Application of Artificial Neural Network in Composite Research

Artificial neural network is a technique with flexible mathematical structure and lots of characteristics such as parallel distributed processing, nonlinear processing and so on. So the artificial neural network becomes a common method to solve complex problems in research of material science by building a model. This article uses BP and RBF neural network to study the impact from components of composite materials, process conditions on properties of composite materials. We establish the relational model among the third element in composition, hot dipping temperature and shear stress which can reflect the joint face strength of Pb-Al composite materials, and give the model verification by using experimental data. The results which show that the neural network model can be used to predict the shear stress when change the third element in composition and the hot dipping temperature.