Abstract
The prediction of polymer properties, based on its composition, it is a complex problem with no easy method to obtain directly and accurately results. Among the tribological properties, the friction coefficient and wear rate are the most interesting ones. The polymers based on epoxy resin, with clay as filler, show different properties depending on the clay concentration. This paper presents an analysis of the polymer properties variation with its filler concentration. Due to the tribological processes complexity, mechanical and thermal properties must be taken into account. The aim of this study is to find an optimal concentration value, with minimal influence on polymer properties, using neural network models. All value properties were used in order to optimize and predict the composite properties. Keywords: polymer properties, tribology, filler concentration, neural network, prediction