AXIAL SHEAR MODULUS OF A FIBER-REINFORCED COMPOSITE WITH RANDOM FIBER CROSS-SECTIONS

A study is made of the effective axial shear modulus of a fiber rein- forced material with random fiber cross-sections so that the micromechanics is governed by stochastic differential equations. A coarse-graining procedure is adopted to investigate the macroscopic behavior of the material. This analysis leads to the formula for the effective axial shear modulus,

been developed for the effective viscosity and the effective thermal conductivity in random suspensions.Einstein's formula [i] is well known as the first theo- retical result for the effective viscosity in dilute suspensions of shperical particles.In recent years, Beran [2] and Miller [3] have used perturbation meth- ods combined with Green's functions to derive the expression for the effective thermal conductivity in suspensions.On the other hand, Hori [4], Hori and   Yonezawa [5][6][7][8] have obtained the rigorous formula for the effective thermal con- ductivity in a random medium using diagrammatic expansions from a field-theoretical view point.In spite of this progress, no results are available for the effective viscosity in more concentrated suspensions.
In a very recent paper, Nagatani [9] has presented a general theory of the macroscopic motions of slow viscous flows in suspensions from the view point of statistical continuum mechanics.With the assumption that the characteristic length scale is much larger than the size of the suspended particles, he derived the coarse-grained equations from the governing equations using the formal pertur- bation methods and coarse-graining procedures.The macroscopic motions on much larger scale than the size of the suspended particles.It has been shown that this formula reduces to the Einstein formula for dilute suspensions.Nagatani's formula for the effective viscosity has a simple closed form and has been found to agree well with experimental results.
In the derivation of the formula, the suspension with its variable physical properties on a microscipic scale, is governed by stochastic differential equations.
The macroscopic behavior of the system could be obtained by ensemble averaging the equations.However, this procedure is usually difficult and an alternative "coarsegraining" procedure is adopted.The coarse-graining means downgrading the spatial resolution to a much larger scale than that of a particle.This is achieved by looking at the particles as random perturbations and representing the field vari- ables by Fourier series in the space co-ordinates.Modes with low wave numbers (lower than the reciprocal of the size of a particle) are retained while those with high wave numbers are repeatedly eliminated through a formal perturbation technique.This yields the expanded form of the effective viscosity, which when formally resumed gives the formula.
We adopt this procedure to determine the effective axial shear modulus of a random composite, with unidirectional fibers.This analysis leads to the formula for the effective axial shear modulus, * i/" i 2c ('2 I)/(2 + I where UI and 2 are the shear moduli of the matrix and fibers respectively and c is the concentration of the fibers less than 0.5.The fibers with their random cross-sections are looked upon as random perturbations over the matrix material, and as such the concentration c of the fibers can not exceed 0.5 if c exceeds 0.5 we can interchange the roles of the fibers and the matrix) a range which is important from the practical point of view.For c 0, it reduces to the shear modulus of the matrix and for c 0.5 it yields the mean of the fiber and matrix moduli.This shows the consistency of the result, since the fiber geometry plays no role in the method, and is random.An important feature of the formula is its simple structure and can be of practical use.Even when the fibers are completely identical (which is difficult to ensure in practice) the formula would be approxi- mately true.
The formula is compared with the bounds obtained by Hashin [I0] and it is shown to exceed his lower bound.For higher values of the concentration, it, how- ever, exceeds his upper bound also.
2. BASIC STOCHASTIC EQUATIONS ON THE MICROSCOPIC SCALE.
We consider a cylindrical specimen of the material with side surface parallel to the fibers.Let the specimen be in a state of shear in the axial direction.
If we introduce co-ordinate axes 0X I and 0X 2 in a cross-section, the shear stress To on it is given by w where w(x), x (Xl, x 2) is the accompanying displacement in the axial direction and is the shear modulus of the material.If i and 2 are respectively the shear moduli of the materials o the matrix and the fibers, takes these values in the two phases.Looking upon the fibers as random perturbations over the matrix, we write where ] ]i + (]a2 i (x) (2.2) (x) 0, in the matrix I, in the fibers.
The function (x) is a random variable with mean (2.3ab) where S is the area of the cross-section and c is the 'concentration' of the fibers.
In the statistical analysis of the problem, we shall also encounter the mean of {(x)} n.Noting from equation (2.3) that its value is that of (x), we also have The stress equation of equilibrium of the material, with equation ( which apparantly is a stochastic differential equation.On the microscale the displacement and stresses must be continuous across the interfaces of the two phases.

COARSE-GRAINED EQUATIONS.
In order to obtain the macroscopic behavior from the stochastic equations (2.1) (2.6), we subject them to the coarse-graining procedure.This envisages downgrading the spatial resolution to a scale much larger than the cross-section of a fiber.To this end we represent the variables in equations (2.1) (2.6) in Fourier series: S-E Xke~k Ujk Plikj Wk4, + (2 Pl S-1/2 i(kj k'.)3 ' Wk-,k   It is apparant that, in the summation on the right hand side, all the wave numbers k' contribute to the component Ujk.
ter corresponding to the lowest wave number k' O, we get Ujk If from this summation we separate out the where we have used the relation 1/2 < >. (3.5) X 0 S Adopting the same procedure for equation (2.6) with (2.2), (3.2) and (3.3), we get In the coarse-graining procedure we repeatedly use equation (3.6) in the right hand side of (3.4) and remove the term corresponding to the high wave numbers k' 0 at each step.Physically this means downgrading the spatial resolution of the composite.As the wave numbers should occur with the diameter of a fiber in a natural way, the resolution would be on a scale much larger that the cross-section of a fiber.At the first step we have If we assume that the composite is statistically isotropic, the second order tensor (kj kj) (k m kl) is approximately an isotropic tensor of order two with The second term on the right hand side of (3.7) thus becomes 2 where we have used the relation (3.5) and the equation (3.3)The expression (4.1) gives the correct value I when c 0. When 2 > i' it increases with c.However, when c approaches (2 + i)/[2(2 I )] > 1/2 it becomes infinitely large.However, we note that we have looked upon the fibers as random perturbations over the matrix material and as such c should not exceed 0.5.
For c 0.5 we get * (I + 2 )/2' which is a consistent result, as the fiber geometry is random.If c exceeds 0.5 we should interchange the roles of the fibers and the matrix and we have This case is,howevernot very important from a practical point of view.
Hashin [i0] has given lower and upper bounds for *.If these are denoted by * and then in our notations L , L * Also the rate of increase It easily follows from (4.1) and (4.2) that * > L of * is higher thmn that of For higher concentrations the deviation from the lower bound will then be pronounced.This behavior is generally in agreement * for with experiments.In relation to the upper bound however, N* We consider Boron fibers in Aluminum matrix.For such a composite H2/ I 25/3.87.We can easily compute the normalized shear moduli given by equations The figure indicates that * is closer to the lower bound for values of c up to 0.35; thereafter, it is closer to the upper bound.
ACKNOWLEDGEMENT.The second author wishes to express his grateful thanks to East Carolina University for partial support of this work.

Call for Papers
As a multidisciplinary field, financial engineering is becoming increasingly important in today's economic and financial world, especially in areas such as portfolio management, asset valuation and prediction, fraud detection, and credit risk management.For example, in a credit risk context, the recently approved Basel II guidelines advise financial institutions to build comprehensible credit risk models in order to optimize their capital allocation policy.Computational methods are being intensively studied and applied to improve the quality of the financial decisions that need to be made.Until now, computational methods and models are central to the analysis of economic and financial decisions.However, more and more researchers have found that the financial environment is not ruled by mathematical distributions or statistical models.In such situations, some attempts have also been made to develop financial engineering models using intelligent computing approaches.For example, an artificial neural network (ANN) is a nonparametric estimation technique which does not make any distributional assumptions regarding the underlying asset.Instead, ANN approach develops a model using sets of unknown parameters and lets the optimization routine seek the best fitting parameters to obtain the desired results.The main aim of this special issue is not to merely illustrate the superior performance of a new intelligent computational method, but also to demonstrate how it can be used effectively in a financial engineering environment to improve and facilitate financial decision making.In this sense, the submissions should especially address how the results of estimated computational models (e.g., ANN, support vector machines, evolutionary algorithm, and fuzzy models) can be used to develop intelligent, easy-to-use, and/or comprehensible computational systems (e.g., decision support systems, agent-based system, and web-based systems) This special issue will include (but not be limited to) the following topics: • Computational methods: artificial intelligence, neural networks, evolutionary algorithms, fuzzy inference, hybrid learning, ensemble learning, cooperative learning, multiagent learning equations (2.2), (3.1), (3.2) and(3.3) in (2.1), we obtain

( 4 .
Fig. I: At the next step of the coarse-graining procedure, we again replace Wk_k,_k,, If we use equates .4) and (2.5) in(3.11),we get the expanded form of * in terms of c.If the infinite series is summed formally, we get the closed form

•
Application fields: asset valuation and prediction, asset allocation and portfolio selection, bankruptcy prediction, fraud detection, credit risk management • Implementation aspects: decision support systems, expert systems, information systems, intelligent agents, web service, monitoring, deployment, implementation