributions for glulam beams with a bumper,where thickness of reinforcement is varied.
Table A4. Coefficients for polynomial curve fits to finite-element stress distributions for glulam beams with a bumper,where stiffness of reinforcement is varied.
BMLENGTH=2 1!! !BEAM LENGTH (It)L=(BMLENGTH)* 12/2 I! !MODELED LENGTH (IN) INCLUDING 3 iNCHES PAST SUPPORTBMWIDTH=5.125 !BEAM WIDTH (IN)W=(BMWIDTHJ2) !MODELED BEAM WIDTH (IN)H=12 !BEAM HEIGHT (IN)HF=3 !HEIGHT OF FINE MESH (IN)HT=5 !HEIGHT OF TRANSITION MESH (IN)LENGTHRP=60 !RP LENGTH (%)LRP=(LENGTHRP/100)*L !MODELED RP LENGTH (IN)LLT=LRP-2LRT=LRP+2HRP=O.14 !RP THICKNESS (IN)LOADDIST=L/3 I DISTANCE FROM CENTER TO LOAD (IN)WDIV=8LOAD=((-1464314)/(WDIV+1)) !LOAD APPLIED AT TOP SURFACE NODES 32-36 IN APART(LB)
L]PE=12LTIPE=8LFRPIPE=O. 15LFWIPE=O. 1DDIV=12
MOEX=2E6
ET, 1,SOLID45
MP,EX,1,MOEX !WOOD PROPERTIESMP,EY,1,(O.059*MOEX) !(AVG ET/EL AND ER/EL FROM WOOD HANDBOOK)MP,EZ,1,(O.059*MOEX) !
MP,PRXY, 1,((O. 169e6)*MOEX) ! (AVG FROM vRL/EL AND vTL/EL FROM BODIG & JAYNE)MP,PRXZ,1,((O.169e6)*MOEX) !
MP,PRYZ,1,O.41 !(WOOD HANDBOOK)MP,GXY,1,O.115E6 'BODIG AND JAYNEMP,GXZ,1,O.115E6MP,GYZ,1,O.012E6 !
BMLENGTH=21 !BEAM LENGTH (fi)L=(BMLENGTH)*12/2 !MODELED LENGTH (IN) iNCLUDING 3 iNCHES PAST SUPPORTBMWIDTH=5. 125 !BEAM WIDTH (iN)W=(BMWIDTHJ2) 'MODELED BEAM WIDTh (IN)HRP=O.07 !RP THICKNESSI{BUMP=1.5 !BUMPER THICKNESSHB=HRP+HBTJMP !DISTANCE FROM TOP OF RP TO BO1TOM OF BUMPERH=12-HBUMP 'HEIGTH ABOVE RPLENGTHRP=60 !RP LENGTH (IN)LRP=(LENGTHRP/100)*L !MODELED RP LENGTH (iN)LLT=LRP-2 !DISTANCE TO FINE MESH LEFT OF END RPLRT=LRP+2 'DISTANCE TO FINE MESH RIGHT OF END RPLOADDIST=L/3 DISTANCE FROM CENTER TO LOAD (iN)WDIV=8LOAD=((-20023/4)/(WDIV+1)) 'LOAD APPLiED AT TOP SURFACE NODES 32-36 IN APART(LB)
LIPE=12LTJPE=8LFRPIPE=O. 15LFWTPE=O. 1DDIV=1 1
MOEX=2E6
ET, 1,SOL1D45
MP,EX,1,MOEX 'WOOD PROPERTIESMP,EY,1,(O.059*MOEX) !(AVG ET/EL AND ERJEL FROM WOOD HANDBOOK)MP,EZJ,(O.059*MOEX)MP,nuXY, 1,((O. 169e6)*MOEX)*.O59 !(AVG FROM vRL/EL AND vTL/EL FROM BODIG &JAYNE)MP,nuXZ,1,((O.169e_6)*MOEX)*.059 !
MP,nuYZ,1,O.41 '(WOOD HANDBOOK)MP,GXY,1,O.115E6 !BODIG AND JAYNEMPGXZ1O115E6 t