Engineering Structural Damage Detection Based On Modal Parameters

With the technological development and the requirement of human self, the modern construction is developing with large-scale,complicated , intelligent and so on. We have the need to found an effective structural health monitoring system, it will be very useful to prevent and discover the damage, it will be also very helpful to insure that the construction runs well. The system also has a great social and economical benefit.In this paper, it expounds the basic theory and the current research situation of structural health monitoring, it studies the theory and methods of structural damage detection based on modal parameters; it studies the theory of neural network especially the theory of BP neural network, illuminates the academic reference that how to use neural network doing structural damage detection. Based on this, bring forward the BP neural network structural damage detection method based on modal parameters. Using MATLAB6.5 BP neural network toolbox, compiles the structural damage detection process with the input parameters based on the modal parameters of frequency and mode shape. And it uses a five-floor frame model to verify this method. The damage of the frame model uses the reducing of elastic modulus to simulate and uses ANSYS8.0 software doing finite element simulation calculation. It calculates the mode parameters of the model under different damage situations to acquire the learning samples and testing samples for the BP neural network. The test indicates that this method can detect the position and degree of the damages.Several conclusions can be drawn from this paper. First, using BP neural network to detect structural damage is feasible, this can be full use of the pattern recognition and self-adaptation ability of neural network. Second, the test of the five-floor frame model indicates that with the input of the frequency square variety and mode shape, using BP neural network, it can be detect the position and the degree of the damages. Third, this paper also test the deductive ability of the BP neural network, it indicates that the trained and optimized BP neural network has a good deductive ability. Fourth, the BP neural network structural damage detection method based on modal parameters can convenient and effectively doing structural damage detection.