Abstract
Using a multiple-layer feed-forward artificial neural network (ANN), the specific wear rate and frictional coefficient have been predicted based on a measured database for short fibre reinforced polyamide 4.6 (PA4.6) composites. The results show that the predicted data are well acceptable when comparing them to the real test values. The predictive quality of the ANN can be further improved by enlarging the training datasets and by optimising the network construction. A well-trained ANN is expected to be very helpful for an optimum design of composite materials, for a particular tribological application and for systematic parameter studies.