Comparison of a theoretical analysis with an

experimental stress analysis of an frp thin-wall composite box beam,

by

CLARK TIMOTHY HARBAUGH
B.S., Kansas State University, 1972

A THESIS

submitted in partial fulfillment of the

requirements for the degree

MASTER OF SCIENCE

Department of Mechanical Engineering

KANSAS STATE UNIVERSITY Manhattan Kansas
,

1989

Approved by:

->/

ft

//,
i
,

^<<^i^^Y [It

Major Professor

.

AllEOfi 3DM3Q2

PREFACE
An oft used purpose of experimental methods and

experimental

analyses is to verify whether the
along with any assumptions

theory
made,

based

calculations,

can be substantiated experimentally.

It was with this

intention

that this project was undertaken.

Although

the methods of analyzing stress and strain in isotropic

materials are well documented,

the analysis of noniso-

tropic materials is not so well established. This work examines the theoretically
stress
in
a thin-wall box beam which

predicted
fabricated

was

from

a nonisotropic composite material.

The

elastic
and

constants of this composite material were unknown,
so

had

to be determined using

experimental

methods.

From these and other intermediate steps, the stress and

corresponding strain were predicted in the box beam due
to a known loading condition.

These were then compared

to the results found by experimental methods

One

difficulty which was encountered
modulus
of elasticity of

involved

finding

the

the

composite

material in compression.
value

The experimentally determined
to

was deemed invalid and so another method had

be used to find its value.
ii

As with most projects,

many others contributed
Appreciation

to the sucessful completion of this work.
is hereby expressed to Dr.

Hugh

S.

Walker for his help
In addition,

and guidance in completing this project.
I

am grateful to Dr.

K.

K.

Hu and Dr. Chi L. Huang for

their

encouragement
In

and

stimulation
the

in

preparatory
of

coursework.

addition,

assistance

Gary

Thornton and Brian Bramel was invaluable in cutting and

machining
fixtures.

the

test

specimens
is

and

in

making
the

test

Gratitude

also due to

Mechanical

Engineering
test

Department for furnishing the material and
for use in this project,

equipment

and use
of

to

the

Civil

Engineering

Department
Finally,
I

for the

their

Riehle test machine.

am indebted to my wife

Olinda and our two children for their enduring patience

while

I

labored on this project.

111

TABLE OF CONTENTS
Page

PREFACE
LIST OF TABLES

ii

v

LIST OF FIGURES
Chapter
I.

vii

INTRODUCTION
METHODS AND PROCEDURES

1 4
4

II.

Description of FRP Composite
Experimental Determination of FRP Composite Properties

10
59

Transformation of Properties
Experimental Determination of Beam Deflections

62

Theoretical Determination of Compressive Modulus of Elasticity Theoretical Beam Stresses and Strains
Experimental Beam Strain
Shear Lag Effect on Beam Strain Measurements
III.

.

67

69

70

79
86
86 87

RESULTS AND CONCLUSIONS
Results

Conclusions
IV.
V.

SELECTED BIBLIOGRAPHY APPENDIX
iv

89
90

LIST OF TABLES Table
1. 2. 3. 4. 5.

Page

Theoretical Beam Stresses Theoretical Beam Strains

.... ....
.

69 70

Strain Gage Multiplying Factors
Strains for Position "A" Strains for Position "B"

72

.... ....
.

79
79
86
1 2
3

6.

Comparison of Strain Values

.

.

7.
8.
9.

X Axis Tension Test, Specimen No.
X Axis Tension Test X Axis Tension Test

90 90

Specimen No.
Specimen No.

91 91 92
92
93 93 93

10.

Y Axis Tension Test Y Axis Tension Test Y Axis Tension Test
X Axis Flexure Test X Axis Flexure Test X Axis Flexure Test X Axis Flexure Test X Axis Flexure Test Y Axis Flexure Test

Specimen No. Specimen No. Specimen No.
Specimen No. Specimen No.

4
5 6

11.
12.

13.
14. 15.

7
8 9

Specimen No.

16.
17. 18. 19. 20.

Specimen No. 10 Specimen No. 11 Specimen No. 12

94 94 94 95

Y Axis Flexure Test Y Axis Flexure Test

Specimen No. 13 Specimen No. 14

95

LIST OF TABLES, continued
Table
21. 22. 23. 24.

Page
X Axis Compression Test, Specimen No. 15 X Axis Compression Test, Specimen No.
16
.

95 96 96 96 97
97

.

X Axis Compression Test, Specimen No. 17 X Axis Compression Test, Specimen No. 18 X Axis Compression Test, Specimen No. 19 Y Axis Compression Test, Specimen No. 20 Y Axis Compression Test, Specimen No. 21

.

.

25.
26.

.

.

27.
28. 29.
30. 31. 32.

.

98 98 99 99

Y Axis Compression Test, Specimen No. 22
Y Axis Compression Test, Specimen No. 23

.

.

Y Axis Compression Test, Specimen No. 24

.

On-Axis Shear Test
Off -Axis Shear Test

100
100

33.
34.

Box Beam Load vs. Deflection Box Beam Strains, Position "A" Box Beam Strains, Position "B" Test Specimen Dimensions

101
101 101 102

35. 36.

vl

Tensile Test Specimen No. Load vs. 16 9. 24 15. Tensile Test Specimen No. Tensile Test Specimen No. 25 Transverse Strain 16.. 22 Longitudinal Strain 13. 26 Longitudinal Strain vii . 3. Tensile Test Specimen No. 13 Longitudinal Strain 6. Load vs. 23 14. 5. . 2. 6. Load vs. Transverse Strain 2. Transverse Strain 1. 18 . 5 8 4. 3. Load vs. Transverse Strain Tensile Test Specimen No. 1. Load vs. 2. 15 8. Load vs. Longitudinal Strain Tensile Test Specimen No. Longitudinal Strain 4. 11 Tensile Test Specimen No. Load vs. Load vs. . Page Detail of Plain Weave 5 Original Fiberglass Reinforcing Cloth Box Beam Fabrication Details X Axis Tensile Test Specimen Dimensions . 21 12. 3. Load vs. Load vs.LIST OF FIGURES Figure 1. Longitudinal Strain Tensile Test Specimen No. Transverse Strain Tensile Test Specimen No.. 5. 11. Tensile Test Specimen No. Load vs. . Y Axis Tensile Test Specimen Dimensions 4. 17 10. 14 7. 5. Tensile Test Specimen No.

27 28 18. Deflection Flexure Test Specimen No. Strain viii 41 30. 42 31. Transverse Strain 6. 31 22. 19. Page Tensile Test Specimen No. Deflection Flexure Test Specimen No.. 39 Compression Test Specimen No. Deflection Flexure Test Specimen No. Load vs. 18. 43 32. Load vs. Deflection Flexure Test Specimen No. 15. 29. Compression Test Specimen Dimensions .. Strain Compression Test Specimen No. Load vs. 12. Load vs. 44 . 36 26. 32 23. Load vs. 20. Deflection 7 .. 37 27. continued Figure 17. Load vs. Deflection 34 25. Deflection Flexure Test Specimen No. Flexure Test Specimen No. Strain Compression Test Specimen No. Load vs. 11. Load vs. Flexure Test Specimen Dimensions Flexure Test Fixture Details Flexure Test Specimen No. Load vs. Strain Compression Test Specimen No. 14. 10.LIST OF FIGURES. 29 Load vs. 16. Deflection 33 24. 13. 8. 9. Load vs. Load vs. 38 28. Load vs. 30 21. Flexure Test Specimen No. 17.

Deflection. 49. 40. 24. 65 . 61 63 46. Strain. 56 Axial Gage 43. 23. continued Figure 33. 49 38. Strain 45 34. Load vs. 66 71 . Strain Compression Test Specimen No. Load vs. 55 Load vs. 21.LIST OF FIGURES. Load vs. 50 52 53 39. On-axis Shear Test. Strain. Strain Shear Test Specimen Dimensions Shear Test Fixture 47 36. Fiberglass Reinforcement and Axes Orientations Box Beam Bending Positions Box Beam Load vs. 41. Strain. Deflection. Gage #3 58 45. Position "A" Box Beam Load vs. Gages #1. Strain. Load vs. Load vs. 48. 19. 46 35. Off -axis Shear Test. Load vs. Compression Test Specimen No. Off -axis Shear Test. 47. Load vs. Load vs. 48 37. Position "B" Box Beam Strain Gage Locations ix . Strain Compression Test Specimen No. Strain Compression Test Specimen No. Strain Compression Test Specimen No. #3 42. 22. Page Compression Test Specimen No. 20. Gage #1 57 44. Load vs. On-axis Shear Test.

Strain.LIST OF FIGURES. #8 Strain. Position "B" Load vs. Position "A" Box Beam Strain vs. Position "B" Load vs. Gages #5. Gage #4 55. 74 Box Beam. Box Beam. 76 Strain. Box Beam. #6. Gage Location. 77 Box Beam. Gage #5. Position "A" Load vs. continued Figure 50. Position "A" Load vs. Gage Location. #3 Strain. #3 54. Gage #1. Gages #1. 84 x . Strain. Position "B" Load vs. Position "B" Shear Lag Side Side 1 2 80 57. Gage #4 52. #8 78 56. Strain. Page Box Beam. Brittle Coating Brittle Coating 60. 73 51. 59. #6. 75 53. Box Beam. 81 82 83 58. Position "A" Load vs. Box Beam Strain vs.

. the Although other types and varieties most commonly used composites consist of a high modulus modulus constituent for strength enhancement and a low binding material known as the matrix. particles. kers. Its primary functions are to provide binding act strength. material The high modulus may be in the form of long directionally oriented fibers. etc.Chapter I INTRODUCTION A composite consists material is one which. the as a filler. tinct of two (or possibly more) phases or materials on a macroscopic exist. support the high modulus material. and provide fracture toughness. by most disscale. short randomly oriented fibers. protect high modulus material from abrasion and environmental agents. reguired to obtain the necessary laminate The matrix is usually some type of resin. composites which use long fibers for reinforceare multiple layers of the fibrous material generally thickness. definitions. spheres. foams. For those ment. although many of the conventional engineering materials can also be used. whis- flakes. transfer the stresses.

noteworthy characteristic of FRP is that it. 1/5 These include a specific gravity to roughly that of steel. like most composites. savings in This property can mean significant since fiber- weight many applications. sensitive. and fiberglass is relatively inexpensive One readily available in many forms. (fiberglass reinforced offers many advantages over steel and most of other metals. However. over 80% of the composite materials produced inforcing plastic) the today still utilize fiberglass FRP as the re- material. reinforced composites can be utilized which have their maximum principal strength oriented in the same direction as the stress. this a characteristic of being nonisotropic necessitates standing of more thorough underhow the the stress-strain relations and 2 . low electrical and thermal conductivity rates.Although much of the attention and development known com- work as today centers on those composite materials "advanced composites" or "high performance posites" because from of the very high specific strengths available these generally unidirectional (UD) fibrous composites. and an ability to obtain physical properties which are directionally Moreover. is nonisotropic. a strong resistance chemical attack and corrosion.

composite's properties vary depending upon the type and orientation materials of the reinforcement. fabri- cating a thin-wall box beam using this FRP composite (with the primary axes of the fiberglass reinforcement axis). combining to these two into the object of this research was done provide some insight into the problems associated with using FRP for structural applications. Thus composite in can offer significant advantages many applications. but a careful analysis of their unigue reguire phy- sical characteristics. Because box beams are important structural elements and FRP is used in many industries. rotated then the relative to the beam longitudinal and comparing experimentally measured strains with and theoretically calculated strains on the upper surfaces of the box beam subjected to a lower load. was This objective the the accomplished of by investigating and comparing finding results and problems associated with elastic constants of a particular FRP composite. simple . also particularly the aerospace industry.

0005 inches in The diameter.020 inches. twisted together. The material was modified tion. The matrix was a thermoset type polyester resin which is manufactured for the automotive company Balkamp. strength between the two principal direc- Each individual yarn consisted of approximately 200 individual. each one being approximately 0. Inc. overall outside diameter of the entire yarn was approximately 0. The purpose of this modification was to increase the differential tions. by removing every third yarn in the Y 9 direc- leaving 1/3 yarns per inch in this direction. glass filaments (or fibers). peroxide catalyst to initiate 4 . It utilizes a MEK (methyl ethyl ketone) the curing process.Chapter II METHODS AND PROCEDURES DESCRIPTION OF FRP COMPOSITE The fiberglass reinforcement used in this study consisted of a plain weave fiberglass material X which (warp) originally contained 16 yarns per inch in the direction and 14 yarns per inch in the Y (fill) (see Figures 1 direction and 2).

Figure 1 Detail of Plain Weave Figure 2 Original Fiberglass Reinforcing Cloth .

This was done so that the physical properties of the composite could inplane the be determined in each of the two principal directions. each layers' X and Y axes were collinear. a nylon peel ply. The vacuum bagging process intended to compress the resin-impregnated fiberglass and air cloth spaces. After all of the reguired number of pieces were properly stacked or wrapped covered release over each other. outer nylon bagging so film used to provide an airtight to enclosure entire that a vacuum could be applied the is fabrication. the catalyzed resin into the fiberglass one piece at a time. and provide a rela- tively smooth surface finish. it was laid into the desired position.The hand layup process consisted of manually- working material. The the flat FRP composite plate from which all of test specimens were cut was made with each ie. layer of the fiberglass material aligned the same. perforated nylon and an an absorbent bleeder material. After each piece of fiberglass cloth was completely impregnated with resin. eliminating voids draw out the excess resin. layers together. Obtaining axis the properties in each of principal composite directions is essential in since a evaluating material properties 6 characteristic . the entire fabrication a was with film.

the piece having its X axis offset from axis by +10 . These correlations are called coefficients of mutual influence by Lekhnitski [2]. Each piece of fiber- cloth was of a predetermined size four necessary with the to make first complete layers around the core.of directionally oriented fibrous composites is that a uniaxial normal stress may produce a shear strain and a pure shear stress may produce a normal strain [1]. This effect will be minimal in this analysis since the box beam dimensions are large comthe pared to the beam wall thickness and amount of bending which occurred was small. beam's of longitudinal the The second piece was proper dimensions to make four layers over the first four. See This procedure produced a box beam wall which would be classified as a balanced. however. except that the X axis of these layers was -10 . The thin-wall box beam was fabricated by wrapping two pre-cut pieces of resin-impregnated fiberglass glass cloth around a wood core. ply laminate. . antisymmetric angle Being antisymmetric will produce a coup- ling effect between bending and extension. which causes warping. offset Figure 3 from the longitudinal beam axis by for details.

1 i i „596"0 A 4-> c <u o a o J-4 e M-< C •r-l 0) M <4-l •p O w M cu CO c <u ^ >— <f- 6 <U O J-i H 4-) CO •H CO O <4-l Q c O 4-) CU c •H QJ >-l U •H CO P 3 O M-l Q) 4-1 O W !-i O •H !-i -Q CO O i—i a) Du >^ CO CO i— -o- 6 CO CU U •H DQ a. H ^ u CD c c •H M-l o CQ X CO O .— CO CU U •H O •H Cu >^ H .

One the drawback of the manual layup procedure was to ensure uniformity in the inability beam wall thickness. Such occurrences are difficult to avoid. wood core. however. The cut ends of the beam and all of the removed. were sealed with resin to eliminate existence of free glass fibers which inevitably from the cutting process. The ends of the beam were and cut the to the dimensions indicated in Figure 3. The and the 8 16 layer flat plate previously described described as long as layers of the box beam wall just can be classified as orthogonal laminates. box beam having a thicker wall than the other side. After a vacuum was applied and the resin had cured overnight. The side of the beam where the outer layer of of bagging film was overlapped had a smaller amount the resin squeezed out during the vacuum bagging This resulted in one side of the and curing phase. specimen edges. all of the nylon covering and bleeder material off was removed. 9 . result The teeth on the sawblade caused those fibers near the surface on the downside of the cut to become disengaged from the matrix as they were cut by the moving sawblade teeth. was test which was split and dowelled in the middle. in a con- tinuous wall structure fabricated manually. except for the compression test the specimens.

EXPERIMENTAL DETERMINATION OF FRP COMPOSITE PROPERTIES The first step necessary for both the theorwas of etical approach and the experimental analysis the to determine independent elastic constants Those which were xy' . . xy' 1/. as determined along formed angle the two inplane principal axes. For an orthotropic material subjected plane stress. G.they are restricted to plane stress conditions. E yx . This transformation will be discussed later. elastic xy constants (any or G xy . ^yx' and G xy Standards were used as guides for all of the experimental analyses. and either G yx ). elastic constants (for isotropic materials. . determined . Another property of nonisotropic materials analysis which is an important part of an experimental is the fact that they have more than two independent E. and V to are related) . experimentally included E x v The ASTM E . 10 . to be trans- to yield the desired elastic properties at some to the principal axes. 1/ 1/ . 3 there are five of E independent . This property allows their elastic properties. this particular FRP composite. yx' l/m .

A CN in Tabs (typ. a different set of experimental tests were performed to determine each value. determining The guide used for the modulus of elasticity in tension and 4 poisson's ratio was ASTM Standard D3039 [3]. Specimen of 4) Figure 4 X Axis Tensile Test Specimen Dimensions 11 . the Figure the approximate dimensions of test ^_ I7~ y x.Determination of E xy and Z^xy Since a characteristic of most fiber-reinforced composites tension is that their modulus of elasticity in is greater than their modulus of elasticity in compression. illustrates specimens.

= modulus of elasticity in tension (psi) = applied tensile load P (lbs) ) A € = test specimen cross-sectional area (in 2 = measured strain (in/in) 12 . except that no yarns were removed from the fiber- glass cloth. Group and a Strain Model P-3500 (S/N 50798). elasticity in tension was found by use of the formula E.. Division. the modulus of in Figures 5-10. Riehle test the Where they tabs were required on tensile fiber- were fabricated from the same glass material and resin as the test specimens and box beam. machine. M-Bond 200 adhesive. specimens. Inc. Measurement Group Switch and Balance Unit. All strain gages used in this and successive applications were manufactured by Micro-Measurements 120LZ-120 gages.000 lb. = WA (1) where E. EA-06All with a gage length strain readings were taken using a Measurements Indicator. (S/N 033433) Model SB-1 The graphs of the load versus strain are shown For each specimen. All tests on the test specimens and the box beam were performed utilizing a 20. They were cut from a larger piece and glued onto the test specimens using Measurements Group. and were of Type 1/8".

—I cu a> •H W C H CU m 8 CM § ^ o o ^^ (lbs Load o 1 o i o 8 o o <\J o o cu M •H Cl.c •H CO M 4-) CO CO C •H 4J •H 60 C O > 3 c •H CO CO O u 4J CO o S2 C CD E •H O CU a CO 4-> W H . 13 .

GO 4-) 0 CM 1 CO H QJ i—i •iH QJ ^0 f-t 1 CO C H 1 1 QJ 1 1 1 1 1 1 1 1 1 1 1 1 1 i I 1 1 1 1 1 \ j / 1 1 1 1 o (0 1-1 QJ 0J rl H iH 1-» * 0) <D N (0 in * (H (M iH U bO Cx.C H CO U rO CO 1 4-> in QJ CO u -0 1 QJ > CO H •0 ID 1 C co u CO > CO 0 in i o h4 u/ i—i 0 5 t 1 c •H CO 5-1 z c a) o 0 (0 1 •P (Z) e *rn o QJ a. T3 CO CO .0 r-H O 14 .

H CO c CO CO c •H T3 U •H GO C O > CO O CO CN u 4J CO Z <u o C E •H O CU CX CO 4-) CO <u H CO C H (U 0) u H 13 co w Xi Oi-h -J — 15 .

CN 0 1 o ui m 3> c O C 0) • o o t 1 03 u 4-J •H go e O 0) • o o (0 1 CO 4-J CO H •o (M 1 cu H •H CD CO C •o 1 o H 00 CD a) 1 1 I 1 1 1 1 1 1 1 1 1 1 I 1 1 i I 1 1 i 1 1 I o o CO o (M o o wt o o o 0) o o CD o o N o (0 in o o * o m o o (M \» o o o o I 5-1 1 P 60 •H J-.H 5-1 c 03 4-> rO 0) 1 CO CD CO 5-1 0) o o CD 1 > CO C 03 0 1 H > T3 03 5-1 • CO O 0 (0 1 O h-3 ^"s. «H HD 03 CO XI <-< O 16 .

c •H cO U en co c •H TJ 4J c O CO > CO O 5 CO C •H <0 u Z C <D o E •H O 0) a. co 4J CO H a) a) .—I H cu co C H o o 0) o o oo o o N o o (0 O o in o o + o o <n o o CM O o «* ON CD U H X! CO CO Cl4 00 JD r-l O 17 .

o o CD O o N O O O O U) o * o o (0 o o (M O O «H P T3 W 60 •H O H 18 .

utilized three as strain gages rather than one or The purpose of the two. With the lack the of a feasible method to ensure that each layer of reinforcing fiberglass cloth. other specimens. Since no bending problem was observed. fibers (filaments). addiof to tional strain was gage was to see if any bending the specimen occurring during the test due mis- alignment of either the specimen or the test fixture.Specimen No. glass yarns incurred some curvature during used With such a loose weave material as that for this project. a problem like this is difficult to completely avoid. on Figure 5. leaving fewer to carry the The second case is probably caused by the takeup of slack by some of the load carrying yarns. The result was a load/ strain 19 curve . one less longi- tudinal gage was used on subsequent tensile specimens. and in others the slope increases. longitudinal did the 1. The first case is most likely caused by the breaking of some of the individual glass load. it was very likely that some of the individual layup. in addition to of the individual remains straight and free any inplane curvature. In some cases the slope decreases. It will also be noted from the referenced load/ strain figure and later figures that some of the curves contain a knee joint where the line abruptly changes slope. yarns.

t . The average value for the y ^ transverse poisson's ratio was found to be 2/.. as E and again using ASTM Standard The test specimen was different dimensionally. however. = 2.586 x 10 psi. the contraction in the Y direction load. properties were found in "J/ the same manner D3039.134. value for the Y was E .513 x 10 psi.„ = 0.. The graphs of the load versus strain for these The average test specimens are shown in Figures 12-17. After determining a tensile modulus of elasti- city for each specimen and then averaging these values.165 Determination of E These two y and Z/. c the result was E x. yx * 20 .which had a sudden increase in slope after this initial slackness was taken up. axis modulus of elasticity in tension = 1.'J" The was (2) resulting TS average value for all of the specimens xy = 0.„ — yx — — . From the strain gage mounted transverse to the specimen's longitudinal axis. Figure 11 has approximate dimensions for these test specimens. was The also measured as a function of the applied poisson's ratio for each specimen was found by appli- cation of the formula ^xy .

~T H<r _L 4"6|" - Tabs ~~7~ 1m 1° Tensile Test Specimen Dimensions can of 00 Figure 11 Y is Axis There another method which used for determining composites. test specimen dimensions. 21 . flexure guide in test. it was what these also employed to see difference would in values would be observed those and if be greater or less than found by using the tension test. This second method utilizes a ASTM Standard D790-86 [4] was used as a Figure 18 gives the performing this test. the modulus of elasticity this reinforced produces Since method generally slightly different values.

-i u •H O 22 . CO 4-1 W H H H CN OJ <U w C T3 CO W XI .c •H CO u 4J CO CO C •H TJ P •H bO J c o 5 C •H co > CO O u CO o C Q) •H E O 0) O.

CO 4-> o o 1 cu CU t— •H c H ——— I I H —— I H CU o o If) o o (M O O 0) O o (0 o o CO CU T3 CO w . O hJ A <!- l*J <. *«-• o o <n i C •H co • M 4-J 2 C CU o CO o o (M i •H cu E a a.Q •H Dt4 O Jw rH 23 .h h t C •H o o (0 1 03 CO CU W CU 8 m i > CO H o o i C CO M • > T3 CO y— >.

C •H CO 4-) CO CO c •H 4-> •H CxO C o co > crj O o C <u •H E O CD a 00 4-) CO H CO QJ C H •H T3 03 co Si O -J — 24 <-< .

CO p CO H •H CO cu C H T3 CO CU un co O J2 i-H CU M Ml •H Ci. 25 .CO >-i 4-) CO cu CO 5-1 CU > CO C CO H CO 5-i > 5 c •H c« 5-1 CO O m z c CU P CO o 6 •H O CU a.

-n O 26 .i c •H CO U CO CO 5-1 c 4-) •H O -J co c > T3 CO O A ^D CO 4-> z c cu O CO •H <U E O Cu CO 4-) CO H i— CO 0) cu •H C H ^D CU <U M &0 •H T3 CO CO JD .

I o o N 1 c •H 03 U 4-> CO o (0 1 CO > CO C CO H CO M > CO O 5*N O CO u 4-1 o c CD CO 6 «H O <u a 00 4-J CO H r— CD cd H CO C H CD <D T3 CO CO O H 43 CO •H 27 .

b = ' 17 bd \ 3 m (3) where E.i T i— ± 7" is Figure 18 Flexure Test Specimen Dimensions A four point bend test was performed on each of these specimens (ASTM Method II). For the curves which reflect the load versus deflection for the E 24. with the center deflection indicator. test specimens. being measured with an Enco model #340 dial Figure 19 gives the details of the loading and support arrangement. see Figures 20- The equation used to calculate the modulus of elasticity in bending was E. . = modulus of elasticity in bending mode (in) (psi) L = support span distance m = slope of the tangent to the initial straight- line portion of the load-deflection curve b = width of specimen (in) 28 .

L. yielding a bending modulus of elasticity of E . = 2.513 x 10 6 psi) . find a value for the modulus These specimens were generally of the same as those for testing E in the X 29 dimensions direction. . = 2..368 x 10 psi. Load Test Specimen Steel Block Dial Indicator Figure 19 The Flexure Test Fixture Details E.2% of the value (E v found for the modulus of elasticity in tension X . This value is 94. The same to test was performed on three in test the Y specimens direction. values found for on the X axis test specimens were averaged.

I l c o •H 4-1 O i — 0) M-l Q > CO QJ co O (J c 0) •H E a cu a c/> 4J CO CU H M X cu CU I— S s ™ o T3 CO § H 30 § ? ™ o CN WD •H O .

'^ i c o •H 4-> O <u i—i 4-1 a . •H 4J S c e O cu O a aj a) i— 4-1 •H a 0) a.— a) • c •H w > T3 CO CO O i—l J oo O * "— X c O . 01 4-1 W H 5-i <D OJ 3 X 0) to <88?^888? CO o (M o QJ JD r-H to O 31 .

CO H cu a) m P X CD .i o •H c a a co c > CO m o J ON O o •I-l c 2 d cu O J-> o a> E •H t— U-l O Q CU a CO 4-> a.—1 Du SS?SoSS?w ~0 CO CO C\l CN CD !-i bO Cu Xi Or-* 32 .

4-1 CO 8 H CD r— fa CO ? CM © 8 T3 8 ? CNl ft >-i CO CO J2> i-H GO •H fa O 33 .s t( o r-l •U a cu r— 4-1 a /— (U • CO c •H > TJ CO CO O o hJ c: • o •H 2 C a> o o o CM o 0) r—l M-l E •H CJ a) Q a» a.

O J2 i-H 34 .1 1 i c o •H 4J O i — <D M-l a co Q) . .»> c •H > TD CO OO O J O • o •H 4-J Z c <u O o CU r— M-l •H 0) Q (U E O ex CO p CO H QJ <D J-i 3 X <1> 1— &c <r CNI § § ° (no T3 CO CO s s ? o CM ci 0) u 3 bO Dl.

[5] indicated that the modulus of elasticity in bending However. = 1. it the values found by tension test was deduced that the slightly higher value found for by the tension test must be reasonably stress which doesn't involve significant accurate bending.496 x 10 psi. value is 94.586 x 10 fc tension test method (E 6 psi). et al. the resulting g calculated to be E . it would is generally greater in composites. Again using Eguation specimen value This was (3) to find the modulus of elasticity of each and then finding the average. then it would which men is seem reasonable to assume that a flexure test creates both tension and compression in a speci- beam will result in a modulus of elasticity less than the modulus of elasticity in tension 35 which and . seem on that a lower value in bending could be the explained and of is basis that bending produces both in the test specimen.The plots of the load versus deflection for these specimens are indicated in Figures 25-27. tension compression If the modulus elasticity of the composite material in compression less than the modulus of elasticity in tension. An interesting sidenote is that Whitney.3% of the value found by the = 1. Since test the values found by using the flexure the method of were very close to being the same percentage method.

w .1 H 0) 1—1 14-1 c o *J o Q > c •r-l CO O on o csj c •H Z C O U O Q> 0) r-H M-l B •H a a) Q (U Ch CO 4-> w H >-i <u QJ 3 X Q) 1— to lO 88?™888? XI Csl o CM o aj 5-1 3 ao •H u.-H O 36 .

i 1 i o H a cu c i— Q CO CU c •H > CO CO O O CO H a i c o • 2 C a) •H" o 4J — cu 4-1 Q cu e a cu a. C/3 P CO H u X CU CU CU I— u. <o CM CU u 3 bO •H X) CO CO to OH 37 JD .

II c o •f-l 4-> o 0) I— Q > CO O C E o CD a. en 05 H i— a) TD CO r-H O 38 .

50" — 0. 1 - 0. particularly since the amount of bending in the box beam was small.156" Figure 28 Compression Test Specimen Dimensions 39 . E in Compression of guide used for determining the modulus in elasticity compression in both the X and Y the directions was ASTM Standard D695-85 [6]. test All of same specimens were approximately the dimen- sionally. bending induced interlaminar shear will tend to reduce the strength and modulus of elasticity of a laminate-type composite.H more than the modulus of elasticity in compression. Determination of E The . In summarizing. stresses Moreover. the higher modulus of elasti- city as determined by the uniaxial tension test method was deemed to be an accurate representation of the true modulus of elasticity of this FRP. Figure 28 gives the dimensions of these test Load specimens.

= 3. instead of the specimen. The modulus of elasticity in compression was calculated for each specimen by use of the equation where E = modulus of elasticity in compression (psi) P = load (lbs) 2 A = specimen cross-sectional area (in ) AL/L = change in specimen length divided by original length (in/in) The resulting load versus strain plots for each of the ten specimens used in this test are 29-38. and E These later = 4. were determined to be substantially below the empiri- cally determined values. the loading the crosshead movement was measured determine actual change in specimen length.68 x 10 5 psi. shown in in Figures The average modulus of elasticity compression values for each of the two principal axis directions were found to be E x c .Since the specimen length was so to small. and .55 x 10 psi. y c values appeared to be abnormally low. One obvious reason for these values the to be suspiciously which the below their actual values is damage by the occurred to the edges of the test specimens sawblade teeth during the process of 40 cutting .

O O

o m
to

CO

4J

CO

CO

o o
(0

>
CO

o

o
in (M

<->

<t

o
x

Ln

z
B H

o
C
<U

8
<M
CO

O
in

U
4J

O
cu

CO

a. CO
4J

o o

H
CO

<1)

C O
•H

o m

CO

CD

ex

o

E

8

O o
U)

O o *

o o
(0

o o
(M

O O

ON
csl

QJ

u
txO

T3
CO

CO

O

h

JD

•H Eh

41

CO

U
4-1

CO

CO

>
X!
CO

o
^D

z
c
oj

o

E
•H

O o
CO

CO

H

V
C o

•1-1

to CO ty

u
a.

u
I

o

B

H

«

1

I

1

1

1

h

o CO
a)
toO

8

8
•H

TD
CO

CO

,Q
r-H

O
-J

^
42

c
•H
CO
5-i

4-)

CO

>
TO

O

z
c b H

o

a
cu

a. en
4-1

w

H

O)

H

c o
CO CO CU

5-1

a,

o
CO

o

B

8 n
T3
TO CO

8 w

8 <«

0)
5-1

JO
i-t

bO •H

O

43

H CO c •P CO CO > CO O oo Z c e O o cu •H QJ CX CO CO H C o •H CO CO cu <u M a E O CN CO CD m H TO CO CO En ^3 i-H -J -^ O 44 .

03 4-> CO > 03 O ON o Z o C -8 C •H •-I a o CD P CO u a. o CO CD o E P •H "O 03 CO XI i-i O Jw 45 . CO w H CO CO CD a) c o •H a.

CO 5-i 4-) CO CO > CO O o CM o x 2 C E >H o 8« CO 5-1 cu a cu a CO 4-) 4J -8 CO CO H C o •H CD CO r-8 CO 0) U Pu O E m cu 5-1 •H T3 CO O •J 03 .£5 .-h ^ 46 .

CT3 4J co > CO O CNI O C CD 6 •H O <u a CO CO H C O CO CO CD 5-1 CD •H CX 6 o CO u So X) co r-H O 47 .

1 * CO M 4J C/3 CO > CO o CNl CM o !S C OJ B •H U cu a en 4-J CO H C O •H CO CO OJ J-l CU a u CO o e 8 n • 8 <>i 0) — co •r-l -o ca o hJ & r— ^ 48 .

8 n CO 4-J o R CO > CO J O -8 a o x CO CN z c •H o 0) E o cu CO a. CO 4-> u 8 M co H cu c o •H -8 CO CO cu u Cu o CO a) o B P TD CO CO GO •H 1X4 Xi O -J — r-H 49 .

c •H CO u CO M o Ifl > X> CO o o o CN1 o 8 * c CD E •H O cu &. GO -8 oo 4-! CO H 0) e o •H w o in en cu M o oo E O a m 0) M 00 T3 W <-< •H o 50 .

mentioned earlier. not a standard method. was ASTM Standard D 4255-83 [7]. The rails are subjected the of to a tensile force. considered inconse- quential. See Figure 39 for details of test specimen dimensions and Figure 40 for a view fixture apparatus. and the concurrent loss of some of the matrix near the edges. in Of the two available this standard. thickness reduction would not be significant for most tests. with the test specimen bolted between them. however. Determination of G and G The guide used in determining the inplane shear modulus methods chosen. it would have a substantial effect in a case where the compressive forces on each end of a test specimen are acting against a relatively small area which As had been further reduced by this cutting damage. The reason for this is that inplane shear test 51 . resulted in a reduction of the actual This test specimen thickness near the specimen edges.test specimens to the proper size. this test data was a result. It the test should be noted that the referenced ASTM that Standard contains a statement regarding the fact it is to be considered a standard guide. the This separation of as glass filaments from the matrix. Method A was It utilizes two pairs of rails.

I5 3 Figure 39 Shear Test Specimen Dimensions 52 .O.

Tensi le Fixture Figure 40 Shear Test Fixture 53 .

The the specimen size and constraints placed on it by clamping mechanism can affect the test results. but this one was selected of the simplicity of the test apparatus and so the ASTM Standards were used. After application of this eguation for calculating 54 each of . a three element On each specimen.) specimens normally fail by buckling out of plane. over. See Figures 41 and 42 for of the load versus strain for the on-axis (G xy and Figures 43 and 44 for the results of the off- axis (G-_J shear test. the with longitudinal axis test separate calculations were necessary to find This maximum shear maximum shear strain. More- no single shear test seems to have gained universal acceptance for determining the shear modulus A number of other shear test of laminated composites.-^^^. methods because that only are available. rectangular rosette strain gage was used to measure the strain. Since cate that the both tests produced results which indi- principal shear stress the axis of was the not coincident specimen. as indicated plots test on Figure 39. - t) ) 2 ( s i the diameter of Mohr's circle [8]. strain was found from the eguation y max which is - v (e.

on co CD &0 O c •H CO 4-1 CO en CO > CO O •J u CO H CD CO CD Xi C/} CO •H < I X o d GO •H fa -O CO CO Xi r-l O 55 .

jQ i-h 56 ..o < c •I-l CO CO •H CO 5-1 u in St. > -a co -8 CO o u u -J in -P W H 5-1 0) CO cu .% CO H X I w <c o CN1 c 8 H itnoiitN i ' i ' i ' <r i ' i ' i ' i i c cu 8 H 8 8 S 8 8 8 u &o •H fa X) w O CO .

M O CO •H CO P 5-1 > CO O •J 4J W H CO c/2 X <c I o CO O rH 57 .

on CD O CO c CO U 5-1 CO CO > CO O -J 4-J CO H CO (U <u X! CO 05 •H < I X 4-1 M-l O T3 CO CO O JD ^h 58 .

the following results were obtained: On-axis specimen: Off -axis specimen: X w max V iuttX = 3615/4 fc w/1000 lb load w/900 lb load = 1654yU£ The shear modulus was then calculated from the equations G = ^ .96 x 10 5 psi yx = 5. "£r = JL ( 6 ) .) L = length of specimen (in. ( 7 where = shear stress (psi) P = load (lbs) t = specimen thickness (in. the following results were obtained: G G = 2.) the principal strains.81 x 10 psi * The larger value for G is could be to expected cause a since a greater force required transverse than shearing of the glass fibers in the X-X direction would be required to produce shear between the matrix and the fibers in this direction.) After substituting in the proper values. and 7JT . TRANSFORMATION OF PROPERTIES Now that all of the on-axis elastic have properties fiber been found for each of the two 59 principal .

(8 = depending on layer) = 3. '12' " m4 n 2 2 2 n . ) m n-mn 3 3 where b = 11 X El E2 - 22 '12 = _ H E 2 1 '66 '12 S 16' S 26 = norma l coupling coefficients m = cos 9 n = sin 9 (8 =±10. depending on layer) ± 10.directions. This method uses .979 x 10 -7 xx J x 2. the power functions of sine© and cosine B being the angle of direction. 2m n 2m n 2 2 2 2 m2 n 2-| m m n 4m n 2 m2 n2 -m n 2 xx m n 4m n 3 2 2 2 3 + m 4^ n 4 2 YY \ ) 2 66 16 -8m n 3 2 2 3 m 2 r.. 3 3 3 Vs 26- 3 n -2m n 2(m n-mn 3 3.513 x 10° 60 . 2 (m -n ( 2 (8) xy ss 2m n -2mn n 2mn 2(mn -m n) mn -m n . method and of compliance transformation developed by [9] Tsai Hahn was used. found was the off -axis elastic properties must fiber ). be for the angle at which the reinforcement the rotated in the box beam (±10 To do this. The rotation from the principal (X) axis process is representated in matrix form as 11 22.

566 x 10 -8 1 u xy = 3.165 2.305 x 10 -7 E y E.378 x 10 -6 2. 1. top view (not to scale) Figure 45 Fiberglass Reinforcement and Axes Orientations After substituting necessary in the proper the values and performing the operations.513 x 10 b J = -6. or E or E 1 2 = 2.1_ yy xy _ ss = 6.96 x 10 Figure 45 illustrates the axes orientations of the fiberglass reinforcement and the box beam.3053 x 10 -7 61 .129 x 10 6 psi = 1.453 x 10 6 S S .6978 x 10~ 7 6. following values were found: S n 22 12 = = 7 4. X ^ Y 2 1 Box Beam.586 x x 10' 0. psi = -1.8835 x 10~ .

963 x 10" for -10° rotation) = 7 7 (-3.667 x 10" for -10° rotation) 3.041 x 10 psi This transverse the shear modulus is the one which shear in would the determine beam deflection due to vertical web or sides of the beam. trans- posing the values of S and xx G xy in S ss .-3. of the composite box beam due known and to plot this deflection versus load to deter- mine whether If the beam behaved in a linear elastic of manner.963 x 10" 7 (+3. then the principle superposition The could have been utilized in the analysis. found to be true.1185 x 10~ .207 x 10 5 psi S S 16 26 7 = .167 x 10" In G-- the box beam analysis. or G 12 = 3. as the resistance of the horizontal top and bottom would be negligible.S 66 = 6 3. machine as indicated The beam's center 62 deflection . is matrix unchanged. test beam was loaded with the Riehle in Figure 46. substituting G for yx yy unchanged since S and leaving S must xy yx The fourth order transformation egual S for symmetry. EXPERIMENTAL DETERMINATION OF BEAM DEFLECTIONS The next step in the analysis was to find to a the deflection load. was also calculated. the transverse shear modulus It was determined by S .655 x 10" or G 91 = 6. with the resulting operation yielding the following value: S ' = 6 5 1.

l i LO CO i o 1 txO .596*0 C •H cd l& o •H •H CO f CQ c o e CO OJ C O CN o> CN U CO O P-i o L_TJ P-I _l 03 o LO oo P O o CO oo oo o o oo o CO oo o CO O ^> CO C O •H •U •H CO o P-I CO i— i— C -o in r-. in c d) DQ E CO 0) CQ cn<r o X O CQ 0) T3 O CO CO <r 0) M cvO 6 cd •H 0) CQ €*-* \W\\\\\WA \V\\\WW\\\ 63 .

bending tests in each position. simply supported center loaded beam was an end loaded cantilever beam. the beam the did a linear elastic manner within applied load region. several minutes were required for the beam to recover the imposed deflecton after the load was removed.163" to 0. it was loaded in both directions as indicated by positions "A" and "B" in Figure 46.001" with the Enco dial load in indicator mentioned previously. However. the deflections varied much from one test to another to be of any use. Since the beam was unsymmetric. 64 . loading To this as 46. although for the cantilever model 75 the equivalent end load would be deflection. simplify the solutions to this problem. As this plot indicates. lbs to produce the same Although always behaving in Hookean too fashion (linear elastic).176".was measured to the nearest . A plot of the versus deflection is given in Figure 47 for the beam Position behave in "A". In all computations. the measured In six deflec- tions ranged from 0. as shown in modelled Figure the center deflection for a 150 lb load is used. in Position The load versus deflection plot for "B" is shown in Figure 48.

c o •H U O CD I— u-\ Q CO > CO O 6 CO cu CQ o 03 X U P •H T3 CO CO JD r-l O 65 .I c o •r-l •H CO O PL.

I i cq C o •H 4-1 •H CO O C O i c o •H 4-> o •H 4-1 O d) — <U <4-l i— LH Q > CD Q <U 05 o s CU DQ o CQ X 00 <t OJ 5-1 H en t3 CO O h 66 .

h .t /2)] 2 2 = (10) The unknown h is the distance from the bottom in surface. 67 the value of .750 t (h . was the for obtained location terms of h. The method which was used employs the principle that the sum of the first moments multiplied by their and respective modulii of elasticity for the tension compression This can be portions of the beam must equal zero [10]. compression. E rearranging in of this equation.965 . this ( 9 By using the appropriate known equation becomes 2.750(. written as E t /y dA + E c yy dA = values.t /2)] + . to the neutral axis. each of the two bending modes. The value and the neutral axis. t- is the upper wall By thickness while t.085 h 2 6 [ 2 + 2. a different approach had to be employed.h) 1 + 2.965 . However.129 x 10 E [. will be different However.085( .) THEORETICAL DETERMINATION OF COMPRESSIVE MODULUS OF ELASTICITY The original and plan involved the use to of beam the deflections strain energy equations find compressive modulus of elasticity. the value of of h.is the lower wall thickness. since the measured deflections varied excessively from one test to another.

452" c = 2.060 x 10 6 psi Each of these two positions represents of the box one of the 46 bending configurations beam. This value also indicates that the previous assumption the was correct regarding the abnormally low value of of elasticity in compression found by modulus experi- mental methods. correct values were: value can be assumed to have been found. 68 .7% of the experimentally determined value for the tensile modulus of elasticity. It's value for can then be found by using an iterative process each bending mode and varying the value of h. which is a reasonable proportion for a fiber-reinforced composite. of E c The and h which were found in this manner Position "A": E h = 0. value is 96.057 x 10 6 psi Position "B": E h = 0. this addition. values of E When the two then the are very close to being equal.E c should be the same in each case. The difference in E In between these two values is only 0.15%. Figure c illustrated the two positions.526" c = 2.

. and h. surfaces at the support of the cantilever their values will be maximized. Theoretical Beam Stresses Compression Tension 3233 psi Position "A" Position "B" 3743 psi 3209 psi 69 3764 psi . previously deter- the stresses can be calculated at the upper and of the beam. where Prior the to performing any strain be calculations. mined. ( in 4 ) By using the values of h. These are determined beams: or t by the following two formulas for composite = My v EtI t + E I C C \r . theoretical strains can be readily The calculated strains to by be using the appropriate will be formulas. stresses must first calculated. determined those maximum compressive and tensile strains on the lower and upper beam. (12) where M = moment (lb. and E. the previously determined values the of E.THEORETICAL BEAM STRESSES AND STRAINS From E . lower surfaces The results are listed in Table 1. Table 1.in) y = distance from neutral axis (in) I = moment of inertia E . or <^ = M Et I / t Ec+ EC T C I (11).

none of of Since the was loaded in both positions. Since desired. be located in the center the gages beam. 2.From these stresses. as shown in this figure. length. the strain at the center of the beam is the measured strain values must be multiplied 70 . etc. on beam's surfaces by the use The gages are of the same type. although they weren't necessarily all on the same side of the beam's center line. Theoretical Beam Strains Compression (M £ ) Tension IM e ) Position "A" 1819 1519 Position "B" 1558 1768 EXPERIMENTAL BEAM STRAIN The next step involved evaluating the the of actual strain strain gages. are found by using the equation 6 The strains = —2E — . strains were found and are listed in Table Table 2. could the actual be in where stresses actual and strains would maximized. the strains could be easily calculated since uniaxial stress were assumed. By using the this equation and the stresses just calculated. Their locations are depicted Figure 49. as those the used beam previously on the test specimens.

L - — 1.075" ©— r— I- Side (Not to scale) 2 Figure 49 Box Beam Strain Gage Locations 71 .L ©^p- 1 - 45 " . 44" 0.| i W ©— m LP) MT O CM C.73" 00 CO LPl J* I i— CO Side 1 rc.p-65: 1 0.40"— -HO.

.38" 1.031 1.147 1.75 10.53" 0. 2 produced erroneous readings.32" 0.127 Strain Gage No.046 1.033 1. by an appropriate correction factor to get the extrap- olated strain value on the beam centerline.34" 0.154 1. The factors were determined according to the eguation R = 10.052 1. The actual strain values and the centerline converted strain values for each strain gage are given in Table 4 for bending in Position "A" and in Table 5 for bending in Position "B" 72 .435" 1.75 .475" 1. Table 3 Strain Gage Multiplying Factors a Gage No 1 3 4 5 R 6 7 8 0.21" 1. so they Plots of the load versus strain for each strain gage and for each bending mode are given in Figures 50-5 5. Since the outer stress and strain vary linearly from zero at the end to a maximum value in the center.a (13) where R = multiplying factor (dimensionless) a = gage distance from centerline (in) Each of the factors which were determined 3 from this formula are listed in Table below. were discarded. the factors were determined solely on the distances shown in Figure 49.

m CO bO O c •H CO TO M 4-1 CO CO > CO J O d o •H •P •H CO O e CO 0) oq o DQ coS?SoSS?S T3 CO CO O lO CU M 00 •r-l £} i-H CL4 O 73 .

o d CO >-l CO •H •U CO CO > X) CO o d o •H CO o e CO cu CQ X O CQ lO 0) u •H X) CO CO O J2 r-t 74 .

B CO cu CQ o CQ X CN LO SS?SoSS?S T3 CO 0) u •H CO XI O H 75 .00 v£> lO CO cu W3 O c •H CO co U •P en CO > CO O o •I-l c 4J •H CO O PL.

m CO d> DO U C •H 03 V-i 4-) en > 03 O -J DQ C O •H P •H CO o PL. 03 0) CQ o CQ X CO 8S?c3o8cD?r3 •H -O 03 CO X> i-t O 76 .

Q) CO O C •H CO 5-1 •P in > CO O cq o •H 4-> •H CO O E CO Q) CQ O aq X 5-1 •H fa CO X3 r-H O 77 .

00 o lO cu bD O C CO CO •H u 4J c/j CO > CO O CQ o CO C O 6 CO d) OQ o CQ X m in cu &0 T5 CO CO ta XI r-H O -J ^ 78 .

only one side of the beam's longitudinal axis needs to be shown.031 1.046 1.127 2002 2231 2082 -1841 -1565 -1578 -1139 These strains are plotted in Figures 56 and 57.147 1.154 1.147 1.033 1.154 1.046 1. 4.031 1. Due to symmetry. 5 i.052 1.052 1. SHEAR LAG EFFECT ON BEAM STRAIN MEASUREMENTS In the ideal case.033 1. Strains for Position "A" Actual Strain R Centerline Strain (tf£) (MO 1 3 4 5 6 7 8 -1781 -2375 -1913 -1875 1355 1333 1026 1. Strains for Position "B" Actual Strain (A"0 R Centerline Strain (/it) 1 3 4 5 6 7 8 1903 2160 2019 -1760 -1364 -1368 -1011 1. normal a strains across the width of recin tangular this beam subjected to a load such as the one of analysis would not vary across the width 79 the .127 -1874 -2453 -1972 1961 1554 1538 1156 Table Gage No. and the one which the is presumed stresses by and the elementary beam theory.Table Gage No.

8 H e T3 0) 6 O d o a CO o u • c c •H CU o C O >H 4-> CD 6 ccj a) DQ in CO + r* + + + +0 N o CJ «+-l o cu o a) =*= bO O o in CO E o • CO o U c CO > C •H CO 5-1 +^ + + + + in eg 4J 4-) CO CO e CO <D CO X O CQ 4o o o (\J H h I O O O (M o o <D 8 o o o o (M o o o m U 01 03 o •H co w 80 .C •H CO c •H CO P CO M c O in 5-1 -U O •H CO CO CU CO <l- + + + o •H C M Pu oo CO C + + + + + + + + CO 0^ ..

C •H CO c M 4-) o in • •H CTS 00 U 4-J c o CO CO CD CO d o •H CO M e in (M c a ii j ™ CQ H II CD o CJ d o •t-i ID © •H CO o ^ o c 6 CO o T3 C CD cjO ^w CD CD o •H hJ p CO CQ 10 h» ^ o o CD m W [* J o o a) &o b0 TD o CO o e 10 *£ CO o o ^ CD > C •H CO a C ca P 00 U 10 (M . 4J CO E CO O Q -H CD CQ X o CQ o CD U bO •H 81 .

To This can be seen on Figures 59 and 60. the stresses (and strains) will the vary from a maximum on each side to a minimum in middle of the beam. stress in the vertical side walls of the box Figure 58 Since the Shear Lag loaded. Shear lag is caused by the fact that the normal stresses in both the and upper lower wall sections are transmitted to them by the shear beam. shear lag phenomenon will be non-existent in this area this but will gradually increase as the distance from area increases. however. middle of a center simply supported beam and the area where a cantilevered beam the is fixed cannot experience this shear deformation. This phenomenon is known as shear lag.beam. In reality. as shown in Figure 58. 82 brittle coating . the presence of shear lag a verify and its effect on the strain gage readings.

. -„ .sy. . i )' „ jjr ii |.. „ —"*»"r* > i a <>.. - .-„.. n - > r 'i c •H 4-1 CO u O CD i-H 4-1 4J •H M cq -a •H CT> m SB ~< ... - v . ! — — a> *^j M H bO fa tftimmmmjii H ( | i^f V y»"{ f #' — 5" v*-» ... :.. I . i -_ * -• - • 13F . .. . . .s^ ..

I 60 C •H 4J CO O .— o CD 4-1 *J •H M CQ CM CD TJ •H C/3 O ^D CD T^tE ' .—^r"—-r^ * •H SK* i* ^^l&fw&i 84 .-^' 5-1 i- i r . - ' — .

a felt tip pen was used to mark some of the stress lines. where shear lag would not be present. first step consisted of applying an aluminum undercoating. the this phenomenon gradually increases until isostatic lines look like arcs of a circle at some distance from the middle of the beam. Tens-lac Type UN-10A. and then applying the brittle coating. except that no quantitative data was recorded. Inc. The curves shown represent isostatic As already which tioned. crack lines. its This was done after all of the other testing The had been completed. After allowed to the beam was loaded in the same manner as before. are lines of constant stress. dry. 59 shows side 1 of the Figure 2. which was Tens-lac Type TL-500-75A lacquer. As the distance from the center of the beam increases. both of which are products of Measurements Group. menthe these lines are generally straight in middle of the actual beam.analysis was employed to get a qualitative look at effect. 85 . beam and Figure 60 shows side Since the cracks were too small to be easily visible.

it can be seen that lines have been drawn which represent an approximation to the average strain gage readings which were recorded during beam loading.0% 2.7% 86 . gage readings on the be appeared low or high based brittle coating These lines of can't the considered average accurate representations true strain some across the width of the beam.5% € ) ) ) 1519 1819 1768 1558 1600 1900 2050 1600 B (Tens. Comparison of Strain Values Average of Strain Gage Measurements (yU £ ) Loading Position A (Tens. were However.3% 4.Chapter III RESULTS AND CONCLUSIONS RESULTS Upon looking again at Figures 56 and 57.) B (Comp. The resulting values are summarized in Table Table 6. the the lines adjusted to be based upon whether somewhat analysis. idea of the approximate but do provide average 6 value. Theoretical Strain {JUL (% of Difference Theoretical) 5. A (Comp. below.) 16.

The dispersion of the strain gage readings can the most beam's likely be attributed to imperfections in As wall due to the hand layup process. so many of the problems associated eliminated. these are very difficult to avoid in a hand layup procedure where the wet resin- impregnated cloth must be handled. Another project was the problem which was encountered in this difficulty for in obtaining in reliable the experimental values use determining compressive modulus of elasticity. men- tioned in the Introduction. need to be utilized. Other methods would as the ASTM Standard used in this analysis results certainly did not produce satisfactory CONCLUSIONS Finally. there is generally good agreement. with handling a sticky material are These materials are known as pre-pregs. resin has also been partially cured (B-staged) by the manufacturer. except in Loading Position "B" on the tension side of the beam. be These problems can in avoided have by using a form of fibers which by the the fibers been The impregnated with resin manufacturer. the the results of this research indicate of using experimental 87 necessity techniques when .. As can be seen from the data in this table.

using a composite material which has been fabricated by hand known. the brittle coating analysis proved to be very helpful in locating anomalies in the walls. fabricated Since readily observed in the walls of this box beam were thin and imperfections in them were present. theoretical analyses composite Since the properties of FRP are dependent upon a number of factors. or one for which the physical properties aren't was this The problems with using a material which were manually project. 88 . which adversley affected the accuracy of the strain gage readings. which contains measures Any additional which provide information critical to the analysis should also be utilized. ever it is important to use what- techniques in are available to obtain the the elastic in a properties structure directions of the composite can stresses materials. This research also suggests that reasonably good comparisons can be obtained between and experimental of technigues materials.

New Jersey. San Francisco. Connecticut. 2.. Annual Book of ASTM Standards. Lekhnitski.. Mechanics of PWS Publishers. . Volume 15. 1963. Massachusetts. 1988 Annual Book of ASTM Standards. S. . 199-204. S. California. Materials 2nd. 280-89.P. . Whitney. Volume 8. Mechanics of Composite Materials Scripta Book Company. Dally. Chapter IV SELECTED BIBLIOGRAPHY 1. W. R. Daniel. Experimental Stress McGraw-Hill Book Company.. 249-50.T. 9. Gere. edition. R. Cliffs. H. Theory of Elasticity of an Aniso tropic Body Holden-Day.C. 1982. 10. 6. . Washington..M. S.B.W. pp.. J. pp 53-56. 1988 pp. 1975. pp. 88-91. .M. . and Timoshenko.M. Introduction to and Hahn.F. pp. pp.G. 1988 Annual Book of ASTM Standards.. 195-204. J. and Pipes. J..W. 4.M..01. . . Volume 15. 1984. Boston. Volume 8. 5. 8. Composite Materials Technomic Publishing Company.01. 1988 Annual Book of ASTM Standards. I. Westport. Mechanics of Fiber Reinforced Experimental Composite Materials Society for Experimental Englewood Stress Analysis. 7. .. 1978. 3.03. pp.03. D. Tsai. 117-121. pp. and Riley. 1980. . 89 . Jones. Analysis 321-23.

u Gage #2 454 814 1154 1495 1830 2208 2598 3000 3368 3770 4152 4554 4932 90 e ) Gage # 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 285 654 1023 1424 1848 2242 2608 3008 3402 3798 4190 4620 5028 . 2 Load (lbs) Gage #1 e -54 119 187 255 324 404 442 530 568 652 688 771 820 Strain (. Specimen No. X Axis Tension Test . Specimen No. X Axis Tension Test. Gage 1 Gage #1 -65 -128 -200 -251 -300 -368 -407 -472 -513 -562 -602 -668 -714 Strain (M 6) Gage #2 Gage #3 360 716 1087 1500 1860 2260 2637 3020 3415 3720 4450 4514 4820 # 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 418 778 1180 1600 2004 2340 2890 3285 3693 4144 4630 5010 5382 260 548 834 1089 1412 1736 2090 2448 2800 3160 3590 3940 4300 Table 8.Chapter V APPENDIX Table Load (lbs) 7 .

3 Load (lbs) Gage #1 316 632 930 1272 1600 1943 2268 2610 Strain (/x Gage #3 -65 -122 -176 -227 -278 -346 -378 -452 t ) Gage #4 310 644 1022 1410 1800 2172 2574 2940 100 200 300 400 500 600 700 800 Table 10. Spe Load (lbs) Strain iM € Gage #1 Gage #2 ) 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 257 502 756 1054 1362 1687 2069 2417 2804 3163 3534 3950 4326 4733 5117 -43 -83 -117 -167 -204 -239 -273 -304 -347 -381 -420 -458 -492 -533 -562 91 . Specimen No.Table 9. Y Axis Tension Test. X Axis Tension Test.

Specimen No. Y Axis Tension Test. 6 (lbs) Strain (yU€ Gage #2 Gage #1 ) 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 348 652 976 1290 1624 2018 2390 2742 3141 3558 3950 4351 4792 5176 5626 -41 -93 -130 -164 -208 -248 -278 -321 -368 -411 -458 -500 -558 -593 -646 92 . 5 Load (lbs) Strain IMZ Gage #1 Gage #2 ) 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 244 562 941 1340 1747 2164 2581 2964 3318 3598 4112 4498 4934 5302 5728 -26 -53 -85 -126 -158 -188 -213 -241 -263 -282 -302 -323 -337 -354 -369 Table 12. Load Y Axis Tension Test.Table 11. Specimen No.

X Axis Flexure Test.046 .463 Table 14.239 .145 .190 . Specimen No.Table 13.096 .395 100 120 140 160 180 0.042 . Specimen No. 9 Load (lbs) 20 40 60 80 Deflection (in) 0.237 . 8 Load (lbs) 20 40 60 80 Deflection (in) 0. X Axis Flexure Test.182 .488 Table 15.395 100 120 140 160 180 0.334 .089 .274 .455 . Specimen No.047 .280 .330 .283 .393 100 120 140 160 180 93 0. 7 Load (lbs) 20 40 60 80 Deflection (in) 0. X Axis Flexure Test.338 .144 .235 .135 .096 .191 .

592 94 .179 . 10 Load (lbs) 20 40 60 80 Deflection (in) 0.333 .187 .192 .143 . Y Axis Flexure Test.046 .278 .237 .298 .056 .386 100 120 140 160 180 0.Table 16.121 .282 . Load (lbs) 20 40 60 80 Deflection (in) 0.147 .456 Table 18.514 100 120 140 160 180 0.098 .331 .363 .238 . X Axis Flexure Test. X Axis Flexure Test.434 .095 .231 .047 . Load (lbs) 20 40 60 80 Deflection (in) 0.391 100 120 140 160 180 0. Specimen No.464 Table 17.

020 .190 .021 0. Specimen No.016 . 15 Load (lbs) Deflection (in) Strain (AL/L) 100 200 300 400 500 0. Y Axis Flexure Test.033 .031 .320 .034 0. X Axis Compression Test.022 0.472 100 120 140 160 0.019 .320 .476 .572 Table 21.Table 19.654 Table 20.256 .193 .256 . Load (lbs) 20 40 60 80 Deflection (in) 0.565 0. Y Axis Flexure Test.396 . Load (lbs) 20 40 60 80 100 120 140 160 180 Deflection (in) 0.060 .036 95 .062 .125 .400 .129 .026 .

009 .0165 .003 .0126 .0024 .0115 .0204 . X Axis Compression Test.007 .0067 .0149 .013 96 0.014 100 150 200 250 300 350 400 450 500 600 0. Specimen No.0081 .010 .0180 .002 .0162 .006 .011 0.0189 100 150 200 250 300 350 400 450 500 600 0.012 .012 0.013 .0209 .0120 .011 .0094 .011 .0188 .0239 0.011 .0099 .Table 22.010 .008 .007 .010 .0222 . 17 Load (lbs) 50 Deflection (in) Strain (AL/L) 0.0066 .015 0.0134 .0179 . 18 Load (lbs) 50 Deflection (in) Strain (AL/L) 0.0047 . Specime: Load (lbs) 50 Deflection (in) Strain (AL/L) 0.004 .0148 .004 .0138 . Specimen No. X Axis Compression Test.006 .0205 .011 .0145 .0198 Table 23.012 .007 .0097 .0183 100 150 200 250 300 350 400 500 0.0256 Table 24.006 .0165 .009 .009 . X Axis Compression Test.012 0.005 .

013 .004 .014 .005 .005 .007 . Specimen No.009 100 150 200 250 300 350 400 450 500 600 0. Load X Axis Compression Test.007 .0162 .0101 .011 .0226 100 150 200 250 300 350 400 450 500 600 0.010 0.0135 .005 .009 .0058 .0284 97 .0049 .0118 .0153 0.0119 .0177 .0147 . 19 Deflection (in) Strain (AL/L) (lbs) 50 0.008 .007 .008 .009 .0082 .0131 .0162 Table 26.015 0.019 0. Specimen No.003 . 20 Load (lbs) 50 Deflection (in) Strain (AL/L) 0.0116 .006 .Table 25.0087 .0099 .0201 . Y Axis Compression Test.0134 .012 .008 .0075 .0216 .

0038 .005 .0069 .0183 .0110 .0189 100 150 200 250 300 350 400 450 500 600 0.0177 98 .0126 .006 .012 .008 .0054 .011 . Specimen No.006 .Table 27.0110 .0096 .010 .010 0.0075 .007 .008 .012 .006 .014 0. 22 Load (lbs) 50 Deflection (in) Strain (AL/L) 0.0084 .003 . Y Axis Compression Test. Specimen No.0096 .0031 .005 .004 .0137 .0213 Table 28.0156 .002 . 21 Load (lbs) 50 Deflection (in) Strain (AL/L) 0. Y Axis Compression Test.0123 -- 100 150 200 250 300 350 400 450 500 0.009 .007 .009 .012 0.0169 .0138 0.

23 Load (lbs) 50 100 150 200 250 300 350 400 450 500 600 Deflection (in) Strain (AL/L) 0.007 .0078 . Y Axis Compression Test.005 .003 .0048 .0142 .0108 .006 .0137 Table 30.007 .0093 .0149 0.003 .0085 .008 .0101 .0062 . Y Axis Compression Test.011 99 .009 0.005 . Specimen No.009 .006 .002 .0112 .0100 .010 0.006 . Specimen No.Table 29.0034 . 24 Load (lbs) 50 100 150 200 250 300 350 400 450 500 600 Deflection (in) Strain (AL/L) 0.0120 0.004 .0132 .007 .007 .001 .0163 0.008 0.0075 .0018 .0039 .0091 .0062 .004 .0124 .009 .

Load (lbs) On-Axis Shear Test Gage #1 -198 -371 -587 -770 -1010 -1190 -1440 -1628 -1904 -2078 -2396 -2586 -2922 -3142 -3514 Strain (>i Gage #2 15 37 60 85 € Gage # 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 108 135 160 184 211 237 262 291 318 350 376 172 317 466 633 784 978 1138 1329 1505 1715 1924 2162 2341 2636 2820 Table 32.) Table 31. Load (lbs) Off-Axis Shear Test Strain (ju £ Gage #2 2 6 ) Gage #1 83 Gage # 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 179 278 378 480 584 692 798 911 1018 1136 1260 1366 1508 1640 7 8 9 9 10 11 12 13 14 15 16 17 20 -63 -155 -238 -335 -418 -528 -612 -704 -800 -902 -1024 -1125 -1232 -1336 -1450 100 .

071 .0495 .114 . Deflection Load (lbs) 20 40 60 80 100 120 140 180 Position "A" Deflection (in) Position "B" Deflection (in) 0.) Gage #1 Gage #3 Gage #4 Gage #5 Gage #6 Gage #8 -250 -480 -710 -950 •1185 •1665 •2140 -334 -664 -975 1278 •1570 •2135 100 140 180 2700 -265 -524 -782 -1042 -1300 -1800 -2290 245 490 738 998 1250 1762 2268 195 370 548 730 915 1280 1640 138 266 400 550 690 970 1250 Table 35.144 0.024 .0735 .) Table 33.0965 .092 . Box Beam Strains.048 . Box Beam Load vs.157 0.025 . Position "A" Strain (/*.120 . Load (lbs) 20 40 60 80 Box Beam Strains.199 Table 34.167 0. Position "B" Load (lbs) 20 40 60 80 Strain {ju. €. e Gage #1 Gage #3 Gage #4 Gage #5 Gage #6 Gage #8 272 539 785 1035 1295 1545 1800 2062 2320 320 616 895 1180 1474 1764 2056 2352 2650 278 550 802 1070 1338 1610 1890 2168 2448 -268 -520 -754 -992 •1225 •1460 •1692 •1925 •2154 100 120 140 160 180 -178 -362 -538 -718 -900 1080 1270 1455 1638 -150 -290 -420 -552 -690 -825 -962 -1100 -1235 101 .

070 7.237 1.154 .550 6.05 On-axis Shear Off-axis Shear .565 6.760 .156 .158 .166 .670 .519 .162 .965 6.648 1.526 .529 .586 .636 .980 .02 3.982 .Table 36.157 .0 6. Test Spec imen Dimensions L (in) Test Specimen No.756 .978 .235 1.153 .651 .532 .649 .157 .065 0.157 .156 102 .166 .946 0.156 . 1 2 3 w (in) t (in) 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 9.974 .154 .159 .512 .156 .151 .760 0.982 .233 .503 0.608 .157 .503 .158 .542 6.938 6.0 9.505 3.978 .154 .0 9.489 .011 7.978 7.156 .930 6.610 .645 5.000 0.152 .158 .155 1.222 .158 6.930 0.475 .935 6.603 .157 .159 .

S. 1972 AN ABSTRACT OF A THESIS submitted in partial fulfillment of the requirements for the degree MASTER OF SCIENCE Department of Mechanical Engineering KANSAS STATE UNIVERSITY Manhattan.. Kansas 1989 . Kansas State University.COMPARISON OF A THEORETICAL ANALYSIS WITH AN EXPERIMENTAL STRESS ANALYSIS OF AN FRP THIN -WALL COMPOSITE BOX BEAM by CLARK TIMOTHY HARBAUGH B.

the For the box beam.ABSTRACT The analysis of stress and strain in box beams composed of an isotropic material However. analysis is not so simple. and the shear modulii. The elastic constants of the FRP were deterfrom mined experimentally utilizing test specimens cut the test specimen plate. is a relatively is elementary exercise. In this study. poisson's ratios. for the other half of the wall thickness. The elastic constants determined by tension test methods included the modulii of elasticity. the a thin-wall box beam and a test specimen plate were hand fabricated of a nonisotropic composite material known as fiberglass reinforced plastic (FRP). Both the box beam and the test specimen plate a utilized modified plain weave fiberglass material as the rein- forcement and a polyester resin as the matrix. . when the box beam fabricated from a nonisotropic composite material. The test specimen plate consisted of 16 layers of fiberglass reinforcing in cloth which each layer's X and Y axes were collinear. the primary axis of +10 fiberglass reinforcing material was rotated and -10 for half of the wall thickness.

in compression was found . but The modulus of however. results . elasticity by using an equi- librium equation.The elastic constants found by compression test methods were to be determined experimentally also. was considerable scatter in the strain but the overall average experimentally the determined strain was within 6% of theoretical strain in three out of four comparisons (the fourth was in error by 16%). were the then transformed using compliance method of Tsai and Hahn elastic These transformed to constants were and then used calculate the the theoretical stresses strains in box beam subjected to a simple load. Since the a box beam's walls contained some irreglarities. were deemed to be invalid. There gage data. Moreover. box beam were The actual strains in the found by loading the beam. The brittle coating analysis did the provide useful information for use in evaluating Overall. after strain gages had been installed. the experimental strain gage data. brittle coating analysis was used to provide additional qualitative information in analyzing the strain gage data. The on-axis elastic constants thus found to off-axis constants by . the effects of shear lag were evaluated with the resultant isostatic lines.

T. technigues.indicate the importance of using experimental methods when evaluating physical or stress and strain in a material with unknown properties and other which may to contain imprecise irregularities anomalies due In fabrication encountered. and Hahn. H.. Connecticut. . 88-91. . Introduction to S. the spite of the problems stress theoretical and experimental analyses were found to be in generally good agreement.W. 1980.. Composite Materials Westport. Tsai. pp.

. and Education: After graduation from Russell High School in Russell. VITA Clark Timothy Harbaugh Candidate for the Degree of Master of Science Thesis: COMPARISON OF A THEORETICAL ANALYSIS WITH AN EXPERIMENTAL STRESS ANALYSIS OF AN FRP THINWALL COMPOSITE BOX BEAM Major Field: Biographical: Mechanical Engineering Born at Great Personal Data: December 11.) Examinations in 1988.. 1972. where a Bachelor of Science in Mechanical Engineering was received in May. attended Fort Hays and then Kansas State UniState University. Ralph Kansas. Became a licensed Professional Engineer in the State of Kansas. Virginia L. son of 1947. R.E. Harbaugh. versity. Professional Organizations: Associate member of the American Society of Mechanical Engineers since 1972. 1989. the longest and most noteworthy being a contract assignment for Northern Natural Gas Company as both a project Mechanical Engineer and a Project Manager Professional Qualifications: Passed the Fundamentals of Engineering (EIT) and the Principles and Practices of Engineering (P. effective January 31. Bend. Kansas. Professional Experience: Have held a number of positions since 1972.