Random Homogenization And Finite Element Analyses Of Thermodynamics Properties For Composite Materials

With the development and application of the composites, composites reinforcedby particles have been attractively paid more attention in the wide use such as optical ormechanical components as well as electronic packaging due to its unique advantages likelow price and superior performance.In this work, thermodynamics properties (thermal physical properties and thermalstress) for particle reinforced composites are addressed. The main work is as follows:(1) Based on the homogenization analytical results of the effective thermalphysical properties (thermal conductivity coefficient and thermal expansioncoefficient), the randomness of the each constituent’s property and the volume fractionas well as the correlation among these random variables are considered simultaneously.The mean values and mean square deviations of analytical results of the thermalphysical properties are derived from Random Factor Method (RFM). Results from theRandom Factor Method and the Monte-Carlo Method (MCM) are compared with eachother through numerical examples, and impacts of randomness and correlation amongrandom variables on the random homogenization results are carefully investigated bytwo methods. It proves that the Random Factor Method is found to deliver rapid resultswith comparable accuracy to the Monte-Carlo approach.(2) As a property of material, thermal expansion coefficient directly gives rise tothe thermal residual stress, and the uncertainty of thermal expansion coefficient willinfluences the magnitude and distribution of thermal residual stress in material as well.ANSYS is used to create a two-dimensional model of the composites, and thermalresidual stress is inspected by changing the parameters of matrix phase and particles’size and the volume fraction as well as shape. The results indicate that thermal residualstress easily occurs during the cooling process because of the difference of thermalexpansions between matrix and particle, and the maximum stress distributes on theside of matrix close to interface in composites.