Thermal Conductivity Prediction Method of Fiber-Reinforced Material with Microstructure Identification

Abstract

An innovative method for predicting the thermal conductivity based on the recognition of microstructure features is established. Digital image processing technology is applied to analyze the electron micrograph of typical long fiber-reinforced composites. A novel method for determining the representative volume element is proposed. Once a set of representative volume elements has been established, actual physical information is taken into account by using a technique of image identification for the actual random distribution of fibers. Using the finite element method, the mean value and the standard deviation of the effective thermal conductivity based on the set of representative volume elements are obtained. The laser flash method was applied to measure experimentally the effective thermal conductivity. The novel numerical method proposed herein has been evaluated against these experimental data. It is found that the accuracy of the effective thermal conductivity prediction is guaranteed by image recognition of scanning electron microscope images. For more accurate information on the actual fibers, a random distribution is introduced into calculations. All the results show that the effective thermal conductivity prediction of the fiber-reinforced composite using microstructure identification is reasonably accurate and applicable. Close agreement with the experimental data is obtained. The average relative error is .