Thermal conductivity estimation in continuous fiber metal matrix composites with random distributions

The real values of thermal conductivity for metal matrix composites reinforced with randomly distributed continuous fibers must be obtained by direct measurement since, for the same weight (or volume) percentage of their components, these values may change-substantially from one sample to another. Typical values supplied by the manufacturers (generally used in many applications) are calculated with approximate expressions that take into account the volume or weight fractions of each component in the composite, without considering the relative positions of fibers in a particular distribution. In this work, the extreme values of thermal, conductivity of a large number of random distributions for different volume percentages have been calculated numerically in a steady-state heat conduction process in order to estimate an approximate mean value. Numerical solution is carried out using the network simulation method, which is based on the equivalence between thermal and electrical finite-difference differential diffusion equations. A 2-D (15x10 volume elements) network model, which contains the assumed isothermal boundary conditions, is designed. Each one of the random distributions leads to a particular model which is run in appropriate standard software. Numerical solution for the thermal conductivity is simply derived from the (stationary) integrated heat flux. Final values of this property are tabulated and depicted for each group of samples and each volume percentage. Examples of continuous fiber metal matrix composites extensively used in industrial applications are studied.