Damage Detection Based On BP Neural Networks For Tall Frame Structures

In recent years, the Artificial Neural Networks has become hit of research, widely used in civil engineering, and taken a very important role in the field of damage detection.Firstly, for state of damage detection based on BP Neural Networks, the basic principle of the Artificial Neural Networks, the design of the model of BP Neural Networks, and four kinds of modified BP algorithms have been discussed. Owing to the advantaged of BP Neural Networks, i.e. non-linear, tolerance and robust, it has been playing a very important role in damage detection.Last, the basic principle of damage detection has been discussed, the four-stage approach based on BP Neural Networks has been proposed for damage detection of tall frame structures. The occurrence, type, location and extent of damage are detected separately in this approach. In order to detect the damage of symmetric frame structures effectively, the input parameter of BP Neural Networks has been discussed, and a composite input vectors include the Damage Signature has been constructed. The effect of filler wall has also been discussed in this paper. The strong ability of damage detection based on BP Neural Networks has been verified in a numerical case study by using an eleven-storey symmetric frame. However, further research should be done to improve the precision of damage detection.