Estimation of in-plane thermal conductivity of copper clad board by inverse analysis using artificial neural networks

Abstract

This study presents a methodology to estimate the in-plane thermal conductivity of copper clad board used in electronic applications. Experiments are performed in vacuum environment and an inverse heat conduction problem (IHCP) is solved employing artificial neural networks to estimate the in-plane thermal conductivity. A comparison of estimates of thermal conductivity as obtained by solving the inverse problem using back propagation artificial neural networks trained using two algorithms namely Levenberg-Marquardt and Scaled Conjugate Gradient are presented.