Experimental and theoretical investigations of stiffened and unstiffened composite panels under uniform transversal loading

Table of Content

List of Figures

List of Tables

1. Introduction

2. Basics of Composite Structures
2.1 Common Fibres
2.2 Glass Fibre
2.3 Common Matrices
2.4 Epoxy Matrix
2.5 Fibre – Matrix Interaction
2.6 Failure Types of Composite Materials
2.6.1 Micro Failure Mechanism
2.6.2 Macro Failure Mechanism

3. Experimental Investigation of a Cantilever Beam
3.1 Setup and Procedure
3.2 Results and Discussion
3.3 E-Modulus
3.4 Conclusion

4. Experimental Investigation of Composite Panels
4.1 Problem Statement
4.1.1 Commonly Used Methods and Former Experiments
4.1.2 Method Adopted
4.2 Specimen Properties
4.2.1 Sample Preparation
4.2.2 Stiffeners
4.2.3 Sensor Setup
4.3 Experimental Program and Instrumentation
4.4 Test Results and Analysis
4.4.1 Panel-1, no Stiffener
4.4.2 Panel-2, one Stiffener
4.4.3 Panel-3, two Stiffeners
4.5 Discussion
4.5.1 Panel-1, no Stiffener
4.5.2 Panel-2, one Stiffener
4.5.3 Panel-3, two Stiffeners
4.6 Discussion and Errors

5. Laminate Theory for the Unstiffened Panel
5.1 Elastic properties
5.2 Analytic Approach with known formulas
5.3 Lamina Strength and Failure Theories
5.3.1 Lamina Strength and Failure Mechanism
5.3.2 Failure Theories
5.4 Laminate Strength and Failure
5.4.1 First Ply Failure
5.4.2 Ultimate Laminate Failure

6. Finite Element Analysis
6.1 Setup
6.2 Results
6.2.1 Panel-1, no Stiffener
6.2.2 Panel-2, one Stiffener
6.2.3 Panel-3, two Stiffeners
6.3 Discussion and Errors

7. Comparison
7.1 Deflections
7.2 Strains
7.3 Failures and Stresses

8. Conclusion

References

Appendix A: Nomenclature

Appendix B: Calculations

List of Figures

Figure 2-1: Chemical Structure of an Epoxy Matrix [14]

Figure 2-2: Failure Appearance in a Bended Cross Ply Composite [15]

Figure 3-1: Cross Section of Laminate Including Strain Gauges

Figure 3-2: Top and Side View of Cantilever Beam with Strain Gauges

Figure 3-3: Specimen before Cutting

Figure 3-4: Structure Failure and Delamination of Layer 1, 2 and 3

Figure 3-5: Load – Strain Plot for all six Gauges

Figure 3-6: Stress – Strain Plot of all six Gauges

Figure 3-7: Plots of all E-moduli

Figure 3-8: Compression Failure at the Bottom of the Specimen

Figure 4-1: Specimen Layout and Strain Gauge Positions

Figure 4-2: Hand Lay-Up Fabrication Process [4]

Figure 4-3: Hand Lay-Up Production of Panel

Figure 4-4: Nominal Stiffener Cross-Sectional Shapes [22]

Figure 4-5: Stiffener Fabrication, (a) Mould, (b) Half, (c) not glued Stiffeners

Figure 4-6: Glued Stiffeners with Clamps

Figure 4-7: Stiffener Geometry according to C. S. Smith [19]

Figure 4-8: Stiffener Position on the Panels

Figure 4-9: Three Wire Cabling of a Quarter Bridge [23]

Figure 4-10: General Stiffener Design and Strain Gauge Positions

Figure 4-11: Test Rig with Clamping, Cabling and Dial Gauges on Panel-1

Figure 4-12: Spots and Numbers of single and double stiffened Panels

Figure 4-13: Top (left) and Bottom (right) of Panel with Hole Enlargement

Figure 4-14: Load - Deflection Plot, Panel 1

Figure 4-15: Demounted unstiffened Panel with Cracked Edges

Figure 4-16: Test Rig with Clamping, Cabling and Gauges on stiffened Panel

Figure 4-17: Stages of Stiffener Failure

Figure 4-18: Stages of Panel Failure

Figure 4-19: Load – Deflection Plot, Panel 2

Figure 4-20: Demounted single stiffened Panel with Cracked Edges

Figure 4-21: Stiffener Failure, Load: 140 KPa

Figure 4-22: Different Stiffener Failure Types

Figure 4-23: Load – Deflection Plot, Panel 3

Figure 4-24: Failure of double stiffened Panel, Load: 220 KPa

Figure 4-25: Demounted double stiffened Panel with Cracked Edges

Figure 4-26: Layer – Strain and Load – Strain, Panel 1, SPOT 1

Figure 4-27: Layer – Strain Plots, Panel-1, SPOT 2 and 4

Figure 4-28: Layer – Strain Plots, Panel 1, SPOT 3 and 5

Figure 4-29: Layer – Strain Plots, Panel-2, SPOT 2 and 4

Figure 4-30: Layer – Strain and Load – Strain, Panel-2, SPOT 1

Figure 4-31: Layer – Strain Plots, Panel-2, SPOT 3 and 5

Figure 4-32: Load – Strain Plots, Panel-2, Stiffener

Figure 4-33: Layer – Strain Plots, Panel-3, SPOT 2 and 4

Figure 4-34: Layer – Strain and Load – Strain, Panel-3, SPOT 1

Figure 4-35: Layer – Strain Plots, Panel-3, SPOT 3 and 5

Figure 4-36: Load – Strain Plot, Panel-3, Stiffeners

Figure 4-37: Maximum Deflection Comparison

Figure 4-38: Load – Deflection at the Centre of the Panels

Figure 4-39: Load – Strain, Centre of Stiffeners

Figure 4-40: Strain Comparison of the SPOT 5 and 3 in Layer-C

Figure 5-1: Off-Axis Loading of a Unidirectional Lamina [22]

Figure 5-2: Reaction Forces along the edges at a load of 200 kPa

Figure 5-3: Reaction Moments along the edges at a load of 200 kPa

Figure 6-1: Stacking Sequence simulated in ANSYS

Figure 6-2: Deflection of Panel-1, in m

Figure 6-3: Strains in 0° direction, Layer A and Layer C, Panel-1

Figure 6-4: Strains in 90° direction, Layer B and Layer D, Panel-1

Figure 6-5: von Mises Stress, Layer A and Layer C, Panel-1

Figure 6-6: von Mises Stress, Layer B and Layer D, Panel-1

Figure 6-7: Shear Stress in Layer D, XY and XZ Direction, Panel-1

Figure 6-8: Deflection of Panel-2, in m

Figure 6-9: Strains in 0° direction, Layer A and Layer C, Panel-2

Figure 6-10: Strains in 90° direction, Layer B and Layer D, Panel-2

Figure 6-11: von Mises Stress, Layer A and Layer C, Panel-2

Figure 6-12: von Mises Stress, Layer B and Layer D, Panel-2

Figure 6-13: Shear Stress in Layer D, XY and XZ Direction, Panel-2

Figure 6-14: Deflection of Panel-3, in m

Figure 6-15: Strains in 0° direction, Layer A and Layer C, Panel-3

Figure 6-16: Strains in 90° direction, Layer B and Layer D, Panel-2

Figure 6-17: von Mises Stress, Layer A and Layer C, Panel-3

Figure 6-18: von Mises Stress, Layer B and Layer D, Panel-3

Figure 6-19: Shear Stress in Layer D, XY and XZ Direction, Panel-3

Figure 6-20: FEA – Deflection Comparison at Failure Loads

Figure 7-1: Deflection Comparison FEA – Experiment

Figure 7-2: Bending – and Membrane Strain Relationship [25]

Figure 7-3: von Mises Layer stresses in Panel-1, load 200 kPa

Figure 7-4: von Mises Layer stresses in Panel-2, load 130 kPa

Figure 7-5: von Mises Layer stresses in Panel-3, load 190 kPa

List of Tables

Table 3-1: Cantilever Beam Characteristics

Table 3-2: E-Modulus from different Data Sources

Table 4-1: C ommon Strain Measuring Methods

Table 4-2: Basic Panel Characteristics

Table 4-3: Mechanical Properties of Glass Fibres used for the Panels

Table 4-4: Mechanical Properties of Epoxy Resin used for the Panels

Table 5-1: Lamina Material Properties

Table 6-1: Results of FEA

Table 7-1: Maximum Strain and Deflection Comparison

Table 7-2: von Mises stresses FEA and Experiment, Panel-1, 200 kPa SPOT 1

Table 7-3: von Mises Stresses and Failure Criteria for Panel-1

1. Introduction

With the increased use of composite materials in various structural applications, the subject of analysing the properties of composite unstiffened and stiffened panels has received widespread attention. Laminated panels are considered as the basic modules of high performance boats, aircraft and many other complex structures, which require less specific weight, better durability, and excellent damage tolerance and are often subject to air-blast loading or under water shock. The failure mode of fibre reinforced composite materials is rather more complex than that of isotropic material. This is because of the different properties of fibres and matrix the composite is made of. There has been a number of failure theories developed since the first industrial usage of composites in the early 1980s. Those theories are based on different failure criteria and often just cover some special set up. In the last years computer simulations based on FEA were set up to predict the failure of composites. Because there is only few experimental data available for comparison, the need for experimental investigations of composites is big.

2. Basics of Composite Structures

The laminated panels are made of fibres reinforced composite material. This means while using two different materials with specific properties a new one is created with the combination and partly improvement of the former separated material properties. When building the composite material normally the final row product is build at the same time. Therefore, the material can be designed in the best way to fulfil the needs of the product. A matrix that is temperature and acid resistant (ceramic or glass matrices) or special fibres with extremely high tensile modulus (boron: 379-414 Gpa) [4] can be chosen. Besides that, the fibre direction can be arranged to take the maximum tensile load.

In a composite, the fibre carries the most of the tensile load of a composite structure and has to take the tensile stress out of the matrix. The Matrix has the job to transfer stresses between and to the fibres, to provide a barrier against an adverse environment and to protect the surface of the fibres from mechanical damage. The matrix plays a minor role in the tensile load carrying of a composite structure. On the other hand the matrix supports the fibre against buckling and provides most of the interlaminar shear strength. In addition, though fibres are strong, they can be brittle. The matrix can absorb energy by deforming under stress. This is to say, the matrix adds toughness to the composite.

2.1 Common Fibres

The most commonly used reinforcing fibre is the glass fibre. Due to the low cost and the easy usage of glass fibres, it has been considered for building the specimens and is described separately. However, more expensive fibres available on market are as following.

The Aramid Fibre, precisely Para-aramid, has an excellent stability against temperature changes and tensile strength up to seven times that of steel wire. That is why this fibre is often used in the car tire industry and for high performance ropes. The most known aramid fibre is Kevlar, which has been used reinforcement in many marine and aerospace applications where high tensile strength and impact damage resistance is needed. The major disadvantages of aramid composites are the low compressive strength and the difficulty in machining and cutting them.

The Graphite Fibre, also called carbon fibre, can be purchased in a wide range of tensile module (from 207 GPa up to 1035 GPa). Their advantages are the high fatigue strength and the exceptional high tensile strength-to-weight ratio. On the other side are the high costs, the low impact resistance and the high electric conductivity.

The Boron Fibre is very expensive but has because of it relatively large diameter a good resistance to buckling which gives boron fibre-reinforced composites a high compressive strength. The tensile modulus ranges from 379 Gpa to 414 Gpa. Ceramic Fibres, like SiC and AL2O3, have the special advantage of being resistant against high temperatures. Both fibres can be used to reinforce metal matrices [4].

2.2 Glass Fibre

The glass fibre is an inorganic, synthetic, multifilament material. Glass fibres are advantageous because of the low cost, non-flammable, electrically nonconductive, corrosion-resistant and strength characteristics. The major disadvantages are the low tensile modulus, the relatively high specific gravity, the sensitivity to abrasion with handling, the relatively low fatigue resistance and the high hardness.

Melting the raw material in a high temperature furnace and then drawing the molten material into filaments makes Glass fibres. Glass fibres are amorphous and isotropic and are long, three-dimensional networks of silicon, oxygen, and other atoms arranged in a random fashion. Glass Fibres Forms can be Fibres, Roving, Chopped Strands, Yarns, Fabrics, Mats and Fillers.

2.3 Common Matrices

The primary consideration in the selection of a matrix is its basic mechanical properties that include tensile modulus, tensile strength, and fracture toughness.

A Polymeric Matrix is made of a polymeric material, which is a collection of large numbers of polymer molecules of similar chemical structure. In solid state, the molecules are frozen in space. In a thermoplastic polymer, the linear molecules have no chemical linking between them and the thermoplastic polymer can be softened, melted and reshaped as often as desired by applying heat and pressure to weaken the intermolecular forces. In a thermoset polymer, the molecules are chemical linked and once these joints are formed the polymer cannot be melted and reshaped again.

Metal Matrix Composites are based most commonly on aluminium and titanium. They are mostly reinforced with continuous fibres, particulates, or whiskers. Metal matrix composites have superior mechanical properties and often-unique physical characteristics. Because of the costs, these materials are rarely used. Applications can be found in aerospace engine components or high-end automobile engine cylinders.

Ceramic Matrix Composites are made of glass or ceramic matrices reinforced with fibres, whiskers or particulates. Currently they have limited high-temperature applications, but because of their low density, high modulus, strength and toughness they have a big potential as a class of advanced engineering structural material [4].

2.4 Epoxy Matrix

Epoxy resin is defined as a molecule containing more than one epoxide groups. The epoxide group also termed as, oxirane or ethoxyline group, is shown in Figure 2-1. The epoxy resin is created in a chemical reaction of a di-epoxide with a diamine where the structure becomes cross-linked.

illustration not visible in this excerpt

Figure 2-1: Chemical Structure of an Epoxy Matrix [14]

These resins are thermosetting polymers and are used as adhesives, high performance coatings and potting and encapsulating materials. These resins have excellent electrical properties, low shrinkage, good adhesion to many metals and resistance to moisture, thermal and mechanical shock.

Viscosity, epoxide equivalent weight and molecular weight are the important properties of epoxy resins. Besides that, they have a high static and dynamic strength and they are electrically nonconductive.

The curing process is a chemical reaction in which the epoxide groups in epoxy resin reacts with a curing agent (hardener) to form a highly cross-linked, three-dimensional network. In order to convert epoxy resins into a hard, infusible, and rigid material, it is necessary to cure the resin with hardener. Epoxy resins cure quickly and easily at almost any temperature from 5-150oC depending on the choice of curing agent. [14]

2.5 Fibre – Matrix Interaction

For the mechanical and physical properties of composites the properties and structure of the fibre – matrix interaction are most important. The large differences in the elastic properties of the two origin materials communicate through an interface. This means the stresses of the matrix are transmitted to the fibres across the interface. This so called interface is a chemical, physical and mechanical bond between the matrix and fibres. In order to use the advantages of the high stiffness and strength of the fibres the bond has to be as strong as possible, but this makes the material very brittle. On the other hand composites with weak interfaces have relatively low stiffness and strength but high fracture resistance.

The interface depends on the atomic and chemical properties of the fibres and the matrix. Therefore it is specific for each fibre-matrix system. The adhesion between the two materials can be wetting, interdiffusion, electrostatic attraction, chemical bonding and mechanical adhesion. Most often the interface is complicated and a combination of all above mentioned bonding forms [15].

2.6 Failure Types of Composite Materials

The failure of a lamina in a laminate does not signify ultimate failure, as the failure of the laminate is progressive. It can be said that under an increasing load the part of the matrix or the fibre with the highest stress will break first. The stress distribution in the composite will change and the load can be increased further until the next most stressed part fails. This change of stress distribution inside the laminate will continue until the ultimate laminate failure. The types of failure can be divided into micro- and macro failures.

illustration not visible in this excerpt

Figure 2-2: Failure Appearance in a Bended Cross Ply Composite [15]

In case of long beams or panels, pressure energy is absorbed through global bending of the composite. This allows the energy to be transferred to locations away from the surface. Often the damage and the loss up to 50% in load carrying capability are undetectable to human eye. It is the intention to receive enough data from the embedded strain gauges to localize strain jumps in the same layer and between neighboured layers. Combined with acoustic and optic observations it should be possible to characterize the type and location of failure.

2.6.1 Micro Failure Mechanism

Fibre and tow breakage appear at the spots of maximum stress because of fibre diameter differences or imperfect fibre straightness. The load of the broken fibres will be redistributed by high interfacial shear stress in the matrix to neighbouring fibres. As a result the local tensile strength of the lamina is lowered in proportion of the number of failed fibres, but the compressive strength is not much affected because of the self steadying of the parted fibre ends.

Micro cracks in the matrix could be due mechanical stress, environmental reasons or fatigue. They can be divided in matrix hair cracks and failure of the fibre – matrix interface. Increasing numbers of those cracks usually start weakening the laminate.

Micro failures are very difficult to analyse. Unless they are in the end stadium and have grown to damages of visible size, a human eye can not see them. Usually cracks and fibre failure can be heard as small cracking noises. Often it is not possible to say where the noise appeared. Larger matrix cracks can be seen as with stripes in the matrix. Tow breakage often show as with strips following the fibre direction. The stress in the surrounding matrix and fibres grows when a fibre breaks. Fibre failure creates a short sharp noise and matrix failure is more an ongoing light cracking sound.

2.6.2 Macro Failure Mechanism

Matrix cracks in single lamina appear between neighbouring fibres of the same lamina over the whole thickness of the ply. Usually they are parallel and start suddenly. The dynamic start and progression (domino effect) often weakens neighboured plies too.

Delamination is the debounding of neighbouring plies. There are two different types of delaminations, inside delamination and delaminations on free edges.

Inside Delamination are often started by impact, voids or micro delaminations. Delaminations grow and the weaker ply will buckle. After that the delamination will grow even faster. The laminate will fail after the redistribution of the stress, then the remaining plies will fail.

Delaminations on free edges start because of different contraction rates of different layers. These can be started at the curing process and be increased by mechanical stress. They lower the static and fatigue strength of the laminate.

Macro failures usually can be seen by the human eye. They normally occur together with larger noise. Delaminations are rather large and are often white spots with round edges often near voids and areas of great shear stress. If a delamination appears between layers close to the surface of the material, often fibre buckling combined with bumps or dents on the surface can be observed. Delaminated layers carry less strain and in the remaining plies growth of stresses can be observed.

3. Experimental Investigation of a Cantilever Beam

With an intention to use strain gauges to predict the ply strains within the composite, the strain gauges and the needed cables are to be embedded in the laminate. The sensors and the cables were embedded in the matrix, in this case an epoxy-resin. The experiment has been organized to investigate and identify possible problems due to embedding strain gauges and cables inside a laminate. Thos problems could be disconnection of the wires or lengthening with the result of resistance changes of the wire because of excessive bending. Besides that it was to get used to the measurement setup and the process of building a composite panel.

3.1 Setup and Procedure

To keep this experiment as simple as possible a cantilever beam with a concentrated loading at the free end has been made. The specimen is made of unidirectional glass fibres combined with an epoxy resin. The laminate consists of five 0° layers made in hand – lay up process. In the laminate 6 strain gauges were included at the location of the biggest bending moment at the clamped end. There were always two gauges per layer, to be able to compare the effect of insulated and non insulated wires.

illustration not visible in this excerpt

Figure 3-1: Cross Section of Laminate Including Strain Gauges

On the top surface the cable of one gauge were inside the resin and the ones of the other gauge were going straight out of the composite, as shown in Figure 3-2.

While applying the incremental load the deflection and the strains in all six gauges were recorded. The strain gauges used, were from Micro-Measurements Group Inc. and of the type CEA-06-250UT-120. These are multi purpose gauges, coated with plastic and large solder taps. The k-factor is 2.07 and the resistance is 120 Ohm. The used cables are copper type and have vinyl insulation. For measuring the strain, a

illustration not visible in this excerpt

portable P-350 instrument from Micro-Measurements Group Inc. was used and the strain gauges were connected in a quarter bridge.

Figure 3-2: Top and Side View of Cantilever Beam with Strain Gauges

illustration not visible in this excerpt

Figure 3-3: Specimen before Cutting

illustration not visible in this excerpt

Table 3-1: Cantilever Beam Characteristics

illustration not visible in this excerpt

3.2 Results and Discussion

While applying the concentrated load the deflection at the loading point and the strains at the point of the biggest bending moment on three different layers were recorded. Light noises of cracking were noticed at a load of 6780g. A short break of 5 minutes was made and after the loading resumed, it was observed that the strain at gauge 5 and 6 had changed from compression to tension. This break shows that the composite material settles and creeps. After increasing the load to 7520g the cracking started again louder and first delaminations on the bottom combined with fibre buckling of layer 1 near the solder taps were spotted. While increasing the load up to 9265g almost constant light cracking noise was heard and thin white strips appeared on the surface near the clamping end. The delamination at the bottom had grown all across the width. After increasing the load beyond 10679g the load steps were about 1400g. As a result the beam totally failed at the load of 14965g. The delamination in layers 1 and 2 increased while applying the load. At a load of 12265g layer 3 the neutral layer started to buckle as shown in Figure 3-4.

illustration not visible in this excerpt

Figure 3-4: Structure Failure and Delamination of Layer 1, 2 and 3

It can be observed that gauges placed at the same layer reported similar strains. Interestingly a change from compression to tension in gauge 6 at the loading of 6780g has been observed. Although close to the neutral layer the change indicates a massive adjustment to the new load conditions in the material as shown in all plots.

The first structural damage appeared at the bottom side in layer 1 near the solder pats of the gauges. Here the connection between layer 1 and 2 was weakest and delamination started. The inter lamina bonding forces were smallest and the fibres of the layer 1 started to buckle. The strain gauges and especially the solder taps weaken the material.

illustration not visible in this excerpt

Figure 3-5: Load – Strain Plot for all six Gauges

On top of layer 5 matrix failure were noticed at the cross section of two cables. This failure did not grow while increasing the load. After the total failure of the beam it could be seen, that the three bottom layers had failed in compression and delamination occurred. Layer 1 and 2 had delaminated and buckled for about 10mm. Even layer 3, the neutral layer had buckled because of the compression.

illustration not visible in this excerpt

Figure 3-6: Stress – Strain Plot of all six Gauges

3.3 E-Modulus

For calculations only the results of gauge 1, 2, 3 and 4 up to the weight of 6870g were used, because there after the deflection was too high. This caused moment changes and nonlinear material behaviour. To check the experimental results, the E-Modulus was calculated based on the deflection and the moment. To compare different ways to calculate the E1-Modulus of the specimen all the stress–strain curves were drawn in. It shows very clearly that the differences are not big. The results are between 11.4 GPa at gauge 2 and 22.1 GPa at gauge 4, see Table 3-2. Using the formulas for the bending of a cantilever beam the E-Modulus could be calculated for all load steps from the measured deflection and the applied moments. The E-Moduli were calculated from the strains as below:

illustration not visible in this excerpt

where s = bending stress, M = bending moment, I = moment of inertia of area, t = thickness, e = measured strain and c is the distance from surface to the strain gauge as shown in Table 3-2. Using the formulas for the bending of a cantilever beam the E-Modulus could be calculated for all load steps from the measured deflection and the applied moments.

illustration not visible in this excerpt

where FN is the gravity force of the weight and y the deflection. The average of only the first six load steps was used for the comparison, because thereafter the deflection was too high causing nonlinear material behaviour. The E-modulus – “Material properties” was calculated using the given material properties of the matrix and fibres as shown in Table 3-1.

The calculation is based on the rule of mixture given below:

illustration not visible in this excerpt

where Ef and Em the E-modulus and Vf and Vm the fibre volume ratios of fibre and matrix respectively. The complete calculation is documented in Appendix B-1 The plots of the stress – strain results in addition to the curves of the calculated E-modulus are shown in Figure 3-7. As shown in Table 3-2 the different E-Moduli differ from 11.4 GPa to 22.1 GPa. These differences can be explained by strain gauge alignment errors, wrong material properties and not precise enough loading assumptions.

Table 3-2: E-Modulus from different Data Sources

illustration not visible in this excerpt

Figure 3-7: Plots of all E-moduli

3.4 Conclusion

The intention of using the strain gauges inside a glass fibre – epoxy laminate to measure the different layer strains is sense full and working. Based on the results it can be concluded that the measured strains and the resulting E-Modulus are in good agreement with the possible range of the material. It can also be stated that the embedded strain gauges show the change of strain and stress over the thickness of the beam. With more accuracy in the application of the gauges it should be possible to obtain more precise measurements. The differences between two gauges at the same layer show that one major error source is the application. The orientation and levelling of the gauge have to be more accurate. In addition the solder taps have to be as small as possible and the cables should not cross each other so that structural weak points are minimized.

There appeared no hindrances with disconnections of gauges and cables within the laminate. Despite of that there seems to be a connection between not insulated and greater strain. The cabling at gauges 3 and 6 was partly not insulated and the cables on gauge 1 had a much shorter way inside the laminate. Those three gauges showed smaller strains then the other three on the same layers. Since the cables change the length while bending, the resistance of wire and with this the measured strain changes. Therefore the use of insulated cables seems to minimise the effect. On the other hand the weakening or strengthening of the structure because of the thicker cables will be higher.

The prediction of failure and the failure type of this composite based on the theoretical approach could not be made, because of the insufficient material data. It can be said that in tension, fibre failure will occur first, because the ultimate strain eTF of the fibre is smaller than the ultimate strain eTM of the matrix in tension.

illustration not visible in this excerpt

The matrix failure on the cable crossing where a cable acted like stiff pieces inside a flexible matrix can be explained by the matrix tensile strength of 55 MPa which is much lower than the occurred stress.

Taking the symmetry over the thickness of the beam with 5 layers, the middle one is the neutral layer. There should not be any stress at all in this layer. As the strain curves measured on top of the middle layer by gauge 5 and 6 show, the maximum longitudinal strain here was small, about a 0.006 or 0.6%, compared to the maximum strain at the surface. Because of the symmetry the compressive stress on the bottom should be the same value as the tensile with a change of sign. The laminate failed in compression as the approximation in equation (3-9) predicted. The calculated ultimate compression strength of the composite is smaller than 352 MPa.

illustration not visible in this excerpt

Figure 3-8: Compression Failure at the Bottom of the Specimen

The failure probably started because of the weakened material near the solder taps. The fibres were not able to run straight and the connection between the layers was not strong because of the strain gauges, as shown in Figure 3-8. The failure type was delamination and macro fibre buckling.

4. Experimental Investigation of Composite Panels

In this experimental investigation three panels made out of fibre reinforced plastic were clamped at all edges, applied with a uniform transversal load and tested for ultimate failure. Since using those panels in aircrafts and ships the possible failure could be catastrophic, too large deflection or cracks over the whole thickness of the panel. The panels were produced with none, one or two stiffeners.

4.1 Problem Statement

Failure mode and damage progression as a function of applied load were also studied, in addition to the strength and the response of the panels during loading. Various geometries and ply orientations were selected for study, and the response of the composites was measured from initial elastic response to the final ultimate failure. It was desired o find:

1. The load-deflection bending response
2. The load-ply failure relation
3. The damage progression before final failure
4. The ultimate load of failure for each specimen
5. The corresponding modes of failure
6. The effect of stiffeners

The possible difficulties of embedding strain gauges and cables in a laminate were investigated in the experimental investigation of a cantilever beam. It was observed that the orientation and levelling of the gauges is probably one major error source of the experiment. In addition the solder taps have to be as small as possible and the cables should not cross each other so that structural weak points are minimized.

4.1.1 Commonly Used Methods and Former Experiments

To investigate the strain behaviour of composite materials under loads several methods can be used. Each has advantages and disadvantages and Table 4-1 shows some of them. Because of the limited equipment, the decision was made to use strain gages and put them inside the laminate.

In former investigations, considerable research has been done to predict the behaviour of laminated composites under in-plane loading conditions such as tension, compression and shear. Most of those investigations based on failure theories and FE Models.

Table 4-1: C ommon Strain Measuring Methods

illustration not visible in this excerpt

Besides that, Echaabi & Trochu [17] made an experimental investigation on failure and damage progression of a graphite-epoxy composite in flexure bending test. They tested small beams made of five layers in three point bending and documented the deflection and the failure mode with C-scan and microscopic investigations. G. S. Padhi [9] investigated “Progressive failure and ultimate collapse of laminated composite plates in bending”. The panels made of E-glass (five unidirectional plies) and polyester, which were clamped on all four sides and uniformly loaded. Because only the deflection was measured and the failure was only observed by ears and eyes this experiment will go further. Lessard & Chang [13] report an experimental investigation in Part 2 of the “Damage Tolerance of Laminated Composites Containing an Open Hole and Subjected to Compressive Loadings”. Although they applied in-plane loads, the failure description was very useful. They used Dye-Enhanced X-Radiography to assess the extent of damage and to observe failure mechanisms. The most comparable investigation was done by Moy, Shenoi and Allen [25] as the Strength and stiffness of fibre-reinforced plastic plates. They used clamped edge boundary conditions and about the same size of specimen proposed in this investigation. Although they investigated more in the direction of fabrication, lay-up and aspect ratios, their observed deflections, surface strains and damage descriptions gave useful information for this experiment.

4.1.2 Method Adopted

To observe failure progression and single ply failure inside a composite it is necessary to measure the strain or stress of every single ply at diverse locations under different increments of loads. Because of the costs the only possible method to achieve this was to use strain gauges to predict the ply strains within the composite. On top the strain gauges and the needed cables had to be embedded in the laminate to measure the strain inside. The gauges and cables were applied in the epoxy-resin between the fibre layers while building the laminate in hand lay-up method. In addition for the visual observation and analyses of the failure progression digital pictures and close-ups of appearing damages were made. All noises and other appearing damage symptoms were recorded in time and kind by video and digital pictures.

4.2 Specimen Properties

Based on the conclusions of the pre-experiment of testing a cantilever beam with embedded strain gauges described in chapter 3, the specimen were prepared in the same way. It was chosen to use 4 plies of unidirectional E-Glass mats. According to the manufacture over 90% of the fibres in a lamina run parallel to one another in a particular direction. Because the specimen had a rectangular shape the longer side was defined as the 0° direction and the shorter as the 90° direction.

Table 4-2: Basic Panel Characteristics

illustration not visible in this excerpt

Figure 4-1: Specimen Layout and Strain Gauge Positions

The composite panel specimens were made out of glass fibre reinforced epoxy resin. The panels made according to the description in chapter 4.2.1. The experiments were meant to be as realistic as possible. That’s why the most common composite materials in ship building were used. A local ship yard, that manufactured the specimen, provided some standard E-Glass fibres and Epoxy resin. The material properties of the used fibres – Advantex E-glass, made by Owens Corning, are presented in Table 4-3 and the physical properties of the resin can be seen in Table 4-4. “WEST SYSTEM Brand” epoxy made from 105 resin and 206 fast drying hardener in the mix ratio of 5:1 were used for the matrix.

Table 4-3: Mechanical Properties of Glass Fibres used for the Panels

illustration not visible in this excerpt

Table 4-4: Mechanical Properties of Epoxy Resin used for the Panels

illustration not visible in this excerpt

4.2.1 Sample Preparation

illustration not visible in this excerpt

The specimens have been built at a local yacht construction business called “Millner Custom Yacht Construction” under controlled conditions in hand lay-up process, see Figure 4-2. Based on the experimental results of Moy, Shenoi and Allen [25] there is no significant effect of manufacture method and panel shape. That is why the simplest and cheapest manufacturing method – the hand lay-up process was chosen. The materials were prepared and cured according to the manufacture’s recommended procedures. Manufactured panels and stiffeners were cut to size using a band saw.

[...]