Thermal conductivities estimation in orthotropic materials making use of global optimization algorithms

Abstract

There are situations where the knowledge of thermodynamic properties such as thermal conductivity for the present case is required. In some of them an additional requirement appears, as the measurement has to be made along the three perpendicular space axes. In the present article, three thermal conductivities that appear in orthotropic materials were predicted by solving an inverse heat transfer problem. This inverse problem was solved using the Cuckoo algorithm, the deterministic Levenberg-Marquardt, and with the new hybrid of these two. It was found that these three strategies produce excellent results when compared to each other. Nevertheless, the hybrid algorithm proved to be more efficient than its precursors in solving the present problem. The hybrid algorithm consumed in average less computing time compared to the metaheuristic algorithm and extended the search range in comparison to the deterministic one, always maintaining precision in its results.