Damage Monitoring Of Composite Structures Based On Neural Networks

This paper starts with the dynamic characteristics of structures, seeks the Damage Sign Parameters(DSP) of composite materials structures, which have essential relations with the damage/healthy status. These relations are ambiguous and aren't able to be formulated today. Making use of the Artifical Neural Network(ANN), those relations between DSPs and damage/health status are ascertained approximatively. By monitoring these DSPs in structures, healthy status could be supervised. The main efforts in this paper are paid out for seeking several appropriate DSPs by ultilizing Modal Analysis(MA) and Finite Element Analysis(FEA) for aeronautic typical structure(such as plate x sheik box), DSPs are computed by PATRAN, and efforts are paid out for simulating the BP Networks by ultilizing MATLAB to process of Structural Health Monitoring(SHM). Analysis and compution proves that those DSPs are precise to identify the damage of structures. So these DSPs could be applied in SHM for composite materials structures.