You are on page 1of 117

Composites

o o o o o o o o o o o o o http://composite.about.com/ http://www.netcomposites.com/ http://www.gurit.com/ http://www.hexcel.com/ http://www.toraycfa.com/ http://www.e-composites.com/ http://www.compositesone.com/basics.htm http://www.wwcomposites.com/ (World Wide Search Engine for Composites) http://jpsglass.com/ http://www.eirecomposites.com/ http://www.advanced-composites.co.uk/ http://www.efunda.com/formulae/solid_mechanics/composites/comp_intro.cfm

Introduction
There is an unabated quest for new materials which will satisfy the specific requirements for various applications like structural, medical, house-hold, industrial, construction, transportation, electrical; electronics, etc. The metals are the most commonly used materials in these applications. In the yore of time, there have been specific requirements on the properties of these materials. It is impossible of any material to fulfill all these properties. Hence, newer materials are developed. In the course, we are going to learn more about composite materials. First, we will deal with primary understanding of these materials and then we will learn the mechanics of these materials.

In the following lectures, we will introduce the composite materials, their evolution; constituents; fabrication; application; properties; forms, advantages-disadvantages etc. In the present lecture we will introduce the composite materials with a formal definition, need for these materials, their constituents and forms of constituents.

Definition of a Composite Material


A composite material is defined as a material which is composed of two or more materials at a microscopic scale and has chemically distinct phases.

Thus, a composite material is heterogeneous at a microscopic scale but statistically homogeneous at macroscopic scale. The materials which form the composite are also called as constituents or constituent materials. The constituent materials of a composite have significantly different properties. Further, it should be noted that the properties of the composite formed may not be obtained from these constituents. However, a combination of two or more materials with significant properties will not suffice to be called as a composite material. In general, the following conditions must be satisfied to be called as a composite material: 1. The combination of materials should result in significant property changes. One can see significant changes when one of the constituent material is in platelet or fibrous from. 2. The content of the constituents is generally more than 10% (by volume). 3. In general, property of one constituent is much greater property of the other constituent. The composite materials can be natural or artificially made materials. In the following section we will see the examples of these materials. than the corresponding

Why do we need these materials?


There is unabated thirst for new materials with improved desired properties. All the desired properties are difficult to find in a single material. For example, a material which needs high fatigue life may not be cost effective. The list of the desired properties depending upon the requirement of the application is given below. 1. Strength 2. Stiffness 3. Toughness 4. High corrosion resistance 5. High wear resistance 6. High chemical resistance 7. High environmental degradation resistance 8. Reduced weight 9. High fatigue life 10. Thermal insulation or conductivity 11. Electrical insulation or conductivity 12. Acoustic insulation

13. Radar transparency 14. Energy dissipation 15. Reduced cost 16. Attractiveness The list of desired properties is in-exhaustive. It should be noted that the most important characteristics of composite materials is that their properties are tailorable, that is, one can design the required properties.

Figure 1.1 Types of reinforcement in a composite History of Composites


The existence of composite is not new. The word composite has become very popular in recent four -five decades due to the use of modern composite materials in various applications. The composites have existed from 10000 BC. For example, one can see the article by Ashby [1]. The evolution of materials and their relative importance over the years have been depicted the Figure 1 of this article. The common composite was straw bricks, used as construction material.

Then the next composite material can be seen from Egypt around 4000 BC where fibrous composite materials were used for preparing the writing material. These were the laminated writing materials fabricated from the papyrus plant. Further, Egyptians made containers from coarse fibres that were drawn from heat softened glass.

One more important application of composites can be seen around 1200 BC from Mongols. Mongols invented the so called modern composite bow. The history shows that the early existing of composite bows dates back to 3000 BC as predicted by Angara Dating. The bow used various materials like wood, horn, sinew (tendon), leather, bamboo and antler. The horn and antler were used to make the main body of the bow as it is very flexible and resilient. Sinews were used to join and cover the horn and antler together. Glue was prepared from the bladder of fish which is used to glue all the things in place. The string of the bow was made from sinew, horse hair and silk. The composite bow so prepared used to take almost a year for fabrication. The bows were so powerful that one can throw the arrows almost 1.5 km

away. Until the discovery of gun-powder the composite bow used to be a very lethal weapon as it was a short and handy weapon. As said, Need is the mother of all inventions, the modern composites, that is, polymer composites came into existence during the Second World War. During the Second World War due to constraint impositions on various nations for crossing boundaries as well as importing and exporting the materials, there was scarcity of materials, especially in the military applications. During this period the fighter planes were the most advanced fighting means. The light weight yet strong materials were in high demand. Further, for application like housing of electronic radar equipments require non-metallic materials. Hence, the Glass Fibre Reinforced Plastics (GFRP) were first used in these applications. Phenolic resins were used as the matrix material. The first use of composite laminates can be seen in the Havilland Mosquito Bomber of the British Royal Air Force.

The composites exit in day to day life applications as well. The most common existence is in the form of concrete. The concrete is a composite made from gravel, sand and cement. Further, when it is used along with steel to form structural components in construction, it forms one further form of composite. The other material is wood which is a composite made from cellulose and lignin. The advanced forms of wood composites can be ply-woods. These can be particle bonded composites or mixture of wooden planks/blocks with some binding agent. Now a days, these are widely used in the furniture and construction materials.

An excellent example of natural composite is muscles of human body. The muscles are present in a layered system consisting of fibers at different orientations and in different concentrations. These result in a very strong, efficient, versatile and adaptable structure. The muscles impart strength to bones and vice a versa. These two together form a structure that is unique. The bone itself is a composite structure. The bone contains mineral matrix material which binds the collagen fibres together.

The other examples include: wings of a bird, fins of a fish, trees and grass. A leaf of a tree is also an excellent example of composite structure. The veins in the leaf not only transport the food and water but also impart the strength to the leaf so that the leaf remains stretched with maximum surface area. This helps the plant to extract more energy from sun during photo-synthesis.

What are the constituents in a typical composite?


In a composite, typically, there are two constituents. One of the constituent acts as a reinforcement and other acts as a matrix. Sometimes, the constituents are also referred as phases.

What are the types of reinforcements?


The reinforcements in a composite material come in various forms. These are depicted through Figure 1.1.

1. Fibre: Fibre is an individual filament of the material. A filament with length to diameter ratio above 1000 is called as a fibre. The fibrous form of the reinforcement is widely used. The fibres can be in the following two forms: a. Continuous fibres: If the fibres used in a composite are very long and unbroken or cut then it forms a continuous fibre composite. A composite, thus formed using continuous fibres is called as fibrous composite. The fibrous composite is the widely used form of composite. b. Short/chopped fibres: The fibres are chopped into small pieces when used in fabricating a composite. A composite with short fibres as reinforcements is called as short fibre composite. In the fibre reinforced composites, the fibre is the major load carrying constituent. 2. Particulate: The reinforcement is in the form of particles which are of the order of a few microns in the diameter. The particles are generally added to increase the modulus and decrease the ductility of the matrix materials. In this case, the load is shared by both particles and matrix materials. However, the load shared by the particles is much larger than the matrix material. For, example in an automobile type carbon black (as a particulate reinforcement) is added in rubber (as matrix material). The composite with reinforcement in particle form is called as particulate composite. 3. Flake: Flake is a small, flat, thin piece or layer (or a chip) that is broken from a larger piece. Since these are two dimensional in geometry, they impart almost equal strength in all directions of their planes. Thus, these are very effective reinforcement components. The flakes can be packed more densely when they are laid parallel, even denser than unidirectional fibres and spheres. For example, aluminum flakes are used in paints. They align themselves parallel to the surface of the coating which imparts the good properties. 4. Whiskers: These are nearly perfect single crystal fibres. These are short, discontinuous and polygonal in cross-section.

5.

6. The classification of composites based on the form of reinforcement is shown in Figure 1.2. The detailed classification further is given in Figure 1.3. The classification of particulate composites is depicted further in Figure 1.4. Some of the terms used in these classifications will be explained in the following paragraphs/lectures.

Figure 1.2: Classification of composites based on reinforcement type

8.

Figure 1.3: Classification of fibre composite materials

10.

Figure 1.4: Classification of particulate composites

Why is the reinforcement made in thin fibre form?


There are various reasons because of which the reinforcement is made in thin fibre form. These reasons are given below. a) An important experimental study by Leonardo da Vinci on the tensile strength of iron wires of various lengths (see references in [2, 3]) is well known to us. In this study it was revealed that the wires of same diameter with shorter length showed higher tensile strength than those longer lengths. The reason for this is the fact that the number of flaws in a shorter length of wire is small as compared longer length. Further, it is well known that the strength of a bulk material is very less than the strength of the same material in wire form.

The same fact has been explored in the composites with reinforcement in fibre form. As the fibres are made of thin diameter, the inherent flaws in the material decrease. Hence, the strength of the fibre increases as the fibre diameter decreases. This kind of experimental study has revealed the similar results [2, 3]. This has been shown in Figure 1.5 qualitatively.

Figure 1.5: Qualitative variation of fibre tensile strength with fibre diameter

b) The quality of load transfer between fibre and matrix depends upon the surface area between fibre and matrix. If the surface area between fibre and matrix is more, better is the load transfer. It can be shown that for given volume of fibres in a composite, the surface area between fibre and matrix increases if the fibre diameter decreases.

Let

be the average diameter of the fibres,

be the length of the fibres and

be the number of

fibres for a given volume of fibres in a composite. Then the surface area available for load transfer is

(1.1) The volume of these fibres in a composite is

(1.2)

Now, let us replace the fibres with a smaller average diameter of

such that the volume of the fibres

is unchanged. Then the number of fibres required to maintain the same fibre volume is

(1.3)

The new surface area between fibre and matrix is

(1.4)

Thus, for a given volume of fibres in a composite, the area between fibre and matrix is inversely proportional to the average diameter of the fibres. c) The fibres should be flexible so that they can be bent easily without breaking. This property of the fibres is very important for woven composites. In woven composites the flexibility of fibres plays an important role. Ultra thin composites are used in deployable structures.

The flexibility is simply the inverse of the bending stiffness. From mechanics of solids study the bending stiffness is EI, where is Youngs modulus of the material and is the second moment of

area of the cross section of the fibre. For a cylindrical fibre, the second moment of area is

(1.5) Thus,

Flexibility

(1.6)

Thus, from the above equation it is clear that if a fibre is thin, that is, small in diameter it is more flexible

Boron Fiber
This fibre was first introduced by Talley in 1959 [15]. In commercial production of boron fibres, the method of Chemical Vapour Deposition (CVD) is used. The CVD is a process in which one material is deposited onto a substrate to produce near theoretical density and small grain size for the deposited material. In CVD the material is deposited on a thin filament. The material grows on this substrate and produces a thicker filament. The size of the final filament is such that it could not be produced by drawing or other conventional methods of producing fibres. It is the fine and dense structure of the deposited material which determines the strength and modulus of the fibre.

In the fabrication of boron fibre by CVD, the boron trichloride is mixed with hydrogen and boron is deposited according to the reaction

In the process, the passage takes place for couple of minutes. During this process, the atoms diffuse into tungsten core to produce the complete boridization and the production of beginning the tungsten fibre of 12 diameter is used, which increases to 12 and . In the . This step induces

significant residual stresses in the fibre. The core is subjected to compression and the neighbouring boron mantle is subjected to tension.

The CVD method for boron fibres is shown in Figure 1.7.

Figure 1.7: Schematic of reactors for silicon carbide fibres by Chemical Vapour Deposition The key features of this fibre are listed below:

These are ceramic monofilament fiber. Fiber itself is a composite. Circular cross section. Fiber diameter ranges between 33-400 and typical diameter is 140 .

Boron is brittle hence large diameter results in lower flexibility. Thermal coefficient mismatch between boron and tungsten results in thermal residual stresses during fabrication cool down to room temperature.

Boron fibres are usually coated with SiC or

so that it protects the surface during contact

with molten metal when it is used to reinforce light alloys. Further, it avoids the chemical reaction between the molten metal and fibre. Strong in both tension and compression. Exhibits linear axial stress-strain relationship up to 650 .

Since this fibre requires a specialized procedure for fabrication, the cost of production is relatively high.

The boron fibre structure and its composite is elucidated in Figure 1.8.

Figure 1.8: Boron fibre structure and its composite

CarbonFiber:
The first carbon fibre for commercial use was fabricated by Thomas Edison. Sixth lightest element and carbon- carbon covalent bond is the strongest in nature. Edison made carbon fiber from bamboo fibers. Bamboo fiber is made up of cellulose. Precursor fiber is carbonized rather than melting.

Filaments are made by controlled pyrolysis (chemical deposition by heat) of a precursor material in fiber form by heat treatment at temperature 1000-3000

The carbon content in carbon fibers is about 80-90 % and in Graphite fibers the carbon content is in excess of 99%. Carbon fibre is produced at about 1300 in excess of 1900 . . while the graphite fibre is produced

The carbon fibers become graphitized by heat treatment at temperature above 1800 Carbon fibers term is used for both carbon fibers and graphite fibers. Different fibers have different morphology, origin, size and shape The size of individual filament ranges from 3 to 147 . to 2000 .

Maximum use of temperature of the fibers ranges from 250

The use temperature of a composite is controlled by the use temperature of the matrix. Precursor materials: There are two types of precursor materials (i) Polyacrylonitrile (PAN) and (ii) rayon pitch, that is, the residue of petroleum refining. Fiber properties vary with varying temperature. Fiber diameter ranges from 4 to10 .

A tow consists of about 3000 to 30000 filaments. Small diameter results in very flexible fiber and can actually be tied in to a knot without breaking the fiber. Modulus and strength is controlled by the process. The procedure involves the thermal decomposition of the organic precursor under well controlled conditions of temperature and stress.

Cross section of fiber is non-circular, in general, it is kidney bean shape. Heterogeneous microstructure consisting of numerous lamellar ribbons. Morphology is very dependent on the manufacturing process. PAN based carbon fibers typically have an onion skin appearance with the basal planes in more or less circular arcs, whereas the morphology of pitch-based fiber in such that the basal planes lie along radial planes. Thus, carbon fibers are anisotropic.

Glass Fibre
Fibers of glass are produced by extruding molten glass, at a temperature around 1200 through holes in a spinneret with diameter of 1 or 2 mm and then drawing the filaments to

produce fibers having diameters usually between 5 to15 The fibres have low modulus but significantly higher stiffness

Individual filaments are small in diameters, isotropic and very flexible as the diameter is small. The glass fibres come in variety of forms based on silica elements to create speciality glass. which is combined with other

What are the different types of glass fibres? What are their key features?
The types of glass fibres and their key features are as follows: E glass - high strength and high resistivity S2 glass - high strength, modulus and stability under extreme temperature and corrosive environment. R glass enhanced mechanical properties C glass - resists corrosion in an acid environment D glass good dielectric properties

Introduction
In this lecture we are going see some more advanced fibres. Further, we will see their key features, applications and fabrication processes.

Alumina Fibre
These are ceramics fabricated by spinning a slurry mix of alumina particles and additives to form a yarn which is then subjected to controlled heating. Fibers retain strength at high temperature. It also shows good electrical insulation at high temperatures. It has good wear resistance and high hardness. The upper continuous use temperature is about 1700 .

Fibers of glass, carbon and alumina are supplied in the form of tows (also called ravings or strands) consisting of many individual continuous fiber filaments. Du Pont has developed a commercial grade alumina fibre and known as Alumina FP (polycrystalline alumina) fibre. Alumina FP fibres are compatible with both metal and resin

matrices. These fibres have a very high melting point of 2100 temperatures up to 1000

. They can withstand

without any loss of strength and stiffness properties at this elevated

temperature. They exhibit high compressive strengths, when they are set in a matrix. The Alumina whiskers are available and they exhibit excellent properties. Alumina whiskers can have the tensile strength of 20700 MPa and the tensile modulus of 427 GPa.

What are the applications of Alumina fibres?


The Alumina has a unique combination of low thermal expansion, high thermal conductivity and high compressive strength. The combination of these properties gives good thermal shock resistance. These properties make Alumina is suited for the applications in furnace use as crucibles, tubes and thermocouple sheaths. The good wear resistance and high hardness properties are harnessed in making the components such as ball valves, piston pumps and deep drawing tools.

Aramid Fibre
These fibres are from Aromatic polyamide, that is, nylons family. Aramid is derived from Ar of Aromatric and amid of polyamide. Examples of fibres from nylon family: Polyamide 6, that is, nylon 6 and Polyamide 6.6, that is, nylon 6.6 These are organic fibers. Melt-spun from a liquid solution Du Pont developed these fibers under the trade name Kevlar. From poly (p-phenylene terephthalamide (PPTA) polymer. Morphology radially arranged crystalline sheets resulting into anisotropic properties. Filament diameter about 12 High tensile strength. Intermediate modulus Significantly lower strength in compression. 5 grades of Kevlar with varying engineering properties are available. Kevlar-29, Kevlar-49, Kevlar-100, Kevlar-119 and Kevlar-129. and partially flexible

Silicon Carbide Fibre (SiC)


Silicon carbide fibres are ceramic fibers. These fibres are produced in similar fashion as boron fibres are produced. The fibres are produced by two methods as follows:

CVD on Tungsten or Carbon Core NICALON by NIPPON Carbon Japan

CVD on Tungsten or Carbon Core:


This fiber is similar in size and microstructure to boron. The fibres are produced on both tungsten and carbon cores. These fibres are relativity stiff due to thicker diameter of the fibres. The diameter of the fibres is about 140 .

The fibres have strength in the range of 3.4 4.0 GPa. Failure strain is in the range of 0.8 - 1%. The Youngs modulus is about 430 GPa. The fibres show high structural stability and strength retention even at temperatures above 1000 .

The CVD with

as the reactant, SiC is deposited on the core as follows:

The SiC fibres produced on a tungsten core with a diameter about 12

. It shows a thin

interfacial layer between the SiC mantle and the tungsten core. In case, when carbon fibre is used the fibre diameter of the carbon fibres is about 33 .

Both type of SiC fibre have smoother surfaces than a boron fibre. This is because there is a deposition of small columnar grains as compared to conical nodules in boron fibres. The SiC fibres produced with carbon core are used in light reinforced alloys. These fibres are produced with a surface coating. The composition of this coating varies from carbon rich from inner surface to silicon carbide at the outer surface.

The fibres that are used to reinforce the titanium have a protective layer which varies from a carbon rich to silicon rich and again to a composition which is rich in carbon at the surface. The outer surface acts as a protective surface and when it comes in contact with molten and highly reactive titanium. The fibres are made by Specialty Materials Inc. under the trade name SCS-6. The coating increases the fibres diameter by 6 .

The fibre has low failure stresses due to surface flaws. The higher strength of fibre is due to the controlled defects at the core-mantle interface. The strength of SiC fibres produced using CVD is seen to be anisotropic. The radial strength is significantly lower than longitudinal tensile strength. When the fibres are heated to above 800 C in air for a long period, they lose their strength due to oxidation of the carbon rich outer layers.

The CVD for SiC fibres is shown in Figure 1.9.

Figure 1.9: Schematic of reactors for silicon carbide fibres by Chemical Vapour Deposition

NICALON by NIPPON Carbon Japan


The fibres are manufactured by the process of controlled pyrolysis (chemical deposition by heat) of a polymeric precursor. The fiber is homogeneously composed of ultrafine beta-SiC crystallites and carbon. The filament is similar to carbon fiber in size. The diameter of the fibre is about 14 The fibres more flexible due to small diameter. The fibres arranged in tows of 250 to 500 filaments per tow. These fibres come in two grades: a. Ceramic Grade: provides good high temperature performance and mechanical properties b. High Volume Resistivity Grade: It is a low dielectric fibre. It has good electrical and mechanical properties. These are used in dielectric structures.

Uses of the NICALON Fibres


These fibres are used to form fibrous products such as high temperature insulation, filters, etc. These fibres have high resistance to chemical attack. Hence, these can be used in harsh environments.

These are also used as a reinforcement in plastic, ceramic and metal matrix composites.

Cross Sectional Shapes of Fibres


The cross sectional shapes of fibre of various types we have studied above are different. The cross sectional shape of the fibres, although is assumed to be circular, is not circular in general. The various cross sectional shapes of the fibre are shown in Figure 1.10.

Figure 1.10: Cross sectional shapes of fibres

Fiber Properties

The following are the important points regarding the fibre properties. Density, axial modulus, axial Poissons ratio, axial tensile strength and coefficient of thermal expansion are some of the important properties. Advanced fibers exhibit a broad range of properties. Properties of carbon fiber can vary significantly depending upon fabrication process. For the advanced fibres studied above one can attain either high modulus (> 700 GPa) or high strength (> 5 GPa) but not both attainable simultaneously. SCS-6, IM8, boron and sapphire fibers offer the best combination of stiffness and strength but have large diameters and thus limited flexibility. However, IM8 fibers are exception for flexibility. The specific stiffness of some of these fibres is almost 13 times of structural metals. Similarly, the specific strength of some of these fibres is almost 16 times of structural metals. Weight saving, when the composites of these fibres are used, is tremendous due to high specific stiffness and strength. Actual properties of composite (fiber + matrix) are reduced. Specific properties are reduced even further when the loading is in a direction other than the length direction of fibers. Tailorable properties. One can get the desired heat transfer or electrical conductivity with proper designing. The increased fatigue resistance is attainable with the use of these fibre composites. Aging effect can be significantly lowered.

Note: The fibres are classified based on their values of modulus as follows: 1. Ultra-high-modulus, type UHM (modulus > 450 GPa) 2. High-modulus, type HM (modulus between 350-450 GPa) 3. Intermediate-modulus, type IM (modulus between 200-350 GPa) 4. Low modulus and high-tensile, type HT (modulus < 100 GPa, tensile strength > 3.0 GPa)

Introduction
In the previous lecture we have introduced various advanced fibres along with their fabrication processes, precursor materials and key features. In the present lecture we will introduce some matrix materials, their key features and applications.

What are the matrix materials used in composites?


The matrix materials used in composites can be broadly categorized as: Polymers, Metals, Ceramics and Carbon and Graphite.

The polymeric matrix materials are further divided into: 1. Thermoplastic which soften upon heating and can be reshaped with heat and pressure. 2. Thermoset which become cross linked during fabrication and does not soften upon reheating. The metal matrix materials are: Aluminum, Copper and Titanium.

The ceramic materials are: Carbon, Silicon carbide, Silicon nitride. The classification of matrix materials is shown in Figure 1.11.

Figure 1.11: Matrix materials

What are the thermoplastic matrix materials? What are their key features?
The following are the thermoplastic materials: 1. polypropylene, 2. polyvinyl chloride, 3. nylon, 4. polyurethane,

5. poly-ether-ether ketone (PEEK), 6. polyphenylene sulfide (PPS), 7. polysulpone. The key features of the thermoplastic matrix materials are: 1. higher toughness 2. high volume 3. low cost processing 4. The use temperature range is upto 225 .

What are the thermoset matrix materials? What are their key features?
The thermoset matrix materials are: 1. polyesters, 2. epoxies, 3. polyimides The key features of these materials are given for individual material in the following.

Polyesters
1. Used extensively with glass fibers 2. Inexpensive 3. Light weight 4. Temperature range upto 100 .

5. Resistant to environmental exposures

Epoxy
1. Expensive 2. Better moisture resistance 3. Lower shrinkage on curing 4. Use temperature is about 175

Polyimide
1. Higher use temperature about 300

2. Difficult to fabricate

What are the problems with the use of polymer matrix materials?
1. Limited temperature range. 2. Susceptibility to environmental degradation due to moisture, radiation, atomic oxygen (in space) 3. Low transverse strength. 4. High residual stress due to large mismatch in coefficients of thermal expansion between fiber and matrix. 5. Polymer matrix cannot be used near or above the glass transition temperature.

Comparison between Thermoplastics and Thermosets:


The comparison between the thermoplastic and thermoset matrix materials is given in Table 1 below: Table 1.1: Comparison between thermoplastics and thermosets.

Thermoplastics
Soften upon heat and pressure Hence, can be repaired

Thermosets
Decompose upon heating Difficult to repair Low strains are required for failure Can not be re-processed Limited shelf life Long curing cycles Tacky and therefore, difficult to handle Fair resistance to solvents Lower processing temperature

High strains are required for failure

Can be re-processed Indefinite shelf life Short curing cycles

Non tacky and easy to handle

Excellent resistance to solvents Higher processing temperature is required.

Hence, viscosities make the processing difficult.

is required.

What are the common metals used as matrix materials? What are their advantages and disadvantages?
The common metals used as matrix materials are aluminum, titanium and copper.

Advantages: 1. Higher transfer strength, 2. High toughness (in contrast with brittle behavior of polymers and ceramics) 3. The absence of moisture and 4. High thermal conductivity (copper and aluminum). Dis-advantages: 1. Heavier 2. More susceptible to interface degradation at the fiber/matrix interface and 3. Corrosion is a major problem for the metals The attractive feature of the metal matrix composites is the higher temperature use. The aluminum matrix composite can be used in the temperature range upward of 300C while the titanium matrix composites can be used above 800 .

What are the ceramic matrix materials? What are their advantages and disadvantages?
The carbon, silicon carbide and silicon nitride are ceramics and used as matrix materials.

Ceramic:
The advantages of the ceramic matrix materials are: 1. The ceramic composites have very high temperature range of above 2000 2. High elastic modulus 3. Low density The disadvantages of the ceramic matrix materials are: .

1. The ceramics are very brittle in nature. 2. Hence, they are susceptible to flows. Carbon

The advantages of the carbon matrix materials are: 1. High temperature at 2200 2. Carbon/carbon bond is stronger at elevated temperature than room temperature. The disadvantages of the carbon matrix materials are: 1. The fabrication is expensive 2. The multistage processing results in complexity and higher additional cost. It should be noted that a composite with carbon fibres as reinforcement as well as matrix material is known as carbon-carbon composite. The application of carbon-carbon composite is seen in leading edge of the space shuttle where the high temperature resistance is required. The carbon-carbon composites can resist the temperature upto 3000 .

The advantages of these composites are: 1. Very strong and light as compared to graphite fibre alone. 2. Low density 3. Excellent tensile and compressive strength 4. Low thermal conductivity 5. High fatigue resistance 6. High coefficient of friction The disadvantages include: 1. Susceptible to oxidation at elevated temperatures 2. High material and production cost 3. Low shear strength Figure 1.12 depicts the range of use temperature for matrix material in composites. It should be noted that for the structural applications the maximum use temperature is a critical parameter. This maximum temperature depends upon the maximum use temperature of the matrix materials

What are the different forms of composites?


1. Unidirectional lamina: o o o o o It is basic form of continuous fiber composites. A lamina is also called by ply or layer. Fibers are in same direction. Orthotropic in nature with different properties in principal material directions. For sufficient number of filaments (or layers) in the thickness direction, the effective properties in the transverse plane (perpendicular to the fibers) may be isotropic. Such a composite is called as transversely isotropic. 2. Woven fabrics: o o o Examples of woven fabric are clothes, baskets, hats, etc. Flexible fibers such as glass, carbon, aramid can be woven in to cloth fabric, can be impregnated with a matrix material. Different patterns of weaving are shown in Figure 1.13.

Typical weaving patterns are shown in Figure 1.13.

Figure 1.13: Types of weave

3. Laminate: 1. Stacking of unidirectional or woven fabric layers at different fiber orientations. 2. Effective properties vary with: 1. orientation 2. thickness 3. stacking sequence A symmetric laminate is shown in Figure 1.14.

Figure 1.14: A symmetric laminate

4. Hybrid composites:

The hybrid composite are composites in which two or more types of fibres are used. Collectively, these are called as hybrids. The use of two or more fibres allows the combination of desired properties from the fibres. For example, combination of aramid and carbon fibres gives excellent tensile properties of aramid and compressive properties of carbon fibers. Further, the aramid fibres are less expensive as compared to carbon fibres.

What are the advantages of the composite materials?


The following are the advantages of composites: 1. Specific stiffness and specific strength: The composite materials have high specific stiffness and strengths. Thus, these material offer better properties at lesser weight as compared to conventional materials. Due to this, one gets improved performance at reduced energy consumption. 2. Tailorable design: A large set of design parameters are available to choose from. Thus, making the design procedure more versatile. The available design parameters are:

1. Choice of materials (fiber/matrix), volume fraction of fiber and matrix, fabrication method, layer orientation, no. of layer/laminae in a given direction, thickness of individual layers, type of layers (fabric/unidirectional) stacking sequence. 2. A component can be designed to have desired properties in specific directions. 3. Fatigue Life: The composites can with stand more number of fatigue cycles than that of aluminum. The critical structural components in aircraft require high fatigue life. The use of composites in fabrication of such structural components is thus justified. 4. Dimensional Stability: Strain due to temperature can change shape, size, increase friction, wear and thermal stresses. The dimensional stability is very important in application like space antenna. For composites, with proper design it is possible to achieve almost zero coefficient of thermal expansion. 5. Corrosion Resistance: Polymer and ceramic matrix material used to make composites have high resistance to corrosion from moisture, chemicals. 6. Cost Effective Fabrication: The components fabricated from composite are cost effective with automated methods like filament winding, pultrusion and tape laying. There is a lesser wastage of the raw materials as the product is fabricated to the final product size unlike in metals. 7. Conductivity: The conductivity of the composites can be achieved to make it a insulator or a highly conducting material. For example, Glass/polyesters are non conducting materials. These materials can be used in space ladders, booms etc. where one needs higher dimensional stability, whereas copper matrix material gives a high thermal conductivity. The list of advantages of composite is quite long. One can find more on advantages of composite in reference books and open literature.

What are the disadvantages of Composites?

1. Some fabrics are very hard on tooling. 2. Hidden defects are difficult to locate. 3. Inspection may require special tools and processes. 4. Filament-wound parts may not be repairable. Repairing may introduce new problems. 5. High cost of raw materials. 6. High initial cost of tooling, production set-up, etc. 7. Labour intensive. 8. Health and safety concerns. 9. Training of the labour is essential. 10. Environmental issues like disposal and waste management. 11. Reuse of the materials is difficult. 12. Storage of frozen pre-pregs demands for additional equipments and adds to the cost of production. 13. Extreme cleanliness required. 14. The composites, in general, are brittle in nature and hence easily damageable. 15. The matrix material is weak and hence the composite has low toughness. 16. The transverse properties of lamina or laminate are, in general, weak. 17. The analysis of the composites is difficult due to heterogeneity and orthotropy.

Introduction
This lecture is dedicated to the use of composites. The use of composites is almost ubiquitous! The use of composites is an unexhaustive list. In the following we cite some important applications.

What are the applications of the composite materials?


The applications of the composites are given in the following as per the area of application.

Aerospace:
Aircraft, spacecraft, satellites, space telescopes, space shuttle, space station, missiles, boosters rockets, helicopters (due to high specific strength and stiffness) fatigue life, dimensional stability. All composite voyager aircraft flew nonstop around the world with refueling. Carbon/carbon composite is used on the leading edges nose cone of the shuttle. B2 bomber - both fiber glass and graphite fibers are used with epoxy matrix and polyimide matrix.

The indigenous Light Combat Aircraft (LCA - Tejas) has Kevlar composite in nose cone, Glass composites in tail fin and carbon composites form almost all part of the fuselage and wings, except the control surfaces of the wing.

Further, the indigenous Light Combat Helicopter (LCH Dhruvh) has carbon composites for its main rotor blades. The other composites are used in tail rotor, vertical fin, stabilizer, cowling, radome, doors, cockpit, side shells, etc.

Missile:
Rocket motor cases Nozzles Igniter Inter stage structure Equipment section Aerodynamic fairings

Launch Vehicle:
Rocket motor case Interstage structure Payload fairings and dispensers High temperature Nozzle Nose cone Control surfaces

Composite Railway Carrier:


Composite railway auto carrier Bodies of Railway Bogeys Seats Drivers Cabin Stabilization of Ballasted Rail Tracks Doors Sleepers for Railway Girder Bridges Gear Case Pantographs

Sports Equipments
Tennis rockets, golf clubs, base-ball bats, helmets, skis, hockey sticks, fishing rods, boat hulls, wind surfing boards, water skis, sails, canoes and racing shells, paddles, yachting rope, speed boat, scuba diving tanks, race cars reduced weight, maintenance, corrosion resistance.

Automotive
Lower weight and greater durability, corrosion resistance, fatigue life, wear and impact resistance. Drive shafts, fan blades and shrouds, springs, bumpers, interior panels, tires, brake shoes, clutch plates, gaskets, hoses, belts and engine parts. Carbon and glass fiber composites pultruted over on aluminum cylinder to create drive shaft. Fuel saving braking energy can be stored in to a carbon fiber super flywheels. Other applications include: mirror housings, radiator end caps, air filter housing, accelerating pedals, rear view mirrors, head-lamp housings, and intake manifolds, fuel tanks.

B. Spray Lay-Up:
Fibre is chopped in a hand-held gun and fed into a spray of catalyzed resin directed at the mould. The deposited materials are left to cure under standard atmospheric conditions. The fabrication method is depicted in Figure 1.16.

The polyester resins can be used with glass rovings is best suited for this process.

Figure 1.16: Wet or hand lay-up fabrication

Advantages:
The spray-up process offers the following advantages: o o o It is suitable for small to medium-volume parts. It is a very economical process for making small to large parts. It utilizes low-cost tooling as well as low-cost material systems.

Limitations:
The following are some of the limitations of the spray-up process: o o o o o It is not suitable for making parts that have high structural requirements. It is difficult to control the fiber volume fraction as well as the thickness. These parameters highly depend on operator skill. Because of its open mold nature, styrene emission is a concern. The process offers a good surface finish on one side and a rough surface finish on the other side. The process is not suitable for parts where dimensional accuracy and process repeatability are prime concerns. The spray-up process does not provide a good surface finish or dimensional control on both or all the sides of the product. o o o o Cores, when needed, have to be inserted manually. Only short fibres can be used in this process. Since, pressurized resin is used the laminates tend to be very resin-rich. Similar to wet/hand lay-up process, the resins need to be of low viscosity so that it can be sprayed.

Applications:
Simple enclosures, lightly loaded structural panels, e.g. caravan bodies, truck fairings, bathtubs, shower trays, some small dinghies.

C. Autoclave Curing:

The key features of this process are as follows: o o o o o An autoclave is a closed vessel for controlling temperature and pressure is used for curing polymeric matrix composites. Composites to be cured is prepared either through hand lay up or machine placement of individual laminae in the form of fibers tape which has been impregnated with resin. Components is then placed in an autoclave and subjected to a controlled cycle of temperature and pressure. After curing, the composite is solidified. One can use the fibres like carbon, glass, aramid etc. along with any resin.

Advantages:
o o o o Large components can be fabricated. Since, the curing of matrix material is carried out under controlled environment the resin distribution is better as compared to hand or spay lay-up processes. Less possibility of dilution with foreign particles. Better surface finish.

Disadvantages:
o o o Initial cost of tooling is high. Running and maintenance cost is high. Not suitable for small products.

Applications:
The process is suitable for aerospace, automobile parts like wing box, chassis, bumpers, etc

D. Filament Winding:
This process is an automated process. This process is used in the fabrication of components or structures made with flexible fibers. This process is primarily used for hollow, generally circular or oval sectioned components. Fibre tows are passed through a resin bath before being wound onto a mandrel in a variety of orientations, controlled by the fibre feeding mechanism, and rate of rotation of the mandrel. The wound component is then cured in an oven or autoclave.

One can use resins like epoxy, polyester, vinylester and phenolic along with any fibre. The fibre can be directly from creel, non-woven or stitched into a fabric form.

The filament winding process is shown in Figure 1.17.

Advantages:
o o o o Resin content is controlled by nips or dies. The process can be very fast. The process is economic. Complex fibre patterns can be attained for better load bearing of the structure.

Disadvantages:
o o o o o Resins with low viscosity are needed. The process is limited to convex shaped components. Fibre cannot easily be laid exactly along the length of a component. Mandrel costs for large components can be high. The external surface of the component is not smoothly finished.

Figure 1.17: Filament winding

Applications:
Pressure bottles, rocket motor casing, chemical storage tanks, pipelines, gas cylinders, firefighters, breathing tanks etc.

E. Pultrusion:

It is a continuous process in which composites in the form of fibers and fabrics are pulled through a bath of liquid resin. Then the fibres wetted with resin are pulled through a heated die. The die plays important roles like completing the impregnation and controlling the resin. Further, the material is cured to its final shape. The die shape used in this process is nothing the replica of the final product. Finally, the finished product is cut to length.

In this process, the fabrics may also be introduced into the die. The fabrics provide a fibre direction other 0. Further, a variant of this method to produce a profile with some variation in the cross-section is available. This is known as pulforming.

The resins like epoxy, polyester, vinylester and phenolic can be used with any fibre.

The pultrusion process is shown in Figure 1.18.

Advantages:
o o o o o o The process is suitable for mass production. The process is fast and economic. Resin content can be accurately controlled. Fibre cost is minimized as it can be taken directly from a creel. The surface finish of the product is good. Structural properties of product can be very good as the profiles have very straight fibres.

Disadvantages:
o o o Limited to constant or near constant cross-section components. Heated die costs can be high. Products with small cross-sections alone can be fabricated.

Applications:
Beams and girders used in roof structures, bridges, ladders, frameworks

Figure 1.18: Pultrusion

Introduction In this lecture we will see some more composites fabrication processes. Further, as we have done in the previous lecture, we will see the advantages, disadvantages and application of these processes.

A. Braiding:
This is an automatic fabrication process. The toes are interlaced together to the final form of the product. Further, this interlacing can be over the mandrel which has the final shape of the product. The toes can be impregnated with the resin. Then the product is cured at room temperature or in autoclave.

Advantages:
o o o Cost effective automated technique for interlacing fibers into complex shapes. Final product is obtained. Structural properties obtained are good.

Disadvantages:
o The initial tooling cost is high.

Applications:
Rope, tubes, narrow flat strips, contoured shapes, solid 3-D shape including I-beams and Tbeams.

B. Vacuum Bagging:
This is basically an extension of the wet lay-up process described above where pressure is applied to the laminate once laid-up in order to improve its consolidation. This is achieved by sealing a plastic film over the wet laid-up laminate and onto the tool. The air under the bag is extracted by a vacuum pump and thus up to one atmosphere of pressure can be applied to the laminate to consolidate it.

Materials Options:
o o o Resins: Primarily epoxy and phenolic. Polyesters and vinylesters may have problems due to excessive extraction of styrene from the resin by the vacuum pump. Fibres: The consolidation pressures mean that a variety of heavy fabrics can be wet-out. Cores: Any.

Advantages:
o o o o Higher fibre content laminates can usually be achieved than with standard wet lay-up techniques. Lower void contents are achieved than with wet lay-up. Better fibre wet-out due to pressure and resin flow throughout structural fibres, with excess into bagging materials. Health and safety: The vacuum bag reduces the amount of volatiles emitted during cure.

Disadvantages:
o o o The extra process adds cost both in labour and in disposable bagging materials. A higher level of skill is required by the operators. Mixing and control of resin content still largely determined by operator skill.

Applications:
Large one-off cruising boats, race car components, core-bonding in production boats.

Figure 1.19: Vacuum Bagging

C. Resin Transfer Molding - RTM


The process consists of arranging the fibres or cloth fabrics in the desired configuration in a preform. These fabrics are sometimes pre-pressed to the mould shape, and held together by a binder. A second matching mould tool is then clamped over the first. Then pressurized resin is injected into the cavity. Vacuum can also be applied to the mould cavity to assist resin in being drawn into the fabrics. This is known as Vacuum Assisted Resin Transfer Moulding (VARTM) or Vacuum Assisted Resin Injection (VARI). The laminate is then cured. Both injection and cure can take place at either ambient or elevated temperature.

In this process, the resins like epoxy, polyester, vinylester and phenolic can be used. Further, one use the high temperature resins such as bismaleimides can be used at elevated process temperatures. The fibres of any type can be used. The stitched materials work well in this process since the gaps allow rapid resin transport. Some specially developed fabrics can assist with resin flow.

Advantages:
o o o o o o o o The process is very efficient. Suitable for complex shapes. High fibre volume laminates can be obtained with very low void contents. Good health and safety, and environmental control due to enclosure of resin. Possible labour reductions. Both sides of the component have a moulded surface. Hence, the final product gets a superior surface finish Better reproducibility. Relatively low clamping pressure and ability to induce inserts.

Disadvantages:
Matched tooling is expensive and heavy in order to withstand pressures. Generally limited to smaller components. Unimpregnated areas can occur resulting in very expensive scrap parts.

Applications:
The applications include the hollow cylindrical parts like motor casing, engine covers, etc.

Figure 1.21: Centrifugal casting

D. Centrifugal Casting:
In this process the chopped fibres and the resin is sent under pressure to the cylindrical moulding. The moulding is rotating. Due to centrifugal action, the mixture of resin and chopped fibres get deposited on wall of the moulding. Thus, the mixture gets the final form of the product.

Advantages:
1. Suitable for small hollow cylindrical products. 2. Economic for small production.

Disadvantages:
3. Complex shape can not be made. 4. Resin with low viscosity is needed. 5. The finish of the inner side of the product is not good.

6. The structural properties may not be good as the chopped fibres are used.

Figure 1.21: Centrifugal casting

Applications:
The applications include the hollow cylindrical parts like motor casing, engine covers, etc.

In this lecture, we are going to introduce some concepts from solid mechanics which will be useful for better understanding of this course. It is presumed that the readers have some basic knowledge of linear algebra and solid mechanics. In solid mechanics, each phase of a material is considered to be continuum, that is, there is no discontinuity in the material. Thus, in this course individual fibres and the matrix of a lamina/composite are considered to be continuum. Further, this results in saying that heterogeneous composite is also a continuum.

In this lecture, we will introduce some of the notations that will be followed for the rest of the course. Hence, the readers are advised to understand them clearly before they proceed to further lectures.

Concept of Tensors
Tensors are physical entities whose components are the coefficients of a linear relationship between vectors. The list of some of the tensors used in this course is given in Table 2.1.

Table 2.1 List of some of the tensor quantities

Quantity
Scalar (zeroth order tensor)

Live subscripts 0 1 2 4

vi ij,ij Cijkl

Vector (first order tensor) Second order tensor Fourth order tensor

It is often needed to transform a tensorial quantity from one coordinate system to another coordinate system. This transformation of a tensor is done using direction cosines of the angle measured from initial coordinate system to final coordinate system. Let us use axes coordinate axes and as the final coordinate axes (denoted here by symbol prime as the initial ). Now, we

need to find the direction cosines (denoted here by aij) for this transformation relation. Let us use the convention for direction cosines that the first subscript (that is, i) of aij corresponds to the initial axes and the second subscript (that is, j) corresponds to final axes. The direction cosine correspondence with this convention in 3D Cartesian coordinate system is given in Table 2.2. The corresponding Cartesian coordinate systems are shown in Figure 2.1.

Table 2.2 Direction cosines for 3D Cartesian coordinate system

From/To

Figure 2.1 Rectangular or Cartesian coordinate systems

Let us derive the direction cosines for a transformation in a plane. Let the coordinate axes x1-x2 (that is, plane 1-2) are rotated about the third axis x3 by an angle figure it is easy to see that between cosines as shown in Figure 2.2. Thus, from the

. A careful observation of the figure shows that the angle . It means that the direction

is not the same as the angle between .

Figure 2.2 Transformation about x3 axis Now, we will find all the direction cosines. The list is given below.

The above can be written in a matrix form as

(2.1)

The matrix of direction cosines given above in Eq. (2.1) is also written using short forms for . Then Equation (2.1) becomes

(2.2)

Note: Some of the books and research articles also use

Note: This matrix is also called Rotation Matrix. Note: The above direction cosine matrix can be obtained from the relation between unrotated and rotated coordinates. For the transformation shown in Figure 2.2 (a) one can write this relation using the geometrical relations shown in Figure 2.2 (b) as

Now the direction cosines are given by the following relation:

Now we will use the direction cosines to transform a vector, a second order tensor and a fourth order tensor from initial coordinate (unprimed) system to a vector, a second order tensor and a fourth order tensor in final coordinate (primed) system. First, let us do it for a vector. Let denote the components of a vector P in unprimed and

primed coordinate axes. Then the components of this vector in rotated coordinate system are given in terms of components in unrotated coordinate system and corresponding direction cosines as (2.3) Now, putting the direction cosines in terms of angles and summing over the repeated index j (=1, 2, 3) in Equation (2.3) we get

(2.4)

med and primed coordinate systems are as shown in Figure 2.2. The transformation matrix for this rotation is given in can be given as

mponents

Note: In two dimensional case, the above transformation is written as

Equation (2.3) can also be written in an inverted form to give the components Pi in unrotated axes in terms of components in rotated axes system as

(2.5)

The rotation matrix aij in Equation (2.2) has a property that

(2.6)

Now, we will extend the concept to transform a second order tensor. Let us transform the stress tensor as follows

The transformation of a fourth order tensor

is given as (2.8)

(2.7)

The readers are suggested to write the final form of Equation (2.8) using similar procedure used to get the last of Equation (2.7).

Deformation of a Body
When a deformable body is subjected to external forces, a body may translate, rotate and deform as well. Thus, after deformation the body occupies a new region. The initial region occupied by the body is called Reference Configuration and the new region occupied by the body after translation, rotation and deformation is called Deformed Configuration. Let us consider a point P in reference configuration. Its position with respect to origin of a reference axes system The point P occupies a new position and its position vector is also given. is shown in Figure 2.3.

Figure 2.3 Reference and deformed configurations The deformation map is defined as (2.9)

Thus, deformation map is a vector valued function. Similarly, for deformation of a point Q to can write (2.10)

, we

We can find the deformation

as (2.11)

where

is called Deformation Gradient. In component form, one can write (2.12)

Now, let us give the deformation map for the displacement of a point. Let us consider the point P in reference configuration again. It goes a deformation can write this deformation as follows (2.13) This gives us the deformation gradient as and occupies a new position . Thus, we

(2.14)

or in component form (2.15)

Now, we will define strain tensor. We are going to find . Thus,

. We know that

(2.16)

where E is Lagrangian Strain Tensor. Now using the last two of Equation (2.16) for we get,

(2.17)

This equation can be written in index form as

(2.18)

where

is given as

. Thus, the strain components are nonlinear in

. Here,

are the displacement components in three directions. For example, let us write the expanded form of strain components .

(2.19)

Similarly, (2.20)

The readers should observe that from the definition of strain tensor in Equation (2.18), the strain tensor is symmetric (that is, ). If the gradients of the displacements are very small the product terms in

Equation (2.18) can be neglected. Then, the resulting strain tensor (called Infinitesimal Strain Tensor) is given as

(2.21)

The individual strain components are given as

(2.22)

The readers are very well versed with these definitions. This strain tensor can be written in matrix form as

(2.23)

Note: The shear strain components mentioned above are tensorial components. In actual practice, engineering shear strains (which are measured from laboratory tests) are used. These are denoted by . The relation between tensorial and engineering shear strain components is (2.24)

The engineering shear strain components are given as follows:

(2.25)

Using the engineering shear strain components, the strain tensor can be written in matrix form as

(2.26)

Stress

Now, we will introduce the concept of stress. The components of stress at a point (also called State of Stress) are (in the limit) the forces per unit area which are acting on three mutually perpendicular planes passing through this point. This is represented in Figure 2.4. Stress tensor is a second order tensor and denoted as . In this notation, the first subscript

corresponds to the direction of the normal to the plane and the second subscript corresponds to the direction of the stress. For example, denotes the stress component acting on a plane

which is perpendicular to direction 2 and stress is acting in direction 3. The tensile normal stress components are positive. The shear stress components are defined to be positive when the normal to the plane and the direction of the stress component are either both positive or both negative. The readers should note that the state of stress shown in Figure 2.4 represents all stress components in positive sense. In this figure, the stress components are shown on positive faces only.

Figure 2.4 State of stress at a point


The stress tensor can be written in matrix form as follows: (2.27) In general, instead of using global 1-2-3 coordinate system, xy-z global coordinate system is used. Further, the shear stress components are shown using notation tensor in this case can be written as . Thus, the stress

(2.28)

Note: The stress tensor will be symmetric, that is only when the there are no distributed moments in the body. The readers are suggested to read more on this from any standard solid mechanics book. In this entire course, we will deal with symmetric stress-tensor.

Equilibrium Equations
The equilibrium equations for a body to be in static equilibrium at a point are given in index notations as (2.29)

where,

are the body forces per unit volume. If the body forces are absent, then the equilibrium

equation becomes (2.30) The equilibrium equations without body forces are written using xyz coordinates as follows:

(2.31)

Boundary Conditions
The boundary conditions are very essential to solve any problem in solid mechanics. The boundary conditions are specified on the surface of the body in terms of components of displacement or stress. However, the combination of displacement and stress components is also specified. Figure 2.5 shows a body, where the displacement as well as stress components are used to specify the boundary conditions.

We define traction vector for any arbitrary point (for example, point P in Figure 2.5) on surface as a vector consisting of three stress components acting on the surface at same point. Here, the three stress components are normal stress written as and shear stress and . The traction vector at this point is (2.32) where 2, then follows is the unit normal to the surface at point P. For example, if this surface is perpendicular to axis and the components of traction acting at a point on this surface are given as

(2.33)

Figure 2.5: (a) A body showing displacement and traction boundary conditions (b) traction vector at any arbitrary point P on the surface of a body

: Constitutive Equations
The relationship between stress and strain is known as constitutive equation. The general form of this equation is (2.34)

Here,

are called elastic constants. This is also referred to as elastic moduli or elastic

stiffnesses. This form of constitutive equation is known as generalized Hookes law. Very soon, we will see this equation in detail for various material types. The inverse of this equation can be written as

(2.35)

where

is known as compliance.

Plane Stress Problem


Plane stress problem corresponds to a situation where out of plane stress components are negligibly small. Thus, we can say that the state of stress is planar. The planar state of stress in x-y plane is shown in Figure 2.6. For the case shown in this figure, the normal and shear stress components in z directions, that is assumes the stress symmetry. are zero. Please note that the state of stress shown in this figure

Note: A careful observation for strain components in z direction (

) reveals that these

need not be zero. This is a common mistake made by many readers. The magnitude of these strain components can be found with the help of constitutive equation given in Equation (2.34).

Figure 2.6: Plane stress problem For plane stress problem the equilibrium equations take the following form

(2.36)

Plane Strain Problem


Plane strain problem corresponds to a condition where all the out of plane strain components are

negligibly small. Here, we denote as out of plane strain components. The readers are again cautioned to note that the out of plane stress components need not be zero. These depend upon the constitutive equation. Further, the equilibrium equation is same as Equation (2.36) and .

Homework
1. Verify the property given in Equation (2.6) for rotation matrix. 2. Using Equation (2.6), show that

3. where the term

, called Kronecker delta, has the value 1 on the diagonal and 0 on the off

diagonal, that is, it represents an identity matrix when represented in matrix form. 4. Using relation for strain components (given in Equation (2.21)) write the expanded form of all strain components and understand the physical significance of all strain components. (The normal strain components denote the stretching of a line element, etc.) 5. Derive the principles of minimum of total potential and total complementary potential energy. 6. Derive the principle of virtual work.

Introduction
In this lecture, we are going to develop the 3D constitutive equations. We will start with the the generalized Hookes law for a material, that is, material is generally anisotropic in nature. Finally, we will derive the constitutive equation for isotropic material, with which the readers are very familiar. The journey for constitutive equation from anisotropic to isotropic material is very interesting and will use most of the concepts that we have learnt in Lecture 5. The generalized Hookes law for a material is given as

(3.1)

where, ij is a second order tensor known as stress tensor and its individual elements are the stress

components.

is another second order tensor known as strain tensor and its individual elements

are the strain components. Cijkl is a fourth order tensor known as stiffness tensor. In the remaining section we will call it as stiffness matrix, as popularly known. The individual elements of this tensor are the stiffness coefficients for this linear stress-strain relationship. Thus, stress and strain tensor has ( ) 9 components each and the stiffness tensor has (=) 81 independent

elements. The individual elements =) 81 are referred by various names as elastic constants, moduli and stiffness coefficients. The reduction in the number of these elastic constants can be sought with the following symmetries.

Stress Symmetry:

The stress components are symmetric under this symmetry condition, that is, are six independent stress components. Hence, from Equation. (3.1) we write

. Thus, there

(3.2)

Subtracting Equation (3.2) from Equation (3.1) leads to the following equation

(3.3)

There are six independent ways to express i and j taken together and still nine independent ways to express k and l taken together. Thus, with stress symmetry the number of independent elastic constants reduce to ( ) 54 from 81.

Strain Symmetry:

The strain components are symmetric under this symmetry condition, that is, Equation (3.1) we write

. Hence, from

Subtracting Equation (3.3) from Equation (3.2) we get the following equation

(3.4)

It can be seen from Equation (3.4) that there are six independent ways of expressing i and j taken together when k and l are fixed. Similarly, there are six independent ways of expressing k and l taken together when i and j are fixed in Equation (3.4). Thus, there are for this linear elastic material with stress and strain symmetry. independent constants

With this reduced stress and strain components and reduced number of stiffness coefficients, we can write Hookes law in a contracted form as

(3.5)

where

(3.6)

Note: The shear strains are the engineering shear strains.

In order for Equation (3.5) to be solvable for strains in terms of stresses, the determinant of the stiffness matrix must be nonzero, that is .

The number of independent elastic constants can be reduced further, if there exists strain energy density function W, given as below.

Strain Energy Density Function (W ):


The strain energy density function W is given as

(3.7)

with the property that

(3.8)

It is seen that W is a quadratic function of strain. A material with the existence of W with property in Equation (3.8) is called as Hyperelastic Material.

The W can also be written as

(3.9)

Subtracting Equation (3.9) from Equation (3.7) we get

(3.10)

which leads to the identity

. Thus, the stiffness matrix

is symmetric. This symmetric matrix has 21 independent elastic constants. The stiffness matrix is given as follows:

(3.11)

The existence of the function W is based upon the first and

second law of thermodynamics. Further, it should be noted that this function is positive definite. Also, the function W is an invariant (An invariant is a quantity which is independent of change of reference).

The material with 21 independent elastic constants is called Anisotropic or Aelotropic Material. Further reduction in the number of independent elastic constants can be obtained with the use of planes of material symmetry as follows.

Material Symmetry:
It should be recalled that both the stress and strain tensor follow the transformation rule and so does the stiffness tensor. The transformation rule for these quantities (as given in Equation (3.1)) is known as follows

(3.12)

where

are the direction cosines from i to j coordinate

system. The prime indicates the quantity in new coordinate system.

When the function W given in Equation (3.9) is expanded using the contracted notations for strains and elastic constants given in Equation (3.11) W has the following form:

(3.1 3)

Thus, from Equation (3.13) it can be said that the function W has the following form in terms of strain components:

(3.14)

With these concepts we proceed to consider the planes of material symmetry. The planes of the material, also called elastic symmetry are due to the symmetry of the structure of anisotropic body. In the following, we consider some special cases of material symmetry.

(A) Symmetry with respect to a Plane:


Let us assume that the anisotropic material has only one plane of material symmetry. A material with one plane of material symmetry is called Monoclinic Material. Let us consider the x1- x2 ( x3= 0) plane as the plane of material symmetry. This is shown in Figure 3.1. This symmetry can be formulated with the change of axes as follows (3.15)

With this change of axes,

(3.16)

This gives us along with the use of the second of Equation (3.12)

(3.17)

Figure 3.1: Material symmetry about x1-x2 plane First Approach: Invariance Approach

Now, the function W can be expressed in terms of the strain components it must be of the form

. If W is to be invariant, then

(3.18)

Comparing this with Equation (3.13) it is easy to conclude that (3.19) Thus, for the monoclinic materials the number of independent constants are 13. With this reduction of number of independent elastic constants the stiffness matrix is given as

Second Approach: Stress Strain Equivalence Approach


The same reduction of number of elastic constants can be derived from the stress strain equivalence approach. From Equation (3.12) and Equation (3.16) we have

(3.21)

The same can be seen from the stresses on a cube inside such a body with the coordinate systems shown in Figure 3.1. Figure 3.2 (a) shows the stresses on a cube with the coordinate system x1, x2,x3 and Figure 3.2 (b) shows stresses on the same cube with the coordinate system Comparing the stresses we get the relation as in Equation (3.21). Now using the stiffness matrix as given in Equation (3.11), strain term relations as given in Equation (3.17) and comparing the stress terms in Equation (3.21) as follows: .

Using the relations from Equation (3.17), the above equations reduce to

Noting that

, this holds true only when

Similarly,

This gives us the

matrix as in Equation (3.20).

Figure 3.2: State of stress (a) in x1, x2, x3 system (b) with x1-x3 plane of symmetry

First Approach: Invariance Approach

We can get the function W simply by substituting

in place of

and using contracted notations for the

strains in Equation (3.18). Noting that W is invariant, its form in Equation (3.18) must now be restricted to functional form

(3.25)

From this it is easy to see that

Thus, the number of independent constants reduces to 9. The resulting stiffness matrix is given as

(3.26)

When a material has (any) two orthogonal planes as planes of material symmetry then that material is known as Orthotropic Material. It is easy to see that when two orthogonal planes are planes of material symmetry, the third mutually orthogonal plane is also plane of material symmetry and Equation (3.26) holds true for this case also.

Note: Unidirectional fibrous composites are an example of orthotropic materials.

Second Approach: Stress Strain Equivalence Approach


The same reduction of number of elastic constants can be derived from the stress strain equivalence approach. From the first of Equation (3.12) and Equation (3.23) we have

(3.27)

The same can be seen from the stresses on a cube inside such a body with the coordinate systems shown in Figure 3.3. Figure 3.4 (a) shows the stresses on a cube with the coordinate system x1, x2, x3 and Figure 3.4 (b) shows stresses on the same cube with the coordinate system stresses we get the relation as in Equation (3.27). Now using the stiffness matrix given in Equation (3.20) and comparing the stress equivalence of Equation (3.27) we get the following: . Comparing the

This holds true when

. Similarly,

This gives us the

matrix as in Equation (3.26).

Figure 3.4: State of stress (a) in x1, x2, x3 system (b) with x1-x2 andx2-x3 planes of symmetry Alternately, if we consider x1-x3 as the second plane of material symmetry along with x1-x2 as shown in Figure 3.5, then

(3.28)

And

(3.29)

This gives us the required strain relations as (from Equation (3.12))

or in contracted notations, we write

Substituting these in Equation (3.18) the function W reduces again to the form given in Equation (3.25) for W to be invariant. Finally, we get the reduced stiffness matrix as given in Equation (3.26).

Figure 3.5: Material symmetry about x1-x2 andx1-x3 planes

:
The stress transformations for this coordinate transformations are (from the first of Equation (3.12) and Equation (3.29))

The same can be seen from the stresses shown on the same cube in x1, x2, x3 and coordinate systems in Figure 3.6 (a) and (b), respectively. The comparison of the stress terms leads to the stiffness matrix as given in Equation (3.26). Note: It is clear that if any two orthogonal planes are planes of material symmetry the third mutually orthogonal plane has to be plane to material symmetry. We have got the same stiffness matrix when we considered two sets of orthogonal planes. Further, if we proceed in this way considering three mutually orthogonal planes of symmetry then it is not difficult to see that the stiffness matrix remains the same as in Equation (3.26).

Homework:
1. Starting with hyperelastic material, first take x2-x3 plane as plane of material symmetry and obtain the stiffness matrix. Is this matrix the same as in Equation (3.20) ? Justify your answer. 2. Starting with the stiffness matrix obtained in the above problem, take x1-x3as an additional plane of symmetry and obtain the stiffness matrix. Is this matrix the same as in Equation (3.26)? Justify your answer.

Transverse Isotropy: Introduction:


In this lecture, we are going to see some more simplifications of constitutive equation and develop the relation for isotropic materials.

First we will see the development of transverse isotropy and then we will reduce from it to isotropy.

First Approach: Invariance Approach


This is obtained from an orthotropic material. Here, we develop the constitutive relation for a material with transverse isotropy in x2-x3 plane (this is used in lamina/laminae/laminate modeling). This is obtained with the following form of the change of axes.

(3.30)

Now, we have

Figure 3.6: State of stress (a) in x1, x2, x3 system (b) with x1-x2 and x1-x3 planes of symmetry

From this, the strains in transformed coordinate system are given as:

(3.31)

Here, it is to be noted that the shear strains are the tensorial shear strain terms. For any angle ,

(3.32)

and therefore, W must reduce to the form

(3.33)

Then, for W to be invariant we must have

Now, let us write the left hand side of the above equation using the

matrix as given in Equation (3.26)

and engineering shear strains. In the following we do some rearrangement as

Similarly, we can write the right hand side of previous equation using rotated strain components. Now, for W to be invariant it must be of the form as in Equation (3.33).

1. If we observe the terms containing is unchanged.

and

in the first bracket, then we conclude that

2. Now compare the terms in the second bracket. If we have (3.32) is satisfied. 3. Now compare the third bracket. If we have satisfied.

then the first of Equation

, then the third of Equation (3.32) is

4. Now for the fourth bracket we do the following manipulations. Let us assume that is unchanged. Then we write the terms in fourth bracket as

and

To have W to be invariant we need to have satisfied.

so that the third of Equation (3.32) is

Thus, for transversely isotropic material (in plane x2-x3) the stiffness matrix becomes

(3.34)

Thus, there are only 5 independent elastic constants for a transversely isotropic material.

Second Approach: Comparison of Constants


This can also be verified from the elastic constants expressed in terms of engineering constants like . Recall the constitutive equation for orthotropic material expressed in terms of engineering constants. For the transversely isotropic materials the following relations hold.

When these relations are used in the constitutive equation for orthotropic material expressed in terms of engineering constants, the stiffness matrix relations in Equation (3.34) are verified.

Isotropic Bodies
If the function W remains unaltered in form under all possible changes to other rectangular Cartesian systems of axes, the body is said to be Isotropic. In this case, W is a function of the strain invariants. Alternatively, from the previous section, W must be unaltered in form under the transformations

(3.35)

and

(3.36)

In other words, W when expressed in terms of replacing by

must be obtained from Equation (3.33) simply by

. By analogy with the previous section it is seen that for this to be true under the

transformation Equation (3.35). We can write

And the transformed strains are given as

(3.37)

Thus, for any angle ,

(3.38)

and therefore, W must reduce to the form

(3.39)

Then, for W to be invariant we must have

Now, let us write the left hand side of above equation using the

matrix as given in Equation (3.34)

and engineering shear strains. In the following we do some rearrangement as

(3.40)

Similarly, we can write the right hand side of the previous equation using rotated strain components. Now, for W to be invariant it must be of the form as in Equation (3.39)

1. From the second bracket, if we propose (3.38).

, then we can satisfy the first of Equation

2. From the third bracket, third of Equation (3.38) holds true when

. 3. The fourth bracket is manipulated as follows:

Thus, to satisfy the second of Equation (3.38) we must have have . From our observation in 2, we can write

. Further, we should

It follows automatically that W is unaltered in form under the transformation in Equation (3.36).

Thus, the stiffness matrix for isotropic material becomes as

(3.41)

Thus, for an isotropic material there are only two independent elastic constants. It can be verified that W is unaltered in form under all possible changes to other rectangular coordinate systems, that is, it is the same function of as it is of when is changed to

Homework:
1. Starting with the stiffness matrix for transverse isotropic material, take the transformations about x1 and x2 and show that you get the stiffness matrix as given in Equation (3.41).

The Lecture Contains: Constitutive Relations for Orthotropic Materials and Stress-Strain Transformations

In the previous lecture we have seen the constitutive equations for various types of (that is, nature of) materials. There are 81 independent elastic constants for generally anisotropic material and two for an isotropic material. Let us summarize the reduction of elastic constants from generally anisotropic to isotropic material. 1. For a generally anisotropic material there are 81 independent elastic constants. 2. With additional stress symmetry the number of independent elastic constants reduces to 54. 3. Further, with strain symmetry this number reduces to 36.

4. A hyperelastic material with stress and strain symmetry has 21 independent elastic constants. The material with 21 independent elastic constants is also called as anisotropic or aelotropic material. 5. Further reduction with one plane of material symmetry gives 13 independent elastic constants. These materials are known as monoclinic materials. 6. Additional orthogonal plane of symmetry reduces the number of independent elastic constants to 9. These materials are known as orthotropic materials. Further, if a material has two orthogonal planes of symmetry then it is also symmetric about third mutually perpendicular plane. A unidirectional lamina is orthotropic in nature. 7. For a transversely isotropic material there are 5 independent elastic constants. Plane 2-3 is transversely isotropic for the lamina shown in Figure 3.7. 8. For an isotropic material there are only 2 independent elastic constants.

Principal Material Directions:


The interest of this course is unidirectional lamina or laminae and laminate made from stacking of these unidirectional laminae. Hence, we will introduce the principal material directions for a unidirectional fibrous lamina. These are denoted by 1-2-3 directions. The direction 1 is along the fibre. The directions 2 and 3 are perpendicular to the direction 1 and mutually perpendicular to each other. The direction 3 is along the thickness of lamina. The principal directions for a unidirectional lamina are shown in Figure 3.7.

Engineering Constants:
The elastic constants which form the stiffness matrix are not directly measured from laboratory tests on a material. One can measure engineering constants like Youngs modulus, shear modulus and Poissons ratio from laboratory tests. The relationship between engineering constants and elastic constants of stiffness matrix is also not straight forward. This relationship can be developed with the help of relationship between engineering constants and compliance matrix coefficients. In order to establish the relationship between engineering constants and the compliance coefficients, we consider an orthotropic material in the principal material directions. If this orthotropic material is subjected to a 3D state of stress, the resulting strains can be expressed in terms of these stress components and engineering constants as follows:

Figure 3.7: Unidirectional lamina with principal material directions

(3.42)

whereas the engineering shear strain components are given as

(3.43)

Here,

are the Youngs moduli in 1, 2 and 3 directions, respectively. Thus,

represents

the axial modulus and

represent in-plane transverse and out-of-plane transverse moduli,

respectively. Note that axial direction is along the fibre direction.

represents the shear moduli. G12,G13 are the axial shear moduli in two orthogonal planes that contain the fibers.G23 represents out-of-plane transverse shear modulus. Further, it should be noted that .

The term

represents the Poissons ratio. It is defined as follows

(3.44)

where

represents the strain in the direction of applied stress and

represents the strain the

associated lateral direction. It should be noted that, in general . We will mimic some (thought) experiments that we actually do in laboratory to extract these engineering constants. For example, we find engineering constants of a transversely isotropic lamina .

Experiment 1: The lamina is loaded in traction along the axial direction as shown in Figure 3.8 (a) and
the strains in along three principal directions are recorded as the load is varied. The slope of the axial stress versus axial strain curve yields the axial Youngs modulus the Poissons ratios respectively.

. The ratios

give

Experiment 2: The lamina is loaded in traction along direction 2. The two views of this loading case are
shown in Figure 3.8 (b). The slope of stress-strain curve in direction 2 gives the in-plane transverse Youngs modulus . Since, the material is isotropic in 2-3 plane, is also equal . The strains in all

three directions are measured. The ratios respectively.

give the Poissons ratios

Experiment 3: The lamina is loaded in shear in plane 1-2 as shown in Figure 3.8 (c). The slope of the inplane shear stress and engineering shear strain curve gives the shear modulus . Please note that if

we load the lamina in 1-3 plane by shear then also we will get this modulus because the behaviour of material in shear in these two planes is identical. Thus, by shear loading in plane 1-2 gives .

Experiment 4: The lamina is loaded in shear in 2-3 plane as shown in Figure 3.8(d). The corresponding
shear stress and engineering shear strain curve yields the shear modulus .

Note: We will see the experimental details to measure some of these engineering constants in a chapter
on experimental characterization of lamina, laminates, fibres and matrix materials.

Constitutive Equation for an Orthotropic Material:


Now, let us assume that we have measured all the engineering constants of an orthotropic material along principal directions. With these engineering constants we know the relation between the strain

and stress components as given in Equation (3.42) and Equation (3.43). Thus, it is easy to see that we can relate the strain components to stress components through compliance matrix. Let us recall from previous lecture the stiffness matrix for orthotropic material (Equation (3.26)). The inverse of this matrix (compliance) will have the same form as the stiffness matrix. Thus, we write the relationship between strain and stress components using compliance matrix as follows

Figure 3.8: Experiments to extract engineering constants for a transversely isotropic material

(3.45)

Now compare Equation (3.42) and Equation (3.43) with Equation (3.45). This gives us the compliance coefficients in terms of engineering constants. The coefficients are given in Equation (3.46).

(3.46)

It should be noted that like stiffness matrix, the compliance matrix is also symmetric. The compliance matrix given in Equation (3.45) is shown symmetric. Note: It is known from our elementary knowledge of linear algebra that inverse of a symmetric matrix is also a symmetric matrix. Since, the stiffness matrix, which is the inverse of compliance matrix, is symmetric; the compliance matrix has to be symmetric. Now, let us derive some more useful relations using the symmetry of compliance matrix. If we compare

and

we get

. Similarly, comparison of

and

and comparison of

and

give two more similar relations. All these relations are given in Equation (3.47).

(3.47)

or one can write this relation in index form as

(3.48)

Constraints on Engineering Constants in Orthotropic Materials


For orthotropic materials there are constrains on engineering constants. These constrains arise due to thermodynamic admissibility. For example, in case of isotropic materials it is well known that the Youngs modulus and shear modulus are always positive. Further, the Poissons ratio lager than half are not thermodynamically admissible. If these constrains are violated then it is possible to have a nonpositive strain energy for certain load conditions. However, for isotropic materials the strain energy must be a positive definite quantity.

In this section, based on the work done by Lempriere we are going to assess the implications of this thermodynamic requirement (positive definiteness of strain energy) for orthotropic materials. The sum of work done by all stress components must be positive, otherwise energy will be created. This condition imposes a thermodynamic constraint on elastic constants. This condition requires that both compliance and stiffness matrices must be positive definite. In other words, the invariants of these matrices should be positive.

Let us look at this condition with physical arguments. For example, consider that only one normal stress component is applied. Then we can find the corresponding strain component from the corresponding diagonal entry of the compliance matrix. Thus, we can say that for the strain energy to be positive definite the diagonal entries of the compliance matrix must be positive. Thus, (3.53) In a similar way, it is possible under certain conditions to have a deformation which will give rise to only one normal strain component. We can find the corresponding stress using the corresponding diagonal entry in stiffness matrix. For the strain energy produced by this stress component to be positive the diagonal entry of the stiffness matrix must be positive. Thus, this condition reduces to

(3.54)

and the determinant of the compliance matrix must also be positive. That is, (3.55) Now, using the reciprocal relations given in Equation (3.49), the condition in Equation (3.54) can be expressed as

(3.56)

This condition also justifies that the Poissons ratio greater than unity is feasible for orthotropic lamina. Poissons ratios greater than unity is sometimes observed in experiments.

The condition in Equation (3.55) can be written as

The terms inside the brackets are positive. Thus, we can write

(3.57)

This condition shows that all three Poissons ratios cannot have large positive values and that their product must be less than an half. However, if one of them is negative no restriction is applied to remaining two ratios. Let us consider the transversely isotropy as a special case. Let us consider transverse isotropy in 2-3 plane. Let

(3.58)

Then, the conditions in Equation (3.56) reduce as

(3.59)

and Equation (3.57) (using reciprocal relations in Eq. (3.49)) becomes as

(3.60)

The condition posed by above equation is more stringent than that posed in Equation (3.59). Note that the

quantities

and

are both positive. Thus, the limits on Poissons ratio in transverse plane are

(3.61)

Further, consider a special case of isotropic material where (3.61) to a well known condition

and

. This simplifies Equation

(3.62)

Stress and Strain Transformation about an Axis

Often it is required to transform the stress or strain tensor from one coordinate axes system to another. For example, if the fibres in a lamina are not oriented along direction x, then we may need to transform the stress and strain components from principal material directions 1-2-3 to global directions xyz or vice-a-versa. It should be noted that the stress and strain tensors are second order tensors. Hence, they follow tensor transformation rules.

In this section we are going to introduce two notations. The subscripts 123 will denote a quantity (like constitutive equation, engineering constants, etc.) in principal material directions, while subscripts xyz will denote the corresponding quantity in global coordinate directions.

Figure 3.9: Unidirectional lamina with global xyz directions and principal material 1-2-3 directions Let us transform the stress and strain components for the case

shown in Figure 3.9. Here, xy plane is rotated about direction z to 12 plane. Here, direction z and direction 3 are in same directions, that is, along the thickness direction of lamina. The direction cosines for this axes transformation are as given in Equation (2.2). However, these are again given below.

Stress Transformation:

Let us do the stress transformation as given in Equation (2.7). In this equation the primed stress components denotes the component in 123 coordinate system. Using the expanded form of Equation (2.7) and stress symmetry, let us obtain component of stress

Thus, substituting the values of direction cosines from above, we get

The remaining five stress terms (using stress symmetry) on the left hand side are also obtained in a similar way. Let us write the final form of the relation as (3.63) Here, and and is the stress

transformation matrix. Thus, comparing all the terms as in Equation (3.63), we can write as

(3.64)

where not symmetric.

. It should be noted that

is

Strain Transformation:
In a similar way, we can transform the strain components from xy plane to 1-2 plane. In this transformation we will use engineering shear strains. Let us find the equation similar to stress transformation and using strain symmetry as using the transformation

(3.65)

Substituting the direction cosines and rearranging, we get

(3.66)

We know from Equation (2.24) that the tensorial shear strains are half the engineering shear strains. Thus, in Equation (3.53) we substitute

On simplification and putting

, we get

(3.67)

The other five strain terms (using strain symmetry) on the left hand side are also obtained in a similar way. Let us write the final form of the relation as

(3.68)

Here,

and

and

is the strain as

transformation matrix. Thus, comparing all the terms as in Equation (3.68), we can write

(3.69)

Note: The transformation matrices,

and

differ by factors 2 in two terms.

Note: The transformation matrices,

and

are not symmetric.

Note: The order of stress and strain components in Equatjion (3.63) and Equation (3.68) is important.
Some books and research articles follow different orders. The readers are cautioned to take a note of it.

Examples: Example 1: Calculate the stiffness and compliance coefficients for transversely isotropic material
AS4/3501 Epoxy. The properties are as given below for a fibre volume fraction of 60%.

Solution:

Unit of all compliance coefficients is 1/GPa. The corresponding stiffness coefficients are calculated by inversion of the compliance matrix.

Unit of all stiffness coefficients is GPa.

Note: Both stiffness and compliance matrices are symmetric. Homework:

1. Write the number of independent elastic constants for 3D hyperelastic, monoclinic, orthotropic,
transversely isotropic and isotropic materials.

2. Are the Poissons ratio


lamina?

and

independent of each other for an orthotropic unidirectional

3. Take the form of stiffness matrix for an orthotropic material as given in Equation (3.26). Using any
symbolic calculation software like Maple or Mathematica, obtain the inverse of this matrix and confirm that the form of compliance matrix written in Equation (3.42) is correct. Further, confirm that this matrix is symmetric. (One should be able to do this using the concepts of linear algebra alone.)

4. Extend the Problem 3 to get the stiffness matrix given in Equation (3.51).
Stiffness, Compliance Transformation and Hygro-thermo-elastic Constitutive Relation

Introduction
In the previous lecture we have derived the stiffness matrix in terms of engineering constants using the compliance relations. Further, we have seen the constraints

over the engineering constants in orthotropic materials. Then we have visited the stress and strain transformation. In the present lecture we will see the stiffness and compliance transformation about an axis. Further, we will address the effect of thermal and hygroscopic actions on lamina constitutive equations.

Stiffness Transformation:
It is required to relate the stress components with strain components in global xyz directions. The stiffness matrix which relates the stress and strain components in global directions is called as transformed stiffness matrix. We will derive an expression for the transformed stiffness matrix as follows. The constitutive equation in principal material coordinates, as given in Equation (3.11), is (3.70)

Now, we express

using Equation (3.63) to transform stresses and

Equation (3.68) to transform strains. Substituting these equations, we get (3.71)

Pre-multiplying both sides by

, we get

(3.72)

where we define the transformed stiffness matrix

as

(3.73)

The transformation matrices

and

can be inverted as follows

(3.74)

The final form of the transformed stiffness matrix is given in Equation (3.75).

(3.75)

The individual relation for

terms of this matrix are determined using . The individual terms are given in Equation (3.76).

and

Note: The transformed stiffness matrix is symmetric in nature. Note: The transformed stiffness matrix given in Equation (3.75) has exactly the
same form as a stiffness matrix for a monoclinic material. Thus, we can conclude that a transformation through an arbitrary angle monoclinic material behaviour. The same can be seen from the plane of elastic symmetry considerations in xyz coordinate system. The given lamina is symmetric only about xy plane. Thus, the transformed stiffness matrix in Equation (3.75) is consistent with monoclinic material. about direction 3, leads to a

Note: Transformed stiffness coefficient terms are fourth order in the sine and
cosine functions. It is very important to use appropriate precision level while calculating (in examinations and writing computer codes) these coefficients.

(3.76)

The constitutive equation

becomes

Compliance Transformation:
We are going to follow a procedure to transform compliance matrix similar to one used for transformation of a stiffness matrix. We have the constitutive equation in principal material direction as in Equation (3.70). We can write this in inverted form as (3.77)

using Equation (3.63) and Equation (3.68) we get

(3.78)

Pre-multiplying both sides by

, we get

(3.79)

where, we define the transformed compliance matrix

as
(3.80)

Alternately, we can find

by inverting the transformed stiffness matrix

. Thus, inverting

from Equation (3.73), we get

After carrying out the calculation for

, it is easy to give its form as follows

(3.81)

Note that

has the same symmetric form as the transformed stiffness matrix.

The individual terms of the compliance matrix are obtained by carrying out multiplication of matrices

as in Equation (3.82) and are given below.

Thermal Effects:
Thermal effects (effects due to change in temperature) are very important in composite materials for various reasons. The analysis of composites with thermal effects and effective thermal properties of the composite are two of the main reasons.

Important issues from analysis point of view:


1. The composite materials are used in environment where thermal gradients are unavoidable. For example, the helicopter containing composite fuselage operates at -50 C during winter at Leh and same helicopter can operate at +50 C during summer in the desert of Rajasthan. Thus, the effect of temperature gradient on the service performance of the composite is very important. In such service conditions, the layers of composite material tend to expand or contract but are restricted due to adjacent layers. Thus, it induces thermal stresses. 2. Most of the fabrication processes of polymer matrix composites have thermal cycles for matrix curing. A typical cycle involves raising the temperature to a certain level and holding it there for specified time and bringing it back to room temperature. It is well known that the fibre and matrix materials have different coefficients of thermal expansion (defined below). This mismatch produces residual thermal stresses because the fibres and matrix material is

constrained in a composite.

Important issues from effective thermal properties point of view:


The second reason for the study of thermal effects is the effective properties of the composite materials. 1. Finding effective thermal properties of the composite theoretically to get an estimate requires sophisticated mathematical modeling when one considers: a. Difference in coefficients of thermal expansion of fibre and matrix materials b. The direction dependence of coefficients of thermal expansion in these materials c. Curing cycle temperature variations. This point is important because for some of the materials the coefficient of thermal expansion changes with temperature. 2. Finding the effective thermal properties for lamina in global direction with oriented fibres as shown in Figure 3.9 requires a special attention. Further, finding these effective properties by laboratory test is also a challenge. Thus, for the various reasons mentioned above the study of thermal effect is very important. In the following, we develop a systematic way to handle effective thermal properties of a lamina along global directions. It is well known that when a material is subjected to thermal gradient, it undergoes a deformation. The strain due to thermal changes is called thermal strain (denoted by superscript ( T)). In general, the thermal strain is proportional to the temperature change . The constant of proportionality is

called coefficient of thermal expansion. Thus, we can write the thermal strains in principal material directions for an orthotropic material as

(3.83)

where

denote the coefficients of thermal expansion in principal

material directions. It should be noted that for an orthotropic material in principal directions there are no shear strains due to thermal effects like in an isotropic material. For an isotropic material the coefficient of thermal expansion is same in any direction. However, for an orthotropic material . The thermal expansion of an elemental cube in principal directions for an isotropic and orthotropic material is shown in Figure 3.10.

These thermal strains will not produce stresses unless these are constrained. The thermal strains which do not produce stresses are known as free thermal strains. However, in case of composites the fibres and matrix are constrained in a lamina and layers are constrained a laminate. Thus, in composite the thermal strains produce the thermal stresses.

The thermal strains are given in principal material directions as given in Equation (3.83). Let us consider that we need to find these strains in a global coordinate system (refer Figure 3.9). We need to transform them from 123 coordinate system to xyz coordinate system by a rotation 3-axis. Thus, similar to Equation (3.68), we can write about

(3.84)

Substituting Equation (3.83) in the above equation,

(3.85)

where

(3.86)

Figure 3.10: Thermal expansion in an isotropic and orthotropic material

On substitution of

in the above equation, we get the following individual terms of coefficients

thermal expansion in xyz directions.

(3.87)

Using Equation (3.87) in Equation (3.85), the engineering thermal strains in global coordinates are given as

(3.88)

Thus, from this equation it should be noted that the transformation of thermal strains in global coordinates gives normal strain components and a shear strain component in xy plane for an orthotropic material with and fiber orientations other than .

Thermo-Elastic Constitutive Equation:


Let us assume that the total strain, in a composite is a superposition of the free thermal strain . Thus,

and the strain due to mechanical loads (also known as mechanical strains)

Now, for mechanical strains we use the constitutive equation as

Thus, we can write the total thermo-elastic strain as

(3.89)

Equation (3.89) can be written as

Premultiplying the above expression by

, we get the stresses as

(3.90)

Equation (3.90) is the basic constitutive equation for thermo-elastic stress analysis. Using Equation (3.90) and similar to Equation (3.72) we can find the stresses due to thermo-elastic effects in global directions as,

(3.91)

where

Equation (3.91) is inverted to give the total strains in terms of the mechanical and free thermal strains as

(3.92)

Effect of Moisture:
The polymer matrix composite materials, during their service can absorb moisture from the environment. The effect of absorption of moisture is to degrade the various material properties of the composite. Further, this results in an expansion. It is called hygroscopic expansion. However, this expansion is again constrained as in thermal expansion. Hence, when dealing with the hygroscopic expansions, a treatment similar to thermal expansion is used.

The hygroscopic strains

are assumed to be proportional to the percentage moisture

absorbed, This percentage is measured in terms of weight of the moisture. The constant of proportionality, is the coefficient of hygroscopic expansion.

Thus, in principal coordinates the hygroscopic strains are

(3.93)

where

(3.94)

denotes the coefficients of hygroscopic expansion in principal material directions. Following a similar procedure for thermal strains, we can write strains due to hygroscopic expansion in xyz coordinates as

(3.95)

We can write .

using values of in 123 directions as

Thus, comparing Equation (3.85) and Equation (3.95), it is easy to conclude that the coefficients of hygroscopic expansion will vary similar to the coefficients of thermal expansion as a function of orientation of fibres.

Hygro-Thermo-Elastic Constitutive Equation:


This is the most general formulation for the mechanical, thermal and hygral effects on stress analysis in composites. Here, we superimpose the strains due to these three effects to give us the total strain as

(3.96)

Using constitutive equation for mechanical strains, we get

(3.97)

The stresses in the composite can be given as

(3.98) These stresses in global coordinates xyz can be written as

(3.99)

where

(3.100)

Examples: Example 2: Transform the stiffness and compliance matrix of Example 1 about axis 3 by an angle
of = 30.

Solution:

Approach 1: One can find the transformation matrices

and do the matrix

multiplication as given in Equation (3.73) for transformed stiffness matrix and then inverse this

matrix or do the matrix multiplication as given in Equation (3.80) to get the transformed compliance matrix. The use of Equation (3.73) and Equation (3.80) is suggested because remembering is not so difficult. Further, their inverse can be easily found with the help of Equation (3.74). For

Thus

Unit of all transformed stiffness coefficients is GPa.

Unit of all transformed compliance coefficients is 1/GPa.

Approach 2: You can write the expanded form for transformed stiffness and compliance
coefficients in Equation (3.76) and Equation (3.82). However, the readers are suggested to use this approach only when they are confident of remembering these terms.

Example 3: The coefficients of moisture absorption for T300/5208 composite material are
. Plot the variation these coefficients between .

Solution:
We have the expression for variation of the coefficients of moisture absorption as

where, . We plot the above variation using a computer code. The final plot is shown in Figure 3.11.

Figure 3.11: Variation of coefficients of moisture expansion with orientation of fibres Homework:

1. Verify the result given in Equation (3.74). 2. Using the invariance property of strain energy density function, show that and 3. 4. Obtain the individual terms of transformed stiffness and compliance matrices using Equation (3.73) and
Equation (3.80), respectively and verify it with Equation (3.76) and Equation (3.82), respectively.

5. Obtain the strain transformation matrix

using tensorial shear strains. Further, using this transformation matrix obtain the transformed stiffness and compliance matrix in the form similar to Equation (3.75) and Equation (3.81). Compare the new matrices and comment on the observations with justifications.

6. Calculate the stiffness and compliance coefficients for following transversely isotropic materials given in Table
3.1.

Table 3.1: Properties for unidirectional transversely isotropic lamina


T300/BSL914C Epoxy E-glass/LY556/ HT907/DY063 S-glass/MY750/ Epoxy HY917/DY063 Epoxy

Property\Material
E1 (GPa) E2=E3(GPa) G12 = G13(GPa) 12= 13 23 1 (10 /C) 2 = 3 (10 /C)
-6 -6

138 11 5.5 0.28 0.4 -1 26

53.48 17.7 5.83 0.278 0.4 8.6 26.4

45.6 16.2 5.83 0.278 0.4 8.6 26.4

6. The stiffness matrix for an orthotropic material is given as

7. Find: a. The compliance matrix b. The engineering constants .

7. Why are the thermal effects important in composite materials? Explain in detail.

8. Plot variation of

between

for T300/BSL914C Epoxy and E-glass/LY556/

HT907/DY063 Epoxy. See Table 3.1 for the required thermal properties. 9. Why are the hygral effects important in composite materials? Explain. 10. Search literature to get the coefficients of moisture absorption for at least two composite materials and plot its variation between 11. Write a computer code to read the properties of a transversely isotropic material and calculate all the terms of stiffness and compliance matrix. Verify your results with the results given in Example 1. Then use this code to get the stiffness and compliance matrices of T300/BSL914C Epoxy and S-glass/MY750/ HY917/DY063 Epoxy. 12. Add another module to the code written for above problem to calculate the transformed stiffness and compliance matrices. Plot all the coefficients between these materials and comment. . Compare the corresponding terms of

Dimensional Lamina Analysis


Introduction
In the previous chapter, we have developed 3D constitutive equations. While analyzing composites, most of the times a planar state of stress actually exists. It is noted that a typical unidirectional lamina has very small thickness compared to its planar (xy) dimensions. Thus, it is appropriate to assume a planar state of stress in a lamina. In this chapter, we are going to derive a constitutive equation for plane stress problem in unidirectional laminar composite.

Plane

Stress

for

Monoclinic

(or

Rotated

Orthotropic)

Material

3D constitutive equation for a single layer of a unidirectional composite with a fiber orientation to the global coordinate

relative

(4.1)

For a state of plane stress, we have

(4.2)

Thus, it is easy to see that the two out of plane shear strains strains using Equation (4.1) as

are zero. We can write these

(4.3)

The out of plane normal strain

is expressed using Equation (4.1) and Equation (4.2) as

(4.4)

Note that this strain component is not zero. In plane components of strain for a plane stress state can be written using Equation (4.1) as

(4.5)

From 3D constitutive equation (Equation (3.72)) for the transformed stiffness, we can write

as

(4.6)

From this equation, we can get the out of plane transverse normal strain as

(4.7)

Thus, the out of plane normal strain is expressed in terms of in-plane strain components and known stiffness coefficients.

Reduced Transformed Stiffness Matrix The equation for in-plane components of stress in terms of the transformed stiffness coefficients is

(4.8)

Substituting

from Equation (4.7) into Equation (4.8) and upon simplification, we get

(4.9)

The above equation is written in matrix form as

(4.10)

where the transformed reduced stiffness coefficients

are defined as

(4.11)

Note: Transformed reduced stiffness matrix is symmetric. Note: It is very important to note that the transformed reduced stiffness terms for plane stress,

, are not simply the corresponding terms, taken from the 3D stiffness matrix. This should be clear from the fact that the inverse of a matrix is different from that of a . This can easily be seen from Equation (4.11). The readers should easily understand that when terms are used to define a constitutive equation, then it is a reduced transformed constitutive equation. Plane Stress for Orthotropic Material Let us recall the constitutive equation for orthotropic material in principal directions.We can write the constitutive equation using compliance matrix as (Equation (3.45))

(4.12)

We have planar state of stress for orthotropic lamina. Then we have out of plane transverse stress components zero, that is, (4.13) Let us write the out of plane transverse shear strains using this information and Equation (4.12) as

(4.14)

Thus, the out of plane transverse shear strains are zero. Now, let us write the out of plane transverse normal strain using Equation (4.12) as

(4.15)

Using and Equation (4.13) in the above equation, we get

(4.16)

and the inplane strain components are given as

(4.17)

This

equation

is

called

reduced

constitutive

equation

using

compliance

matrix.

We have the 3D constitutive equation using stiffness matrix in principal material directions as (Equation . (3.26))

(4.18

We have from the condition of plane stress problem that we can write

. Thus, using Equation. (4.18),

(4.19)

This leads to non-zero transverse normal strain

as

(4.20)

Using Equation. (4.18), we can write the inplane stress components as

(4.21)

Putting the expression for

from Equation. (4.20) in above equation, we get

(4.22)

This equation is written in matrix form as

(4.23)

where, the

terms can be written using index notations as follows

(4.24)

Note: The reduced stiffness matrix is symmetric.

Note: The readers should again understand the difference between not the same.

and

terms. They are

The inversion of Equation (4.24) should give us Equation (4.17), that is,

Let us compare this equation with corresponding 3D equation (Eq. (4.12). It can easily be seen that the compliance terms of the constitutive equation are identical for 3 D and plane stress problems. Thus, we can write for the plane stress problems as

(4.25)

It is easy to invert a

matrix. In fact, you need to invert a

matrix. Thus, we can write the

individual reduced stiffness entries in terms of compliance entries as

(4.26)

Compliance

and

Stiffness

Coefficients

Using

Engineering

Constants

Let us write the compliance and stiffness matrices using engineering constants. It is easy to see that the individual entries of the compliance matrix in plane stress problem are the as 3 D compliance. Thus, we write for the plane stress problem the compliance entries as

(4.27)

Here, we have used the property that compliance matrix is symmetric, that is,

. Using

, we can develop the reciprocal relationship for 2D case as

(4.28)

Note: It is easy to see for a plane stress problem of an orthotropic material that there are only four of the five material constants independent. We can write the individual terms of reduced stiffness matrix in principal material directions by using Equation (4.26) and Equation (4.27) as

(4.29)

Note: For a transversely isotropic material there is no reduction of the number of independent constants for plane stress problem.

One can write the constitutive equation in material coordinates, using Equation (3.42), Equation (3.43) and introducing the corresponding reduction in out of plane direction as

(4.30)

It is easy to write the compliance coefficients in Equation (4.27) from these relations. Further, we can write the above relations in inverted form as

(4.31)

These relations lead to individual reduced stiffness coefficients given in Equation (4.29).

2D

Transformations

about

an

Axis:

In planar stress condition we need to transform the stresses in plane. Let us write, similar to Equation (3.63), the transformation equation for stresses as

(4.32)

where

is the transformation matrix for stress tensor. For the above equation, using Equation

(3.64), this matrix can be written as

(4.33)

Similarly, we can write the strain transformation equation in the following form.

(4.34)

where

is the transformation matrix for strain tensor. We can find this matrix using Equation (3.39)

and the above relations as

(4.35)

Note: The transformation matrices in two entries of these matrices.

and

are not symmetric. There is a difference of factor 2

Note: The transformation matrices

and

can be inverted using following relation

(4.36)

The

readers

should

verify

this

result.

Note: We have used the same matrix notation for stress and strain transformation matrices (

and

) in 3D case and plane stress case. However, the readers should note the corresponding differences.

Lamina

Constitutive

Relations

in

Global

Coordinates:

The plane stress constitutive equation in principal material coordinates is

(4.37)

Let us write the stresses and strains in terms of components in global directions using Equation (4.32) and Equation (4.34). The above equation can be re-written to give stresses in global coordinates as

(4.38)

We define the plane stress transformed reduced stiffness matrix

as

(4.39)

Introducing this definition in Equation (4.38), we get

(4.40)

The above equation is written in expanded form as

(4.41)

Note:

is a symmetric matrix. Further, it is a fully populated matrix with non zero

coefficients. Thus, using Equation (4.33), Equation (4.35 in Equation (4.39), we can write the individual terms in expanded form as

(4.42)

Note: Reduced stiffness coefficients are fourth order in the sine and cosine functions.

Note:

are very important in that they define the coupling both in-plane normal and shear

responses. Figure 4.1 shows response of an isotropic and orthotropic material under traction. The behaviour of orthotropic lamina loaded along fibre direction and perpendicular to fibre direction is essentially similar to an isotropic material. However, for an off axis lamina, the behaviour clearly shows the coupling between normal and shear terms.

Figure 4.1: Normal shear coupling in orthotropic lamina The transformed plane stress constitutive equation can also be given in inverted form of Equation (4.40) as

(4.43)

where

(4.44)

Using

from Equation (4.33) and

from Equation (4.35) in the above equation, we get the

individual coefficients of transformed reduced compliance matrix as

(4.45)

Note: The same notation has been used for compliance matrices in principal directions

and

transformed directions

in 2D and 3D. This is because the corresponding terms are identical.

However, for the stiffness coefficients these are different in 2D and 3D. Note: One can see the difference between the stiffness values by algebra involved. The inverse of the compliance matrix for plane stress case is different from the inverse of the case. matrix for 3D

Thermal Effects:
Thermal strains in principal material coordinates are proportional to the temperature change These are given using coefficient of thermal expansion as .

(4.46)

where

Transformation of the thermal strains

to the strains

in global coordinates gives

(4.47)

Let

. Thus, Equation (4.47) becomes

(4.48)

or

(4.49)

Thermo-Elastic Constitutive Equation:


The total strain due to mechanical and thermal loading in principal material directions is given as (4.50)

We can write for the mechanical strains as

(4.51)

Thus, Equation (4.50) becomes

(4.52)

Re-arranging the above equation, we can write for the stresses as

(4.53)

The above equation can be written in global coordinate system as

Moisture

Effect:

The hygroscopic expansion in principal material direction is proportional to the amount of percentage weight of moisture absorbed. Further, the hygroscopic expansion will be in principal normal directions only. This expansion will not lead to any shear. Thus, we write the hygral strains in principal directions for planar problem as

(4.54)

Here, problem

denotes the coefficient of hygroscopic expansion in principal material directions for planar and denotes the amount by percentage weight of moisture absorbed.

Now let us transform the hygroscopic strains in global coordinate system as

(4.55)

Using Equation (4.54), we can write

(4.56)

where

(4.57)

It is clearly seen from Equation (4.48) and Equation (4.57) that way.

and

behave in a similar

Hygro-Thermo-Elastic

Constitutive

Equations:

When hygral and thermal effects are present along with mechanical strains, then the total strain in principal material direction is given as

(4.58)

Using Hookes low for mechanical strain and solving for stress the hygro -thermal constitutive equation, we get

(4.59)

Equation (4.59) can be written to give stresses in global directions as

(4.60)

where

and

are as given in Equation (4.48) and Equation (4.57), respectively.

Homework:

1. Verify the result given in Equation (4.36).


2. Using the invariance property of strain energy density function, show that:

and
3. Using the relation between engineering constants, show that and as given in Equation (4.24) and in terms of

are as given in Equation (4.29).

4. Write the compliance coefficients in Equation (4.45) in terms of engineering constants. 5. Using Equation (4.24) in Equation (4.42) obtain the individual terms of 6. For fibre orientation 7. The and obtain in terms of .

matrix for materials given in Table 3.1. is given as

matrix for a composite with fibre orientation of

8. Find all engineering constants for this material. 9. Write a computer code to calculate reduced transformed stiffness and compliance matrix for any angle of fibre orientation with respect to global coordinate system. 10. Extend the code written for the above problem to plot the variation of fibres between terms for orientation of

. Plot the variation for materials given in Table 3.1.

11. Write a computer code to plot the variation of thermal and hygroscopic expansion coefficients with fibre orientation between for T300/5208 composite.

Lamina Engineering Constants


Introduction
In the previous lecture we have derived constitutive equation for planar state of stress in a lamina. We have derived these constitutive equations in principal material and global directions. In this lecture, we are going to see a practical application of the planar constitutive equations in industry. Although, the engineering constants in principal material directions are known, it is difficult to comment on the engineering constant for off axis lamina, instantly. When laminae are used for designing a structure, the engineering constants in global directions become very useful for a quick estimate of the behavior of the structure under certain loads. Thus, for practical application purpose, the various lamina engineering constants are obtained and their variation for fibre orientation between for a range of

composite materials is given together. Thus, a designer can use the required lamina with appropriate fibre orientation and material.

Here we are going to obtain engineering constants for any off axis lamina as a function of engineering constants of that lamina in principal material directions and fibre orientation. This can be done with the help of lamina constitutive equation with appropriate one dimensional state of stress.

We have constitutive equation in global directions as given in Equation (4.5)

Thus, for a given state of stress in global directions we can find the strains in global directions from this equation.

Axial Modulus:
Consider a unidirectional off axis lamina. This lamina is subjected to the loading as shown in Figure 4.2. and

Thus, from Equation (4.5) for this state of stress we can write the axial strain as

(4.61)

The Youngs modulus in x-direction is now defined as

(4.62)

Thus, from Equation (4.61), we can write

(4.63)

Figure 4.2: Off axis lamina loaded in traction along x direction

In the above equation,

is written using compliance terms in principal material directions as

(4.64)

Further, it can be writing compliance terms in principal directions using engineering constants doing some rearrangements as

(4.65)

From this expression it is easy to see that the modulus . The variation of the modulus

when

and

when

with fibre orientation for AS4/3501-6 Epoxy material is

shown in Figure 4.3. The variation of the modulus is identical in nature.

for both positive and negative fibre orientations

Figure 4.3: Variation of axial modulus with fibre orientation for AS4/3501-6 Epoxy

Axial

Poissons

Ratio:

The axial Poissons ratio defined as

can also be obtained for above loading condition. This Poissons ratio is

(4.66)

Using Equation (4.61) we can write

(4.67)

Further, expressing the compliance terms in above equation in terms of engineering constants we can write

(4.68)

In-plane

Shear

Modulus:

The in-plane or axial shear modulus for an off axis lamina can be obtained when it subjected to a pure shear loading as shown in Figure 4.7. Thus, for this loading condition we have . and .

For this loading, we define the in-plane shear modulus

as

(4.75)

With the help of Equation (4.5) we rewrite this equation as

(4.76)

And in terms of engineering constants, it becomes

(4.77)

The variation of

with fibre orientation between

to

for AS4/3501-6 Epoxy material is .

shown in Figure 4.8. From this figure it can be seen that shear modulus is maximum when At the value of shear modulus is

(4.78)

Figure 4.7: Off axis lamina loaded in pure shear

Note: When the material is isotropic, that is


reduces to the familiar relation

and

, then the above expression

(4.79)

The minimum value of shear modulus is seen when the lamina is loaded in shear in principal material directions and its value becomes

(4.80)

Note: It is very important to note that the shear modulus of the lamina is a minimum when lamina is in
principal directions and a maximum when fibre orientation is lamina under same pure shear for fibre orientation fibre orientation of in the later section. or . Further, the behavior of a

is significantly different from that of lamina with

. The physical significance of this phenomenon is explained in greater details

Figure 4.8: Variation of shear modulus with fibre orientation for AS4/3501-6