Research On The Method Of Structural Damage Detection Based On GA-BP Neural Network

In recent years the structure damage detecting has become one of the hottest topics both in engineering and academic world, and has been done broad research by the researcher all the world. In practical engineering, there always is some extents nonlinear quality in the engineering structure with damage, which would bring much difficulty for solving the problem. And there may be not a wanted result. At present, artificial neural network (ANN) and genetic algorithm (GA) etc are being used for the structural damage detecting. In this thesis, from the point of the combination of both NN and GA, to research the structural damage detecting making use of the merit of NN and the GA is conductive to the practical application of engineering, and also with the stronger background for engineering.In this thesis, three structural damage detection methods, including: static state identification approaches, dynamic identification approaches, and ANN combined with GA aptitudinal identification approaches, are summarized, on the basis of analysis of the data about the structural damage detection, ANN and GA. The theory about ANN and GA is summarized, and the two method's property is studied. Then the GA-BP neural network is proposed, based on the combined ANN and GA, which is used to study the structural damage detection. Through numerical verification of the damage case on a simple beam and a four-story frame, the proposed method, is applied to detect the two structural' damage cases. It was proved the proposed method has a good precision in detecting, in the single damage and multi-damage, and the result is worth to forward study.