Neural network based prediction on mechanical and wear properties of short fibers reinforced polyamide composites

The artificial neural network technique was applied to predict the mechanical and wear properties of short fiber reinforced polyamide (PA) composites. Two experimental databases were used to train the neural network: one consisted of 101 independent fretting wear tests of PA 4.6 composites; the other one was from a commercial company and included 93 pairs of independent Izod impact, tension and bending tests of PA 6.6 composites. The predicted property profiles as a function of short fiber content or testing conditions proved a remarkable capability of well-optimized neural networks for modeling concern.