PREDICTION OF EFFECTIVE THERMAL CONDUCTIVITY OF POLYMER COMPOSITES USING AN ARTIFICIAL NEURAL NETWORK APPROACH

The effective thermal conductivity (ETC) of polymer composites is studied using artificial neural networks. Artificial neural networks are a form of artificial intelligence, which attempt to mimic the function of the human brain and nervous system. Artificial neural networks provide a great deal of promise but they suffer from a number of shortcomings, such as knowledge extraction, extrapolation, and uncertainty. This paper presents the use of the artificial neural network for prediction of ETC of metal-filled polymer composites due to their increasing importance in many fields of engineering applications and technological developments. Artificial neural networks models are based on a radial basis with the training function: the more efficient design radial basis network (NEWRB) and the feedforward backpropagation network with training functions conjugate gradient with Powell-Beale restarts, Levenberg-Marquardt, one-step secant, random order incremental, and resilient backpropagation. The volume fraction and thermal conductivity of continuous (matrix) and dispersed (filler) phases are input parameters to predict the ETC. The resultant predictions of ETC using the different models of artificial neural networks agree well with the available experimental data.