Prediction of fatigue strength of composite laminates by means of neural networks

In this paper a first attempt on the use of a neural network for the prediction of fatigue strength of composite materials is carried out, First some information about the composite used are given: it is on a graphite-peek composite with an average fibre content of 63%, Next a brief explanation about neural networks and how they work are given; in particular we have used a back propagation neural network. Finally the results of the application of the neural network to our problem are discussed; in particular we have established that the neural network technique is a very attractive tool to model a set of experimetal data, but it should be used very carefully to avoid wrong predictions.