EXPERIMENTAL STUDY OF LAMINATED COMPOSITE

TUBES UNDER BENDING

by

PUNEET SAGGAR

Presented to the Faculty of the Graduate School of

The University of Texas at Arlington in Partial Fulfillment

of the Requirements

for the Degree of

MASTER OF SCIENCE IN MATERIALS SCIENCE AND ENGINEERING

THE UNIVERSITY OF TEXAS AT ARLINGTON

May 2007
Copyright © by Puneet Saggar 2007

All Rights Reserved

 ACKNOWLEDGEMENTS

From core of my heart I would like to thank Dr. Wen S. Chan for his expert

guidance and patience he displayed during the time period of my stay with him. His

unique way of teaching was very precious to me and that’s what helped me in building

fundamentals in the field of composites.

I would like to thank Mr. Skip Pankewich and Mr. Mike Finn at United Sport

Technologies, Inc. for supplying specimens needed for this study.

I would also like to thank Dr. Pranesh B. Aswath for his help that I needed

regarding calibration and trouble shooting of the MTS machine.

I would like to extend my appreciation to Dr. Shashank Priya for serving as a

committee member.

I would like to thank my father, brother, Dr. Guna Selvaduray and Santosh for

the moral support they gave me all through in my Master’s.

My sincere thanks to all my friends- Ramoun, Vaneet and Chethana for their

much needed help they provided in various stages of research and documentation in this

thesis.

April 16, 2007

iii
 ABSTRACT

EXPERIMENTAL STUDY OF LAMINATED

COMPOSITE TUBES UNDER BENDING

Publication No. ______

Puneet Saggar, M.S

The University of Texas at Arlington, 2007

Supervising Professor: Dr. Wen S. Chan

Bending stiffness of laminated composite tubes was determined experimentally

using four-point bending test. The experimental results were compared with the results

obtained by the laminated plate and smear property approaches. The results indicate that

experimental values are closer to prediction of the laminated plate approach compared

to smear property approach. Effects of tube radius and stacking sequence and fiber

orientation of the walled thickness laminate on bending stiffness and strength of

composite tubes were studied.

Fracture analysis was conducted to investigate the failure process of the tube.

Both techniques of x-ray radiography and optical microscopy were used in this study. It

is found that damage is initiated at the loading point. The failure process depends on the
iv
fiber orientation and ply stacking sequence of the walled laminate of the tubes. The

tubes finally failed in compression. Fiber breakage and delamination were observed as

the prominent damage modes.

v
 TABLE OF CONTENTS

ACKNOWLEDGMENTS ......................................................................................... iii

ABSTRACT .............................................................................................................. iv

LIST OF ILLUSTRATIONS..................................................................................... ix

LIST OF TABLES..................................................................................................... xi

Chapter

1. INTRODUCTION AND LITERATURE REVIEW ..................................... 1

1.1 Composites and their advantages............................................................. 1

1.2 Applications of composite tube(s) ........................................................... 1

1.3 Fabrication of composite tubes................................................................ 2

1.4 Literature survey and previous work on composite tubes in bending ..... 3

1.5 Objective and outline of this study .......................................................... 5

2. EXPERIMENTAL PROGRAM.................................................................... 7

2.1 Description of Test Plan...................................................................... 7

2.2 Characterization of principal material properties (E1, E2, G12, ν12) ..... 7

2.2.1 Specimen preparation ............................................................... 8

2.2.2 Testing for basic material properties ........................................ 11

2.3 Four-point bends testing of composite tubes ...................................... 11

2.3.1 Geometry and lay-up of composite tubes ................................. 11

2.3.2 Fabrication of composite tube and inspection method ............. 13

vi
 2.3.3 Test Fixture design and test set up............................................ 14

2.3.4 Specimen preparation ............................................................... 16

2.3.5 Test procedure.......................................................................... 17

3. REVIEW OF ANALYTICAL MODELS FOR

BENDING STIFFNESS EVALUATION
OF COMPOSITE TUBES.............................................................................. 18

3.1 Geometry and loading of composite tube ........................................... 18

3.2 Lamination theory .............................................................................. 19

3.2.1 Ply stress-strain relationships in material coordinates ............ 19

3.2.2 Ply stress-strain relationships in laminate coordinates ............. 20

3.2.3 Laminate constitutive equations ............................................... 21

3.3 Laminated plate approach ................................................................... 22

3.4 MATHEMATICA Program................................................................ 25

4. EXPERIMENTAL RESULTS ...................................................................... 27

4.1 Test results and discussion of principal material properties ............... 27

4.2 Strength of composite tube and their load-deflection curves.............. 29

4.2.1 Sudden failure of composite tubes ........................................... 30

4.2.2 Gradual failure of composite tubes........................................... 31

4.2.3 Effect of radius ......................................................................... 32

4.2.4 Effect of 0˚ ply location............................................................ 33

4.2.5 Effect of fiber orientation of angle plies .................................. 35

4.3 Curvature effect on bending stiffness of composite tube.................... 36

4.3.1 Discussion for possible error in approaches ............................. 36

vii
 5. FAILURE INVESTIGATION........................................................................ 38

5.1 Failure Process and X-ray Radiography ............................................. 38

5.1.1 X-ray radiographic procedure................................................... 39

5.1.2 Analysis of x-ray images .......................................................... 39

5.2 Optical Microscopic Analysis............................................................. 44

5.2.1 Experimental Procedure............................................................ 45

5.2.2. Delamination failure in tubes................................................... 46

6. CONCLUSIONS……….. ............................................................................. 52

Appendix

A. MATHEMATICA CODE FOR STIFFNESS MATRICES… ...................... 55

REFERENCES .......................................................................................................... 59

BIOGRAPHICAL INFORMATION......................................................................... 61

viii
 LIST OF ILLUSTRATIONS

Figure Page

2.1 (a) 0° coupon (E1 and ν12), (b) 90° coupon (E2),
(c) 10° coupon (G12), (d) side view of coupons .............................................. 10

2.2 Geometry of the composite tube and section A-A΄......................................... 12

2.3 Typical set up for four point bending test ....................................................... 15

2.4 Base plate for the fixture ................................................................................. 16

2.5 Top plate attached to load cell………………………………………………...16

3.1 Infinitesimal section of composite tube .......................................................... 23

3.2 Flowchart for Bending Stiffness matrix computation of composite tube ....... 26

4.1 Failed 0˚, 90˚ and 10˚ off-axis coupons .......................................................... 28

4.2 Load vs. Displacement for sudden failure specimens ..................................... 30

4.3 Load vs. Displacement for 2C1 specimen....................................................... 31

4.4 Load vs. Displacement for 2B2 specimen....................................................... 32

4.5 Deflection vs. Load curves for various inner radii.......................................... 33

4.6 Deflection vs. Load data on number of zero-degree plies............................... 34

4.7 Effect of fiber orientation in symmetric lay-ups in Deflection vs. Load ........ 35

4.8 2B3 specimen under four-point bending......................................................... 37

5.1 X-ray image of specimen 2A1 ........................................................................ 40

5.2 X-ray image of specimen 2A2 ........................................................................ 41

ix
5.3 X-ray image of specimen 2A3 ........................................................................ 41

5.4 X-ray image of specimen 2B1......................................................................... 42

5.5 X-ray image of specimen 2B2......................................................................... 43

5.6 X-ray image of specimen 2B3......................................................................... 43

5.7 X-ray image of specimen 2C1......................................................................... 44

5.8 X-ray image of specimen 2C2......................................................................... 44

5.9 Specimen for Optical Microscopy investigation............................................. 45

5.10 Fractured tube specimen 2A1 at 150X magnification..................................... 46

5.11 Fractured sample 2A2 at 70X magnification .................................................. 47

5.12 Fractured sample 2A3 at 350X magnification ................................................ 47

5.13 Specimen 2A2 at 100X magnification showing

shear effect and failure between +45 and -45 plies......................................... 48

5.14 Specimen 2A3 at 300X magnification showing

disastrous shear failure in the middle............................................................. 49

5.15 Specimen 2A1 at 275X showing delamination and

crack-propagation in oblique manner from layer to layer as well................... 49

5.16 Crack propagation in 2A3 at 350X magnification .......................................... 50

5.17 Failed sample 2B1 at 200X magnification showing crack propagation.......... 50

5.18 Failed sample 2C1 at 100X showing delamination and

crack propagation from layer to layer as well ................................................. 51

x
 LIST OF TABLES

Table Page

2.1. Test Plan for determination of material properties.......................................... 8

2.2 Lay-up and radius of specimens...................................................................... 13

2.3 Effects studied by lay-ups ............................................................................... 13

4.1 Material property test data............................................................................... 28

4.2 Failure load and first damage load of tube specimens .................................... 29

4.3 Bending stiffness of various lay-ups of composite tubes................................ 36

xi
 CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

1.1 Composites and their advantages

A composite by general definition means a combination of two or more

materials combined together in such a way that it gives better properties than the

individual material system in it. Structural composites are a blend of two or more

components, one of which is stiff, long fibers and the other, a binder or ‘matrix’ which

holds the fibers in place. Fibers are usually much stronger than the matrix material.

When fibers and matrix are joined to form composite they retain their identities but both

directly influence composite’s final properties. Due to the high strength and load

bearing capacity possessed by fibers (reinforcement phase), composite usually exhibits

better mechanical properties in the direction of fibers. So a structural fiber reinforced

composite is a layered structure having fiber orientation in each layer designed to get

maximum benefit in different directions for desired application. These advantages are

not just tamed by aircraft industry but also in everyday use of golf shafts and tennis

rackets.

1.2 Applications of composite tube(s)

For the lucrative benefits of high specific strength and high specific stiffness

offered by fiber reinforced composites, they are put into wide variety of applications.

Until 1991 about 10% of the money invested in advanced composites used to come

1
from sports industry [1]. But slowly there has been active use of composites in most

industries and now the arena is not just limited to defense sector only. Composite tubes

found their application in golf shafts. Composites in rolled or tubular form are used by

printing industry a lot these days. Being light and strong the printer rollers retain their

strength but light weight gives them benefit in print quality than steel rollers and also

lesser vibrations. Composite tubes with uniform circular cross-section or tapered cross-

sections find use in drive shafts because of extra stiffness, they get higher whirling

speeds. Very stiff shafts for lathe, very strong for trucks and well insulated shafts for

trolley buses and generators are some applications where composite tubes are put into.

Tubes with square cross-section are being used for robotic applications where light

weight of robot enhances its performance speed and extra stiffness adds to its

mechanical properties.

1.3 Fabrication of composite tubes

Composite tubes can be fabricated in several ways, namely- filament winding,

pre-preg rolling (sheet wrapping) and pultrusion. These techniques can be utilized either

directly or indirectly to aid in tube construction. Choice of technique is dominated by

properties required, cost and quantity. Of the four types of methods listed, filament

winding is the most common for tubes.

Pre-preg rolling is the most popular method. In this technique, layers of material

are rolled around a mandrel, by hand or machine, prior to consolidation or cure. This

process is ideal for smaller quantities and smaller tubes where increased material costs

2
are less significant. This technique is adopted in this study and will be explained in

more detail in chapter 2.

Pultrusion is a continuous process in which fibers in form of roving, mat or

fabric are impregnated with resin and pulled through a heated die of the required shape,

molding through the inside and outside diameters. This technique uses raw materials in

their most basic forms and lowest cost forms. Due to high tooling cost, large quantities

need to be produced to make this process cost effective. The main drawback of this

method is the choice of fiber angle is very limited.

Circular cross-sections are typical of filament winding with an added advantage

of several fiber angles and shapes. On the other hand, it can create high fiber volume

fraction parts with high quality as well. In this process, fiber bundles, after impregnating

with resin, are wound on mandrel ate the angle required to produce the mechanical

properties as well. Fibers can be positioned at any angle within the tube, with different

layers at different angles to carry the various internal and external loads applied. Tubes

are seldom made of pure 0˚ or pure 90˚ fibers as they would easily split. Tubes

produced in this way, have a molded inside diameter and outside diameter can be

machined if required, after curing.

1.4 Literature survey and previous work on composite tubes in bending

Considerable efforts have been conducted in studying properties of composite

laminated beams in bending. Beams with I-section and hat-section under bending been

studied. Increasing use of composites in civil structures and sporting goods, various

sizes of composite tubes have attracted many interests in understanding their structures

3
response. Fam and Rizkalla [2] studied large diameter glass-fiber reinforced polymer

(GFRP) under bending to see the effects of concrete filling, central hole, and laminate

structure for strength to weight ratio and ductility of the tubes. Reddy and Binienda [3]

explained the bending phenomena in composite beams by new theory which includes

anisotropic nature of composites. Taheri et al. [4] carried out comprehensive numerical

investigation to evaluate the response and energy absorption capacity of hybrid

composite tubes made of unidirectional pultruded tube over wrapped with ±45° fiber

reinforced plastic. Parametric studies conducted examined the effect of tube’s length,

thickness and type of braid and loading conditions as well on crushing behavior of the

tubes. Chan and Demirhan [5] evaluated bending stiffness of composite tubes using the

modified lamination theory of composites. In their approach, change of fiber orientation

along the circumference of the tube is included. Hu et al [6], using numerical analysis,

evaluated macroscopic properties under biaxial bending of thin walled composite tube.

Stockwell and Cooper [7] investigated collapsing behavior of moderate wall thickness

composite tubes by finite element analysis. Padmanabhan et al.[8] examined the

mechanical behavior of an AlSiC metal matrix composite tubular samples with 17.8%

of volume fraction of 3µm SiC particles in a 2124 aluminum matrix. The tubes were

under ratios of combined tension and torsional loads. Grediac et al. [9] presented a

method to determine the four principal material constants of the through-thicknesses of

thick laminated composite tube. Ellyin and Maser [10] investigated the effects of

moisture absorption and exposure to elevated temperature on the mechanical properties

of glass fiber reinforced epoxy filament wound composite tubes. They performed multi-

4
axial tests and observed that for all biaxial stress ratios, strength and stiffness decreased

to some extent with presence of moisture and elevating the temperature. Nixon [11]

determined twist deformations for the design of full-scale extension-twist coupled tilt-

rotor blades by conducting static torsion and axial tension tests on extension-twist-

coupled circular tubes. Chen et al. [12] experimentally investigated the impact damage

tolerance of thin walled composite struts made of both brittle and toughened epoxy.

They used two different impactor sizes and evaluated damage parameters like barely

visible surface damage, internal damage and residual strength against impact energy.

Krafchak et al. [13] presented experimental results to assess the effects of barely visible

internal damage on fatigue life of thin walled composite tubes. Undamaged composite

tubes and impact damaged composite tubes were tested in compression-compression

fatigue and they observed that fatigue life degraded because of impact damage. Barely

visible impact damage was predominantly due to matrix cracking and delamination in

their study. Jensen and Pickenheim [14] identified failure mechanisms and measured

compressive strength and stiffness of specimens that modeled micro modulations as

their focus was to determine compressive performance of a filament-wound cylinders.

1.5 Objective and outline of this study

After reviewing the past work in composite tubes, little work has been

conducted in evaluating the bending stiffness of composite tubes. Accurate evaluation

of bending stiffness is important for better prediction of deflection, buckling loads and

vibration response of structures. The purpose of this research was to experimentally

evaluate bending stiffness of uniform diameter composite tube. Four-point bending tests

5
are conducted using MTS machine. Chapter 2 mainly discusses the experimental test

plan and testing procedures in detail that were used in this study.

Chapter 3 reviews analytical model for evaluation of the bending stiffness of

composite tubes. A MATHEMATICA program is developed based on this model.

Chapter 4 presents and discusses the results obtained by tests. Chapter 5 investigates the

failure process of composite tube under bending by using x-ray radiograph and optical

fractography methods. The conclusions of this study are summarized in chapter 6.

6
 CHAPTER 2

EXPERIMENTAL PROGRAM

This chapter covers test plan devised to obtain principal material properties of

material and the bending stiffness of the tube. The experimental program also aims to

evaluate the effect of bending stiffness and bending strength due to stacking sequence,

fiber orientation and radius of the tube.

2.1 Description of Test Plan

Four point bending was primarily used as a test method to evaluate bending

stiffness of composite tubes. The whole test program was divided into two parts. Part

one was related to evaluate the material properties needed to incorporate in the

calculation of bending stiffness by analytical methods. Part two was mainly to obtain

data from strain gage and dial gage to assess the bending stiffness of the tubes.

2.2 Characterization of principal material properties (E1, E2, G12, υ12)

Determination of material properties is very important to evaluate bending

stiffness of composite tubes. For a thin layer of composite material, four basic material

constants are required to fully characterize the material structural response. They are

elastic modulus along the fiber, E1, along the transverse fiber direction, E2, shear

modulus in the plane, G12 and the Poisson’s ratio, υ12. These four constants were

evaluated using the specifications as per ASTM standards. Table 2.1 lists the typical

specimen sizes, specimen type and the ASTM standard specifications that were used for

7
each test. As indicated in Table 2.1, no ASTM standard specification was adopted for

determining the shear modulus in this study. In measuring shear modulus of fiber

reinforced composites, ASTM D4255 standard specification is often used. In this

specification, picture-frame or rail-shear fixture is needed. In this study, a 10˚ off-axis

tension test is used. This test is simple and requires no special fixture. This test has been

often used by industries.

Table 2.1 Test Plan for determination of material properties

Specimen Specimen No. of plies Material property ASTM standard

type dimension determined specification
0˚ coupons 0.5˝ X10.0˝ 5 E1 and υ12 D3039/D3039M-
00(2006)
90˚ 1.5˝ X 10.0˝ 10 E2 D3039/D3039M-
coupons 00(2006)
10˚ 1.0˝ X 10.0˝ 5 G12 NONE
coupons

2.2.1 Specimen preparation

Composite panels were made by T700S/G91 graphite/epoxy prepreg. The panels

were cured at 270 F and manufactured by Sawyer Composites in Fort Worth, Texas. As

per the ASTM standard-D3039/D3039M-00(2006) described in Table 2.1, required

coupons were made of unidirectional reinforced composite and these coupons were

tested in MTS machine to generate the data which was then assimilated to evaluate the

material properties. The following procedures were adopted and specimens were made

ready to be tested.

8
 1. Cut the cured panels into desired dimensions specified in Table 2.1.

2. Attach aluminum end tabs on each coupon using two-part epoxy

adhesive.

3. Measure the width and thickness of each coupon in the test zone at three

different locations.

4. Attach strain gage(s) on each coupon using epoxy solution giving

adequate curing time to ensure perfect bond between gage and specimen.

Figure 2.1 shows dimensions and gage locations of different coupons used to obtain all

of the principal material properties.

9
 1.5" 1.5"

10"

10"
0.5" 1.5"
(a) (b)

1.5"
1.5”

10”
10"

tlam
0.5”

(c) (d)
Figure 2.1 (a) 0˚ coupon (E1 and υ12), (b) 90˚ coupon (E2), (c) 10˚coupon (G12), (d) side
view of coupons.

10
2.2.2 Testing for basic material properties

The MTS machine was calibrated on its 10KN load cell first. The test specimen

was properly placed in the grip. The load is applied at a rate of 0.02 inches per minute.

During the loading, the loads and their corresponding displacements and strain data

were recorded through the data acquisition system. Strain gage data were only obtained

at a prescribed load interval. During loading specimens are observed visually with aid

of magnifying lens for any damage that may occur. The specimens were loaded until

total failure.

2.3 Four-point bending test of composite tubes

Flexural behavior of structural composites is characterized by bending test. Two

types of bending tests are usually used, namely three-point bending test and four-point

bending tests. In the three-point bending test, the moment along the specimen length is

linear and reaches the maximum at the loading point. In this case, the transverse shear

along the specimen length is constant between the loading and supported points. This

test is often used to study the transverse shear behavior of the composites. In the four-

point bending test, the bending moment between two loading points of the specimen is a

constant. This provides a convenient way to evaluate the bending stiffness of the test

sample. Bending stiffness of the uniform cross-section circular composite tubes is

evaluated by this test.

2.3.1 Geometry and lay-up of composite tubes

Table 2.2 lists the laminate lay-up of tubular wall, radius and number of the test

specimens. Four different radii of the composite tubes, R=0.25, 0.375, 0.5 and 0.75 inch

11
with the identical lay up of walled laminates were used to investigate the effect of

radius on the bending stiffness of the tube.

For R=0.375 inch, a set of [0/-45/+45]S with different stacking sequences were

used to study the bending stiffness and their failure process. Laminates with [0/-

15/+15/+15/-15/0]T, [0/-45/+45/+45/-45/0]T and [0/-75/+75/+75/-75/0]T were used to

investigate the effect of fiber orientation on bending stiffness and strength of composite

tube under bending.

Table 2.3 tabulates the lay-ups that are used for effects studied.

2P

P P
e=5.0” e=5.0”
A΄

ε top
A

L=14.0”
ε
bottom

Dial gage

Section A-A΄ [0/-45/+45/+45/-45/0]T

Figure 2.2 Geometry of the composite tube and section A-A΄.

12
 Table 2.2 Lay-up and radius of specimens
Specimen type Lay-up Inner radius (in)
2A1 [0/-45/+45/+45/-45/0]T 0.375
2A2 [-45/+45/0/-45/+45/0]T 0.375
2A3 [0/0/+45/-45/+45/-45] T 0.375
2B1 [0/-45/+45/+45/-45/0]T 0.25
2B2 [0/-45/+45/+45/-45/0]T 0.5
2B3 [0/-45/+45/+45/-45/0]T 0.95
2C1 [0/-15/+15/+15/-15/0]T 0.375
2C2 [0/-75/+75/+75/-75/0]T 0.375

Table 2.3 Effects studied by lay-ups

Effect studied Lay-up(s) used
Stacking sequence 2A1, 2A2, 2A3
Radius 2B1, 2A1, 2B2, 2B3
Fiber orientation 2A1, 2C1, 2C2

2.3.2 Fabrication of composite tube and inspection method

In general composite tubes can be fabricated by several different ways. Sheet-

wrapping, filament winding and pultrusion are among the popular methods to

manufacture the composite tubes. The composite tube specimens were supplied by

United Sports Technologies. The tubes were manufactured by sheet-wrapping process

in which the unidirectional fiber-reinforced prepregs were hand laid and then machine

rolled on steel mandrel. To apply pressure from outside poly-propylene tape was used

and steel mandrel gave it pressure from inside while curing. Curing temperature for the

material chosen is 270 F and curing cycle for the same is 70 minutes. The tubes were

cured in oven vertically.

13
 The main advantage of this process over filament winding is its ability to lay

prepregs at desired angle. Another most common procedure to fabricate uniform

circular tube is by pultrusion in which constant cross-section area is the main

requirement.

After the tubes were fabricated, they were inspected visually to see any damages

or initial curvature induced. As the tubes were hung vertically during curing, the

induced curvature of the tubes was minimal. Any surface cracks were inspected visually

using paper paint.

2.3.3 Test Fixture design and test set up

Figure 2.2 shows the tube geometries and a typical lay-up of composite tube. As

shown in the figure, the distance between the two fixed plates is 14 inches and

composite tube is 18 inches in length with 2 inches of hanging length on either side was

given. The composite tube is placed on fixture which was designed for this experiment

and loaded until failure to record for deflection and strain at specified locations. In order

to perform the four-point bending test fixture was designed so that the test set-up can

allow to measure data needed for calculations. A typical set up for four-point bending

test is shown in Figures 2.3.

14
 Figure 2.3 Typical set up for four-point bending test.

On the base of the machine were designed two aluminum plates screwed inside

firmly with roller supports on top. The point pin support allows the tube free rotated and

gives no additional moment to the tube. Figure 2.4 shows the detail of the base plate.

Upper plate shown in Figure 2.5 was designed to screw into 10 KN capacity load cell

with two rollers on either ends which are 4 inches apart. The test load was applied to the

composite tube through these rollers. To avoid sudden failure at upper loading pins a

firm polymeric cushion was attached on the rollers. Test runs were performed to select

right thickness of cushion.

A dial gage (see Figure 2.3) was used to obtain deflection right below loading

point. The gage was calibrated to obtain maximum deflection in the center of the tube to

obtain bending stiffness. The strain gage was also used to record strain in tube during

15
loading. Two strain gages were mounted along the longitudinal direction on the top and

bottom surfaces of the tube. During testing important observations were noted down

carefully to evaluate the behavior which is of certain importance during their failure

analysis.

Figure 2.4 Base plate for the fixture.

Figure 2.5 Top plate attached to load cell.

2.3.4 Specimen preparation

1. Measure the inner and outer diameter at 4 different locations to obtain wall

thickness of the tube.

2. Inspect the test specimen for any damage that may occur during handling.
16
 3. Mount strain gage along the longitudinal direction on the top and bottom

surfaces of the test tube.

4. Mount strain gage using two part adhesive solution mentioned for the strain

gage type and allow it to cure for required time span to get perfect bonding.

2.3.5 Test procedure

1. Place the test tube in the proper position on the fixture.

2. Both upper pins are in perfect contact with the tube.

3. Calibrate the 10KN load cell along with other data acquisition devices like dial

gage and strain gage indicator for accurate results.

4. Load the tube at loading rate of 0.02 inches/min.

5. Set the zero reading of dial gage indicator and strain gage indicator before the

load application.

6. Record strain gage and dial gage reading at the pre-set intervals.

7. Any peculiar behavior is recorded and cracking behavior on loading is

important to observe and document and specimen is preserved at its fractured

regions for its failure analysis under microscope.

8. Test is stopped when the tube fails or no longer carry the load.

17
 CHAPTER 3

REVIEW OF ANALYTICAL MODELS FOR BENDING STIFFNESS

EVALUATION OF COMPOSITE TUBES

In this chapter analytical model for calculating bending stiffness of uniform

circular composite tube is reviewed. In evaluating bending stiffness of laminated

composite tube, two approaches were often considered, one with using smear property

approach. Both approaches employed lamination theory to calculate the properties of

the laminate. A brief description of these approaches is included in this chapter.

3.1 Geometry and loading of composite tube

The composite tube considered here has a uniform circular cross-section with a

radius R. R is measured from the center of the tube to the mid-plane of the wall-

thickness. The length of the tube is much larger as compared to its radius. The tube is a

layer structured with plies oriented in preferred orientations. The wall laminate of the

tube can be in any general lay-up. The tube is subjected to pure bending moment, Mx. It

should be noted that unlike the conventional strength of material, Mx points to the

transverse to the longitudinal direction of the tube (x-direction). Mx is the resultant

moment due to σx. It is assumed that plane cross-section remains plane and the circular

cross-section remains unchanged after the deformation.

18
 3.2 Lamination theory

A laminate contains multiple layers in different fiber orientations perfectly

bonded together. To analyze this kind of laminate, an equivalent plate theory is

developed. It is so-called laminated plate theory or lamination theory. The lamination

theory is used to calculate the stresses and strains in different plies of laminated

composite under load. It consists of a collection of mechanics-of-materials type of stress

and deformation hypothesis which is described in this section. Using this theory, we

proceed from basic building blocks, the lamina, to structural laminate. It is used to

calculate the following terms:

• Stiffness matrices for the laminate

• Mid-plane strains and curvatures for each lamina

• In-plane stresses and strains for each laminas

Since the laminate is thin compared to other dimensions, the theory of plate is

adopted. Because of thin lamina, the plane stress condition is assumed. That is,

σ z = τ xz = τ yz = 0

3.2.1 Ply stress-strain relationships in material coordinates

The stress-strain relation for the orthotropic plies under plane stress can be

expressed in terms of four independent elastic parameters, Q11, Q12, Q22 and Q66 as:

 σ 1   Q11 Q12 0  ε 1 
    
 σ 2  =  Q 21 Q 22 0  ε 2  Or [σ 1 − 2] = [Q1 − 2][ε 1 − 2] (3.1)
τ 12   0  γ 12 
0 Q 66 
   

19
 The subscripts 1, 2 and 6 refer to the principal coordinate system that is along

the fiber, transverse to the fiber and shear direction, respectively. The components of

the [Q] matrix is given as

E1
Q11=
1 −ν 12ν 21

E2
Q22= (3.2)
1 −ν 12ν 21

E 1ν 21 E 2ν 12
Q21=Q12= =
1 −ν 12ν 21 1 −ν 12ν 21

Q66=G12

Where, E1 and E2 are the moduli of lamina along the fiber and transverse

direction respectively. G12 is the shear modulus of lamina in 1-2 planes and ν12 is the

Poisson’s ratio of lamina due to the loading along the fiber direction.

3.2.2 Ply stress-strain relationships in laminate coordinates

Each layer in the laminate has its own principal material coordinate system. A

coordinate system, x-y-z common to all of laminas is selected. These coordinates are

usually set at the mid-plane of the laminate. The strains in any given lamina (kth lamina)

can be expressed in terms of strain at the mid-plane and the curvature of the laminate as

follows:

ε x  ε x  κ x 

  
  
ε y  = ε y  + z κ y  (3.3)
γ s  γ
 κ 
   s  s
20
σ x  Qxx Qxy Qxs  ε x
 Qxx Qxy Qxs  κ x 
σ  = Q     
 y   yx Qyy Qys  ε y
 + z Qyx Qyy Qys  κ y  (3.4)
 τ s  k  Qsx Qsy Qss  γ s
 
 Qsx Qsy Qss  κ s 
k k

[Qx-y]k is the stiffness matrix of the kth lamina referred to the laminate coordinate

system, x-y-z.

3.2.3 Laminate constitutive equations

The resultant forces, [N] and moments, [M] of the laminate can be obtained by:

n zk
[ N ] = ∑ ∫ [σ ]k .dz
k =1
zk − 1

n zk
[ M ] = ∑ ∫ [σ ]k .z.dz (3.5)
k =1
zk − 1

After performing the integration, we have:

 N   A B  ε

M  = B D .   (3.6)
    κ 

Where,

n ^
[ A] = ∑ [Q ] ( zk − zk −1 )
k =1 k

1 n ^
[ B ] = ∑ [Q ] zk2 − zk2−1
2 k =1
( ) (3.7)
k

1 n ^
[ D ] = ∑ [Q ] zk3 − zk3−1
3 k =1
( )
k

21
 ^
zk refers to the z coordinate of the upper interface of the kth laminate. [Q ] k

matrix is so- called reduced stiffness matrix of the kth layer. The matrices, A, B and D

are 3x3 matrices. They refer as extensional, extensional-bending coupling and bending

stiffness matrices, respectively. Equation 3.6 is often called as “Laminate Constitutive

Equations”.

3.3 Laminated plate approach

The determination of bending stiffness for composite tubes by laminated plate

approach was developed by Chan and Demirhan [5]. In their approach, an infinitesimal

plate section of the tube laminate is considered as shown in Figure 3.1. This section has

its axis x-y-z and is inclined at an angle θ with respect to axis of the composite tube.

The plate section is rotated about x to position parallel to the y΄-axis. The stiffness of

the plate calculated by the lamination theory is translated to the axis y΄ according to

parallel axis theorem. The overall stiffness of the composite tube is then obtained by

integrating over the entire θ domain.

22
 z ''

y ''

x’, x”, x

z” z

θ α
R
R cosθ y

y”

Figure 3.1 Infinitesimal section of composite tube.

23
 The overall stiffness matrices,  A ,  B  ,  D  of the tube can be expressed as:

2π
 A =
  ∫ [ A '].R.dθ
0

2π
B =
  ∫ [ B '].R.dθ
0

2π
D =
  ∫ [ D '].R.dθ
0

Where,

 A'  = [ A]

 B '  = [ B ] + R.cos θ .[ A]

 D '  = [ D ] + 2 R.cos θ .[ B ] + ( R.cos θ ) [ A]

2
(3.8)

The [A΄], [B΄], [D΄] matrices are the universal stiffness matrices per unit section

of the composite plate as shown in figure with respect to x-y-z axis shown. Substituting

equations and rearranging equations gives the final equations for calculation of the total

extensional, coupling and bending stiffness matrices as follows:

2π n  2π ^ k
 n ^k  
Aij = ∫  ∑ Q ij ( zk − zk −1 ) .R.dθ = R.∑  ∫ Q ij .dθ  ( zk − zk −1 )
0  k =1  k =1  0 

2π
R n  ^k 
(
B ij = ∑  ∫ Q ij .dθ  zk2 − zk2−1
2 k =1  0
) (3.9)


24
 2π n  2π ^ k
R n  ^k  3 3 
D ij = ∑  ∫ ij
3 k =1  0
Q .dθ  k
. (
z − z k −1 )
+ R 3
∑  ∫ Q ij .cos θ .dθ  ( zk − zk −1 )
2

 k =1  0 

^ k
Q ij is a function of rotation angle about x-axis, θ, fiber orientation angle, β, and the

elastic material constants as shown in appendix.

The effective bending stiffness of the composite tube can be expressed as:-

1
Dx = (3.10)
d 11

where d 11 is the (4, 4) element of the inverse matrix of  A B D  as shown as

−1
a b A B
 =  (3.11)
bT d   B D 

3.4 MATHEMATICA Program

In order to have efficient calculations on a public available software program,

MATHEMATICA was used to resolve the equations used in laminated plate approach

to get the value of bending stiffness. Features of this program include solving

mathematical equations using constants and variables. Variables are assigned as

functions in this program and there are certain codes which are followed to achieve the

task to be done like integration, differentiation and solving complex equations as well in

matrix form for this typical example. In this program, equations are input in certain

format which is similar to programming language and program then calculates when it

is run. The program coded for approach used in this work is described in flow chart

shown in Figure 3.2 and detailed code is shown in appendix A.

25
 E1, E2, ν12
G12, tply

[Q ] 1-2

Rotated about z-axis

[ Q ]x-y

Rotated about x-axis

[Qˆ ] x-y

Summed up for all laminas

[A], [B], [D]

Integration along the circumference

[ A],[B],[ D ]

Dx = 1
d 11
a b
 T  Dx
b d

Figure 3.2 Flowchart for Bending Stiffness matrix computation of composite

tube

26
 CHAPTER 4

EXPERIMENTAL RESULTS

4.1 Test results and discussion of principal material properties

The principal material properties of lamina, E1, E2, υ12 and G12 are measured in

this section. The thickness of the cured lamina is often considered a principal material

property since it depends on cured process and relates to the fiber volume fraction. E1

and E2 are directly obtained from the slope of the stress-strain curves of 0˚ and 90˚

coupons, respectively (see equation 4.1). The stress-strain curve is constructed from the

load-strain data which was stored in data acquisition system during the test. Poisson

ratio was estimated from strain readings read by horizontal and vertical strain gages

attached on zero-degree coupons (see equation 4.2). A 10˚ off axis coupon under

tension test is used to measure the shear modulus of 0˚ lamina. Shear modulus (G12) was

calculated by using equation 4.3. In equation 4.3, Ex is obtained from the 10˚ coupon

test. E1, E2 and υ12 in the right hand side of equation 4.3 is obtained from 0˚ and 90˚

coupon test data.

∆σ
E1 = E2 = E x = (4.1)
∆ε

ε lateral
ν 12 = − (4.2)
ε longitudinal

1 m2 2 n2 2 m2 n2
= (m − n 2ν 12 ) + (n − m 2ν 21 ) + (4.3)
Ex E1 E2 G12

27
 E2
Here, m= cos10˚, n= sin10˚ and ν 21 = ν 12 .
E1

Table 4.1 lists the material properties of the lamina. The cured ply thickness is

obtained by average measuring thickness of each coupon at three locations divided by

number of plies in that coupon.

Table 4.1 Material property test data

Property Tested Test Data Average Numbers
E1 (msi) 20.0, 20.5, 19.8 20.1
E2 (msi) 1.77, 1.79, 1.76 1.77
G12 (msi) 1.1, 1.13, 1.1 1.11
υ12 0.313 0.313
tply (inch) 0.0058 0.0058

Figure 4.1 shows images of broken 0˚, 90˚ and 10˚ off axis coupon, respectively.

0˚ coupon

90˚ coupon

10˚ coupon

Figure 4.1 Failed 0˚, 90˚ and 10˚ off-axis coupons.

28
 4.2 Strength of composite tubes and their load-deflection curves

Table 4.2 shows the final failure load of various lay-ups tested and load at

which first damage initiated. The first damage load was recorded when sound of crack

was heard. The characteristics of the load-deflection curves of composite tubes and their

strength are described below.

Table 4.2 Failure load and first damage load of tube specimens

Failure Load Average

Specimen Lay-up Ri (lbs) Load(lbs)
Code (outer to inner layer) (inch) First Final 1st Final
Damage Failure
180 275
2A1 [0/-45/+45/+45/-45/0]T 0.375 200 295 197 290
210 300
200 330
2A2 [-45/+45/0/-45/+45/0]T 0.375 225 380 217 367
225 390
180 315 190 317
2A3 [0/0/-45/+45/-45/+45]T 0.375 200 320
140 200
2B1 [0/-45/+45/+45/-45/0]T 0.250 150 215 150 208
160 210
110 300 117 325
2B2 [0/-45/+45/+45/-45/0]T 0.500 125 350
125 350
2B3 [0/-45/+45/+45/-45/0]T 0.950 125 350 128 367
135 400
190 215
2C1 [0/-15/+15/+15/-15/0]T 0.375 190 220 193 220
200 225
150 280
2C2 [0/-75/+75/+75/-75/0]T 0.375 160 300 162 310
175 350

29
4.2.1 Sudden failure of composite tubes

Figure 4.2 shows a typical load-displacement curve. As indicated, the curve is

linear elastic to failure. No yielding is observed before total failure. This behavior was

observed on specimens 2A1, 2A2, 2A3, 2B1 and 2C2. For these specimens, visual

damage was not seen by naked eye until close to failure load. Fiber breaking and

internal damage was heard with minor sounds at loads marked in Table 4.2 above.

300 Sudden Failure Specimens

250
Load (lbs)

200

150

100

50

0
0.0 0.1 0.2 0.3 0.4

Displacement (inch)

Figure 4.2 Load vs. Displacement for sudden failure specimens.

30
4.2.2 Gradual failure of composite tubes

Figure 4.3 shows load vs. displacement curve for tube specimens 2C1 that

exhibited load drop. Specimens 2B2, 2B3 and 2C1 showed gradual load drop behavior

before final failure. For the case of 2C1 specimens, the cross-section of tube remains

almost circular after load drop. The significant load drop is due to damage occurring in

0˚ply and delamination at the interface of +15˚ and -15˚plies.

250
R=0.375" 2C1 [0/-15/+15/+15/-15/0]T

200
Load (lbs)

150

100

50

0.0 0.1 0.2 0.3 0.4 0.5

Deflection (inch)

Figure 4.3 Load vs. Displacement for 2C1 specimen.

For the case of 2B2 and 2B3, change in the shape of tubular cross-section was

prominent at much lower loads as compared to failure load. Due to large diameter, these

31
specimens were able to sustain high loads with less deflection. Figure 4.4 shows load-

displacement curve for specimen 2B2.

300
[0/-45/+45/+45/-45/0]T 2B2 (R=0.5")

250

200
Load (lbs)

150

100

50

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Deflection (inch)

Figure 4.4 Load vs. Displacement for 2B2 specimen.

4.2.3 Effect of radius

Figure 4.5 shows load-deflection curves of tubes with various radii under

bending. All of the tubes considered here have the identical lay-ups with radius ranging

from 0.25 inches to 1.0 inches. As indicated, the strength of composite tube increases as

its radius increases. Conversely, the deflection decreases as the tube radius increase.

However, this is not the case for composite tube. For isotropic material, the strength of

32
the tube is proportional to the ratio of bending stiffness which is related to moment of

inertia (I).

400
[0/-45/+45/+45/-45/0]T 2B1(R=0.25")
360 2A1(R=0.375")
2B2(R=0.5")
320
2B3(R=1.0")
280
Load (lbs)

240

200

160

120

80

40

0.0 0.1 0.2 0.3 0.4 0.5 0.6

Deflection (inch)

Figure 4.5 Deflection vs. Load curves for various inner radii.

4.2.4 Effect of 0˚ ply location

Figure 4.6 shows the comparison of the tube deflection for the position of 0˚ply

in the tubular wall laminate. The results indicate that there is an insignificant difference

in the slope of the load-deflection curve. This implies that the bending stiffness of the

tube does not appear significant influence due to the change of the position of 0˚-ply. It

is well known that placing 0˚ ply away from the mid-plane of the laminate gives higher

33
bending stiffness of the laminate. However, the bending stiffness of the laminate tube,

shown in equation 3.8 constitutes two parts. One is due to the bending stiffness with

respect to its mid-plane bending axis and the other part is from the axis shift. Examining

the equation, we found the latter part of the bending stiffness is dominant since it is

related to the extensional stiffness of the laminate, [A]. It is known that position of

0˚ply in laminate affects [D] matrix, not [A] matrix. Therefore, the bending stiffness of

the tube is insignificant difference.

450
2A1[0/-45/+45/+45/-45/0]T R=0.375"
400
2A2[+45/-45/0/+45/-45/0]T
350 2A3[0/0/-45/+45/-45/+45]T

300
Load (lbs)

250

200

150

100

50

0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

Deflection (inch)

Figure 4.6 Deflection vs. Load data on number of zero-degree plies

34
4.2.5 Effect of fiber orientation of angle plies

Coupons with ∓ 75˚, ∓ 45˚ and ∓ 15˚plies placed between two 0˚ plies in the

wall laminates were used to study the effect of bending strength and the deflection. The

load vs. deflection curves are shown in Figure 4.7. The results indicate that increasing

fiber orientation of angle plies results in decrease of the bending stiffness but increase

of the bending strength. The coupons with ∓ 75˚ plies placed have the highest bending

strength and the largest deflection of the tube among those coupons studied. Intuitively,

the coupons with ∓ 15˚ should carry the higher load to failure compared to the other

two sets of coupons. Carefully examining the failure of the coupons, it is found that

significant shear failure and delamination occurs at the interface of -15˚/+15˚ layers.

350 2C1[0/-15/+15/+15/-15/0]T R=0.375"

2A1[0/-45/+45/+45/-45/0]T
300 2C2[0/-75/+75/+75/-75/0]T

250
Load (lbs)

200

150

100

50

0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

Deflection (inch)

Figure 4.7 Effect of fiber orientation in symmetric lay-ups in Deflection vs. Load
35
 4.3 Curvature effect on bending stiffness of composite tube

Bending stiffness of uniform circular tube with isotropic properties can be

evaluated by 2D-model using solid mechanics equations. Laminated composites being

transversely isotropic materials as discussed in chapter 3 need to be correctly evaluated

using lamination theory. Table 4.3 shows the calculated bending stiffness based on both

experimental methods and analytical methods described in chapter 3.

Table 4.3 Bending stiffness of various lay-ups of composite tubes

Lay-up Dial-Gage Plate Approach Smear Strain Gage

Results Approach Results
2A1 4.553 x104 5.611 x104 5.98 x104 5.779 x104
2A2 4.645x104 5.604 x104 5.975 x104 5.715 x104
2A3 4.709x104 5.602 x104 5.973 x104 6.22 x104
2B1 1.721x104 1.866 x104 1.99 x104 1.86 x104
2B2 8.702 x104 1.488x105 1.586x105 none
2B3 1.206x105 7.84x105 8.358x105 none
2C1 6.634 x104 1.142x105 1.16x105 1.243x105
2C2 4.092 x104 4.877 x104 5.154 x104 5.032 x104

4.3.1 Discussion for possible error in approaches

As seen in Table 4.3 above, bending stiffness for specimen 2B1 is closest match

between experimental approach and plate approach. 2B1 has smallest outer diameter

and hence change of cross-section from uniform circular to elliptical is minimal. In

specimens 2A1, 2A2 and 2A3 outer radius is same with different stacking sequence.

Bending stiffness calculated as per plate approach differs from experimentally

calculated due to change of cross-section observed. Below is table 6 showing horizontal

and vertical diameter data for some cases that confirm the change in shape from circular

as assumed in plate approach model in chapter 3. 2B2 and 2B3 are the largest diameter

36
tubes and shape not just changed from uniform circular to elliptical but deformation

mechanism was much more complicated as seen in Figure 4.8. In these specimens

change of shape was localized as well and plate approach was not capable to determine

it’s bending stiffness.

Figure 4.8 2B3 specimen under four-point bending.

37
 CHAPTER 5

FAILURE INVESTIGATION

Failure analysis is an important tool used in to study their fracture behavior

under loading and their physics behind the failure. Cause of failure is determined using

fractography and an insight of failure mechanism can be obtained by using this tool. In

study of any failure, broad spectrum of possibilities for the failure to occur must be

carefully analyzed. Two failure inspection methods were often adopted. They are:

1. X-ray radiography

2. Optical Microscopy

In this chapter, we will discuss the failure of composite tube by using these two

inspection methods.

5.1 Failure Process and X-ray Radiography

X-ray radiography is a non-destructive method for damage inspection. It was

used to get basic information of the failure such as the size of the failure zone. Damage

size was important information which needed to be addressed before the destructive

method is conducted. In this technique, specimen preparation was important procedure

to ensure successful results. Since the graphite/epoxy composite is transparent to x-ray,

a solution opaque to x-ray embedded in the composite is needed to enhance the damage

of the specimen. Zinc iodide was chosen for this matter to enhance damage size. The

solution was diluted with distilled water.

38
5.1.1 X-ray radiographic procedure

1. The damage area of the composite tube specimens were soaked in the

solution for approximately 20-30 minutes to ensure complete seepage of zinc

iodide into the specimen.

2. Excess solution of zinc iodide on the specimen was removed by using clean

lint free cloth or some other media.

3. Polaroid film (52-type) was placed below specimen on specified markings to

get proper imaging of fracture region. This was critical step in getting

relevant image.

4. Cabinet x-ray machine was used at a voltage of 12kV and specimens were

exposed to x-rays for 10 minutes to generate results.

5. After the exposure voltage was turned down at slow rate to avoid element

malfunctioning.

6. Film taken out is thus developed in the Polaroid film developer.

5.1.2 Analysis of x-ray images

Figure 5.1 through 5.3 shows the radiographs of 2A1, 2A2 and 2A3 specimens,

respectively. These specimens have the same diameter but different stacking sequence

(see Table 4.2). Specimens 2A1, 2A2 and 2A3 have one 0˚, none and two 0˚ plies on

the outer surface of the wall laminate, respectively. The dark region shown in Figure 5.1

is where zinc iodide solution penetrated in cracks and illuminated it on the film. It is

clearly seen that crack on the circumference is running around in slant fashion and is

localized clearly around the loading point. This justifies presence of hidden possible

39
delamination as well. The cracks and delamination are in the inner layers and can’t be

viewed just using x-ray radiography and need to be opened up to view them at higher

magnifications. This specimen failed at load around 300 lbs and no visual damage was

seen until the end. Cracking sound was prominent after 220 lbs and dark black spot is

the actual loading point in this specimen.

Damage Zone

Shadow

Figure 5.1 X-ray image of specimen 2A1.

Figure 5.2 is the image of specimen 2A2 and dark region shows the presence of

visual crack after tube failed. This specimen sustained load until about 350 lbs and

damage was again sudden as in previous case and we see in this one as well that crack

started at the loading point and continued from layer to layer around the circumference

of tube in slant fashion.

40
 Damage
Zone

Shadow

Figure 5.2 X-ray image of specimen 2A2.

In Figure 5.3, damage of specimen 2A3 was predominantly the same as the

previous specimens shown. The crack is also observed to be from layer-to-layer in

oblique manner indicating possibility of shear failure.

In viewing of the above three specimens, we observed that shear cracks are more

pronounced in specimen 2A2. this is because of more ±45˚ plies and none 0˚ ply near

the outer surface of the wall laminate.

Damage Zone Shadow

Figure 5.3 X-ray image of specimen 2A3.

41
 Figure 5.4 is the same lay-up as 2A1 but with smaller diameter. Sudden failure

and fiber bundles breakage was finally observed. Cracking sound was heard but not

significantly loud. This suggests there are some matrix cracks and delamination

occurring.

Shadow

Damage Zone

Figure 5.4 X-ray image of specimen 2B1.

In Figures 5.5 and 5.6, specimens 2B2 and 2B3, respectively, are the tubes with

same lay-up as 2B1 but larger diameters. Large deformation was observed while the

tube was loaded over 225 lbs and then the cross-section was no longer circular. The

peak loads in both the cases was close to 350 lbs owing to change of shape involved and

failure was not sudden and load drop was consistent from peak load to the point test

stopped.

42
 Shadow
Damage Zone

Figure 5.5 X-ray image of specimen 2B2.

Damage
Zone

Figure 5.6 X-ray image of specimen 2B3.

Below is Figure 5.7 showing image of damage region in specimen 2C1. Crack in

this case also initiated near the loading point and continued in shear fashion across the

circumference. Failure in this case happened at load of 225 lbs and cracking sound

started at approximately 200 lbs and sample failed with sudden quick tearing sound at

225 lbs.

43
 Through thickness crack

Damage zone

Figure 5.7 X-ray image of specimen 2C1.

In Figure 5.8, specimen 2C2 failed at load of 300 lbs with sudden failure and no

significant cracking or tearing sound was heard in this case.

Damage Zone

Figure 5.8 X-ray image of specimen 2C2.

5.2 Optical Microscopic Analysis

Analysis of x-ray radiography can not provide the detailed information of failure

process, layer location of delamination. To meet this need, optical microscopy is used to

obtain this information. Optical microscope with magnification between 50X-400X was

used to investigate the fracture surface. For that purpose the failed tubes have to be cut

at the specified location. The selection of location and cutting process are critical to

44
preserve fracture surface as well. Figure 5.9 shows the specimen taken from the tube for

optical investigation. A procedure is discussed in the next section.

5.2.1 Experimental Procedure

1. Failed tube was marked for cutting locations and that was decided upon

analysis from x-ray images.

2. The full-length tube was cut into small round piece containing fracture area

to eliminate the size being barriers to further cut open the tubes.

3. Cut was made by using diamond-edged micro cutter to open the failed sample

into two halves showing fracture on both halves and direction of cut was

such chosen to have minimal or no effect on fracture area.

4. The opened specimen was then mounted on sample holder to firmly place it

just below microscope to view it under low and then higher magnifications.

Loading Point Loading Point

B A

B’ A’
Cutting line

Figure 5.9 Specimen for Optical Microscopy investigation.

45
5.2.2 Delamination failure in tubes

Reinforcement in laminated composites is in the plane and none in the thickness

direction. Hence, failure often occurs in between layers, namely delamination. Figure

5.10 shows the fractured specimen 2A1 which has 0˚ fibers in its innermost and

outermost layers followed by symmetrically placing -45˚ and +45˚ fibers.

Delaminations are observed at the interfaces between 0 and -45˚ layers and +45 and -

45˚ layers of the innermost layers.

LOADING
POINT

CRACK
PROPAGATION

FIBER
BREAKAGE

Figure 5.10 Fractured tube specimen 2A1 at 150X magnification

In the Figure 5.11, the fiber failures were occurred at the outermost layers of

+45 and -45 plies and the innermost layers of 0˚ and 45˚ plies. The delamination is also

observed at the interface of 0 and -45 plies on the mid-plane of the laminate.

46
 DE-LAMINATION

LOADING
POINT

FIBER
BREAKAGE

Figure 5.11 Fractured sample 2A2 at 70X magnification

DE-LAMINATION

Figure 5.12 Fractured sample 2A3 at 350X magnification

Figure 5.12 shows the fracture image of specimen 2A3 which has all of 0˚ plies

in the outermost layers. Delamination was observed in the mid-plane of the laminate.

47
 In summary, for all of 2A1, 2A2 and 2A3 specimens, the delamination occurred

mostly at the inner surface in the localized area near the loading point. Figure 5.13 and

5.14 shows extensive damage of fiber breakage and delamination for specimen 2A2 and

2A3, respectively. It should be noted that fracture surface shown is located at the mid

plane of the entire tube where the highest shear occurs. Hence, the failure can be

attributed to the shear failure.

LOADING
POINT

CATASTROPHIC
FAILURE

Figure 5.13 Specimen 2A2 at 100X magnification showing shear effect and failure
between +45 and -45 plies

48
 FIBER BREAKAGE LOADING
POINT

Figure 5.14 Specimen 2A3 at 300X magnification showing disastrous shear failure in
the middle

Figure 5.15 and 5.16 shows that the crack initiates between the layers where the

maximum shear mismatch (positive and negative shear), interface of +45 and -45 plies,

but propagates in an oblique manner from layer to layer as well.

CRACK
PROPAGATION
LOADING
POINT

DE-LAMINATION

Figure 5.15 Specimen 2A1 at 275X showing delamination and crack propagation in
oblique manner from layer to layer as well.
49
 CRACK
PROPAGATION

FIBER
BREAKAGE

RESIN
FAILURE

Figure 5.16 Crack propagation in 2A3 at 350X magnification.

Figure 5.17 shows specimen 2B1 symmetric laminate with smaller radius has

similar failure process as in specimen 2A1.

DE-LAMINATION
LOADING
POINT

CRACK PROPAGATION

Figure 5.17 Failed sample 2B1 at 200X magnification showing crack propagation.

50
 In the Figure 5.18 below is symmetric laminate with angle plies at +15˚ and -15˚

with 0˚ plies on outer most and innermost surface and delamination is observed in the -

15 and +15 plies as that is the location where the highest shear mismatch occurs and

crack tends to propagate from layer to layer as well in an oblique manner.

DELAMINATION

Figure 5.18 Failed sample 2C1 at 100X showing delamination and crack propagation
from layer to layer as well.

51
 CHAPTER 6

CONCLUSIONS

Four-point bending tests were conducted to evaluate the bending stiffness and

strength on various sizes of the composite tubes. The walled laminates of the tubes with

various stacking sequence of [0/-45/+45/+45/-45/0]T graphite/epoxy. Coupons of the

material used for the tube were also tested to obtain the principal material properties.

Bending stiffness of the composite tube was calculated by analytical methods.

Among analytical methods, both laminated plate method and smear property method

were used. The bending stiffness of the tube was also calculated based upon the

measurement data. Experimental bending stiffness is calculated from the longitudinal

strains at both upper and bottom surface. It is also calculated from the dial gage

readings.

From this study we found that:

• For all of the tube studied, bending stiffness obtained from laminated

plate method is closer to the one obtained by strain gage data than from

smear property method.

• The composite tube with smallest radius has least difference in bending

stiffness obtained by laminated plate approach and experimental method

for the tube with [0/-45/+45/+45/-45/0]T laminate.

52
 • For large radius of the tubes, the experimental values of the bending

stiffness deviate from the analytical prediction due to deformation in

shape from circular cross-section to elliptical shape under load.

• Stacking sequence of tube walled laminate does not appear to affect the

bending stiffness of composite tube, in particular for larger radius of the

tubes.

• Bending strength of tubes with identical radius increases as the 0 layers

in laminates are placed close to the mid-plane of the wall laminate.

• Unlike isotropic material, the bending stiffness of the tube is not

proportional to R4. It is a function of R3 and R as well as the material

property of the tube.

• Among the tubular walled laminates such as [0/-15/+15/+15/-15/0]T, [0/-

45/+45/+45/-45/0]T, [0/-75/+75/+75/-75/0]T , the tubes have the highest

bending strength with [0/-75/+75/+75/-75/0]T laminate and have the

lowest strength with [0/-15/+15/+15/-15/0]T laminate.

Fracture analysis was also conducted by using both x-ray radiography and

optical microscopy to investigate the failure process of the tube under bending. No

attempt was focus on identification of fracture modes because of catastrophic and

multiple failures observed. The following observations were obtained during and after

test:

53
 • Damage is initiated at the loading point then propagates in the shear

direction along the circumference.

• The final failure is in compression. No failure in the tension side of the

tube is observed.

• The failure process depends on the fiber orientation and ply stacking

sequence of the walled laminate of the tubes.

• Fiber breakage and delamination were observed as the prominent

damage mode.

For a better estimate on bending stiffness of the composite tubes undergoing

deformation in shape, closed form solution incorporating deformation of cross-section

from circular to elliptical is needed. Future scope of this study can involve estimate of

bending stiffness of tube specimens involving shape change by using modified model of

plate approach.

54
 APPENDIX A

MATHEMATICA CODE FOR STIFFNESS MATRICES

55
 (*Closed form solution for stiffness matrices of composite tubes*)

E1=20.0x106; E2=1.77x106; G12=1.1x106; υ12=0.313; tply=0.0058;

E2
ν 21 = ν 12 ;
E1

E1 E2 ν 12 E2
Q11 = ; Q22 = ; Q22 = Q21 = ; Q66 = G12 ;
1 −ν 12 .ν 21 1 −ν 12 .ν 21 1 −ν 12 .ν 21

h= Table [i, {i,-3, 3}] tply;

π
θ= {0, −1, +1, +1, −1, 0} ;
4

R=0.39;

 6  3  
A11 = 2π R  ∑   ( Cos[θ [[i]]]) Q11 + ( Sin[θ [[i]]]) ( Cos[θ [[i ]]]) ( Q12 + 2Q66 ) + ( Sin[θ [[i ]]]) Q22  ( h[[i + 1]] − h[[i]])  
4 2 2 4

 i =1   8  

 6  2 3  1  
A12 = 2π R  ∑   ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q11 + Q22 − 2Q66  + ( Cos[θ [[i ]]]) + ( Sin[θ [[i ]]]) Q12  ( h[[i + 1]] − h[[i ]])  
2 4 4
 i =1   8  2  
  

 6  3  
A22 = 2π R  ∑   ( Sin[θ [[i]]]) Q11 + ( Sin[θ [[i ]]]) ( Cos[θ [[i]]]) ( Q12 + 2Q66 ) + ( Cos[θ [[i]]]) Q22  ( h[[i + 1]] − h[[i]])  
4 2 2 4

 i =1   8  

 6  3 1  1 3  
A16 = 2π R  ∑   ( Sin[θ [[i ]]])( Cos[θ [[i ]]])  Q11 − Q12 − Q66  + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q12 − Q22 + Q66   ( h[[i + 1]] − h[[i ]])  
3
 i =1   2   2 8   
  

 6   1  31 3  
A26 = 2π R  ∑   ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q11 − Q12 − Q66  + ( Sin[θ [[i ]]])( Cos[θ [[i ]]])  Q12 − Q22 + Q66   ( h[[i + 1]] − h[[i ]])  
3
 i =1   2  2 8  
  

 6  
 i =1 
 
2


3
8
 1
 2
( 4 

)
A66 = 2π R  ∑   ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q11 + Q22 − Q12 − Q66  + ( Sin[θ [[i]]]) + ( Cos[θ [[i ]]]) Q66  ( h[[i + 1]] − h[[i ]])  
2 4




56
  6 
 i =1 
2 2 3
8
4 

2
( 2 
B11 = π R  ∑  ( Cos[θ [[i ]]]) Q11 + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]]) ( Q12 + 2Q66 ) + ( Sin[θ [[i ]]]) Q22  ( h[[i + 1]]) − ( h[[i ]]) 
4


)

 6  2 
 i =1 
 
2 2


3
8
 1
 2


2
(
B12 = π R  ∑   ( Sin[θ [[i]]]) ( Cos[θ [[i ]]])  Q11 + Q22 − 2Q66  + ( Sin[θ [[i ]]]) + ( Cos[θ [[i ]]]) Q12  ( h[[i + 1]]) − ( h[[i ]])  
4 4


)

 6  2 

 i =1  
4 2 2 3
8
4 

(
B 22 = π R  ∑   ( Sin[θ [[i]]]) Q11 + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]]) ( Q12 + 2Q66 ) + ( Cos[θ [[i ]]]) Q22  ( h[[i + 1]]) − ( h[[i ]])  
2


)

 6  2 
 i =1 
 
3


1
2


1
2
3
8


( 2
)
B16 = π R  ∑   ( Sin[θ [[i]]])( Cos[θ [[i ]]])  Q11 − Q12 − Q66  + ( Sin[θ [[i]]]) ( Cos[θ [[i]]])  Q12 − Q22 + Q66   ( h[[i + 1]]) − ( h[[i]])  
3



 6  2 
 i =1 
 


1
2


31

 2
3
8

 
( )
B 26 = π R  ∑   ( Sin[θ [[i ]]]) ( Cos[θ [[i]]])  Q11 − Q12 − Q66  + ( Sin[θ [[i ]]])( Cos[θ [[i]]])  Q12 − Q22 + Q66   ( h[[i + 1]]) − ( h[[i]])  
3 2



 6  2 
 i =1 
 
2


3
8
 1
 2
( 4 4
) 

(
B 66 = π R  ∑   ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q11 + Q22 − Q12 − Q66  + ( Sin[θ [[i ]]]) + ( Cos[θ [[i ]]]) Q66  ( h[[i + 1]]) − ( h[[i ]])  
2 2


)

2π R  6   3 
D11 =
3  i =1  
4 2 2 3
8
4 

(
 ∑   ( Cos[θ [[i]]]) Q11 + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]]) ( Q12 + 2Q66 ) + ( Sin[θ [[i ]]]) Q22  ( h[[i + 1]]) − ( h[[i ]])  
3


)
 6  1 3 5  
+2π R 3  ∑   ( Cos[θ [[i ]]]) Q11 + ( Sin[θ [[i]]]) ( Cos[θ [[i]]]) ( Q12 + 2Q66 ) + ( Sin[θ [[i ]]]) Q22  ( h[[i + 1]] − h[[i ]])  
4 2 2 4

 i =1   2 4 16  

2π R  6  3 
D12 =
3  i =1 
2


3
8
 1
 2
( 4 4 

) 3
(
 ∑  ( Sin[θ [[i ]]]) ( Cos[θ [[i ]])  Q11 + Q22 − 2Q66  + ( Sin[θ [[i ]]]) + ( Cos[θ [[i ]]]) Q12  ( h[[i + 1]]) − ( h[i ]]) 
2


)
 6  
 
i =1
2 21

 2
5
16
3
2
 3
 8
( 

)
+2π R3  ∑  ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q11 + Q22 − Q66  + ( Sin[θ [[i ]]]) + ( Cos[θ [[i ]]]) Q12  ( ( h[[i + 1]]) − ( h[[i ]]) ) 
4 4

2π R  6   3 
D 22 =
3  i =1  
4 2 2 3
8
4 

(
 ∑   ( Sin[θ [[i]]]) Q11 + ( Sin[θ [[i]]]) ( Cos[θ [[i]]]) ( Q12 + 2Q66 ) + ( Cos[θ [[i]]]) Q22  ( h[[i + 1]]) − ( h[[i]])  
3


)
 6
 1 3 5  
+2π R 3  ∑   ( Sin[θ [[i ]]]) Q11 + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]]) ( Q12 + 2Q66 ) + ( Cos[θ [[i]]]) Q22  ( ( h[[i + 1]]) − ( h[[i]]) )  
4 2 2 4

 i =1   2 4 16  

57
 2π R  6   3 
D16 =
3  i =1  
3


1
2


1
2
3
8

 ∑   Sin[θ [[i ]]] ( Cos[θ [[i ]]])  Q11 − Q12 − Q66  + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q12 − Q22 − Q66   ( h[[i + 1]]) − ( h[[i ]])  
3


3


( )
 6
   1 3 3   3 5 3   
+2π R 3  ∑   Sin[θ [[i]]] ( Cos[θ [[i]]])  Q11 − Q12 − Q66  + ( Sin[θ [[i ]]]) ( Cos[θ [[i ]]])  Q12 − Q22 + Q66   ( ( h[[i + 1]]) − ( h[[i ]]) )  
3 3
 i =1   2 8 4   8 16 4   
  

2π R  6   3 
D 26 =
3  i =1  
3


1
2


1
2
3
8

 ∑   Cos[θ [[i ]]] ( Sin[θ [[i ]]])  Q11 − Q12 − Q66  + ( Cos[θ [[i ]]]) ( Sin[θ [[i ]]])  Q12 − Q22 − Q66   ( h[[i + 1]]) − ( h[[i ]])  
3


3


( )
 6  31 3 3  3 5 3  
+2π R  ∑   Cos[θ [[i ]]] ( Sin[θ [[i ]]])  Q11 − Q12 − Q66  + ( Cos[θ [[i ]]]) ( Sin[θ [[i ]]])  Q12 − Q22 + Q66   ( ( h[[i + 1]]) − ( h[[i ]]) )  
3 3
 i =1  2 8 4  8 16 4  
  

2π R  6   3 
D 66 =
3  i =1  
2


3
8
 1
 2
4
( 4 
 ∑   ( Sine[θ [[i ]]]) ( Cos[θ [[i ]]])  Q11 + Q22 − Q12 − Q66  + ( Sin[θ [[i]]]) + ( Cos[θ [[i]]]) Q66  ( h[[i + 1]]) − ( h[[i ]])  
2


3


) ( )
 6  
 i =1 
 
 21

2
5
16
3
4
3
4
 3
 8
4
(4 
+2π R 3  ∑   ( Sin[θ [[i]]]) ( Cos[θ [[i]]])  Q11 + Q22 − Q12 − Q66  + ( Sin[θ [[i]]]) + ( Cos[θ [[i]]]) Q66  ( ( h[[i + 1]]) − ( h[[i ]]) )  
2

 

)

ABD = {{ A
11, }{ }{ }{ }{ }{
A12, A16, B11, B12, B16 , A12, A22, A26, B12, B 22, B 26 , A16, A26, A66, B16, B 26, B 66 , B11, B12, B16, D11, D12, D16 , B12, B 22, B 26, D12, D 22, D 26 , B16, B 26, B 66, D16, D 26, D 66 }}

Comp = Inverse [ABD];

Print [“ Resultant ABBD matrix”]
Comp // MatrixForm

“Bending Stiffness”
BS = 1/comp[[4]][[4]]

58
 REFERENCES

1. Peters, S.T (1998). Chapter 49 of Handbook of composites (2nd edition). p.1044

Chapman and Hall

2. Fam, A.Z. and Rizkalla, S.H, “Flexural behavior of concrete-filled fiber-

reinforced polymer circular tubes,” Journal of composites for construction, Vol.
6, No.2, May 2002, pp.123-132.

3. Reddy, E.S and Binienda, W.K., “Prediction of crack initiation in unidirectional

composite beams subject to four-point bending,” Composites Engineering, Vol.
4, No. 7, 1994, pp.703-714.

4. Han, H., Taheri, F., Pegg, N. and Lu, Y, “A numerical study on the axial
crushing of hybrid pultruded and ±45˚braided tubes,” Composite structures,
Volume 80, Issue 2, September 2007, pp.253-264.

5. Chan, W.S. and Demirhan, K.C, “A simple closed-form solution of bending

stiffness for laminated composite tubes,” Journal of reinforced plastics and
composites, Vol. 19, No. 4, 2000, pp.278-291.

6. Hu, G., Bai, J., Demianouchko, E. and Bompard P, “Mechanical behavior of

±55˚Filament-Wound Glass-Fibre/Epoxy-Resin Tubes-III. Micromechanical
model of the macroscopic behavior of tubular structures with damage and failure
envelope prediction,” Composites Science and Technology, 58 (1998), pp.19-29.

7. Stockwell, A. E. and Cooper, P. A, “Collapse of composite tubes under end

moments,” AIAA-92-2389-CP.

8. Padmanabhan, R, Macdonald, B.J and Hashmi, M.S.J, “Mechanical behavior of

SiC reinforced aluminum thin walled tube under combined axial and torsional
loading,” Journal of Materials Processing Technology, Volumes 155-156,
November 2004, pp.1760-1763.

9. Grediac, M., Pierron, F. and Zhavoronok, S, “Identification of the through-

thickness properties of thick laminated tubes using virtual fields method,”
International Journal of Solids and Structures, Volume 37, Issue 32, 8 August
2000, pp.4437-4453.

59
10. Ellyin, F. and Maser, R, “Environmental effects on the mechanical properties of
glass-fiber epoxy composite tubular specimens,” Composite Science and
Technology, Volume 64, Issue 12, September 2004,pp. 1863-1874.

11. Nixon, Mark W, “Extension-Twist Coupling of Composite Circular Tubes with

Application to Tilt Rotor Blade Design,” AIAA/ASME/ASCE/AHS 28th
Structures, Structural Dynamics and Materials Conference. Pp 295-303.

12. Chen, G-S, Bidinger, G.M, Lou, M.C, “Impact Damage Tolerance of Thin Wall
Composite Struts,” AIAA-93-1398-CP.

13. Krafchak, T.M, Petra, J-M, Davidson, B.D, Chen, G-S, “Effect of Impact
Damage on the Compression Fatigue Behavior of Composite Tubes,” AIAA-93-
1399-CP.

14. Jensen, David W, Pickenheim, Timothy R, “Compressive Behavior of

Undulations in Filament-Wound Composites,” AIAA-9-1516-CP.

60
 BIOGRAPHICAL INFORMATION

Puneet Saggar was born on June 02, 1979 in India. He came to The University

of Texas at Arlington in fall of 2004 after he took transfer from San Jose State

University in California. He first took his composite courses with Dr. Wen S. Chan in

spring of 2005 and upon showing his interest to work in the field of structural

composites; Dr. Wen S. Chan took him as his student. He showed dedication and hard

work towards completion of this study and endeavors to be a Materials Engineer

working in the field of composite materials used in structures.

61