AE behaviors evaluation with bp neural network

This paper reports our experiment on training a three layer forwards neural network with backpropagation algorithm (BP) to memorise acoustic emission (AE) behaviors of magnesium alloy during fatigue test process and perform the mapping from AE behaviors to fatigue crack propagation. This study reveals the potential possibility of using artificial neural networks for automated AE testing. Simulation results are very encouraging and conclusively in favor of this attempt.