Abstract
This work presents a computational method for the identification of the thermal conductivity of a powdered composite. The thermo-physical properties of powdered composites depend not only upon the intrinsic material properties of the filler and the matrix, but also upon several other parameters including the packing density, the particle shape factor and the particle size. In this paper, a genetic algorithm-based model is proposed for the identification of the effective thermal conductivity of a Bakelite-graphite powdered composite. A comparative analysis is also developed between the genetic algorithm and the experimental, theoretical and finite element results. In comparison with the experimental observations, the genetic algorithm model was found to have error values of approximately 5%, whereas the errors resulting from the theoretical models were up to 12%.