Structural Damage Monitoring Based On Vibration Modal Analysis And Neural Network Technique

Structural health monitoring is a hot issue both in engineering andacademic domain. In the thesis, a lot of domestic and foreign papers about structural damage detection based on vibration modal analysis have been analyzed, induced and concluded systemically. Dividing structural damage detection into three sub-modules is presented originally, which are structural damage identification module, structural damage localization module and structural damage severity determination module, respectively. Associated to three modules, different signatures for damage detection, which are sensitive to damages, are presented.Signatures for damage detection formed by modal parameters always need solve intricate mathematic iteration problem, which makes it difficult to realize real-time and on-line structural damage identification. Neural network features in its anti-linear mapping ability, which can change inverse problem into forward problem. Vibration modal analysis is integrated with neural network in the thesis. Damage signatures for damage detection formed by vibration modal parameters are inputted to neural network as eigenparameters for structural health monitoring.Being aimed at the main drawbacks of slowly learning convergent velocity and easily converging to local minimum of the Backward-propagation (BP) network is optimized. Later by applying software emulation and numeric emulation, good effects are gained. By adopting the optimized BP network to process damage detection on a rectangular beam, satisfactory outcome is gained. Above research result indicates that the integration between neural network methods and vibration modal analysis is feasible and presents a new technological way for structural health monitoring.