Damage detection in smart structures using neural networks and finite-element analyses

An important aspect of the smart structures concept is the automated structural health monitoring of aircraft structures. This requires detecting damage and assessing its effect on structural performance. The authors present a new approach to detecting and defining large area damage on a structure. The approach is based on using a neural network to deduce the damage size and location from measured strain values at discrete locations. The neural network is trained using results from finite-element analyses. Several examples illustrating this approach are presented.