| There are hundreds of millions of monitoring data available in real-time monitoring of bridge.It is difficult to establish real-time monitoring system to locate the damage position and damage state of bridge structure.currently,the application of several methods,due to the reasons for the error,the result is not ideal.Based on the finite element analysis software,this thesis simulates the mechanical and modal parameters of a large number of truss simply supported beam bridges when they are damaged and intact.The damage parameter model is established by the neural network regression analysis,cluster analysis and multilayer perceptron.The main work is as follows:(1)Based on the theory of modal,frequency,matrix,statistics and genetic algorithm of bridge structure damage identification,through comparative analysis,the damage mechanism of bridge structure and the related parameters of damage evaluation are studied,and the selection of evaluation parameters is completed.(2)Based on the truss simply supported beam bridge,the structure model is built by ANSYS software.The modulus of elastic modulus of single element steel and the modulus of elastic modulus of multiple unit steel are simulated respectively,and the mechanical data and modal data of the structural model are obtained.(3)Based on a large number of data simulated by finite element,using neural network regression analysis,clustering analysis and multilayer perceptron theory,the variable mechanical model of single unit single variable and variable modal model of single element variable,variable mechanics model of multiple unit and multivariable and variable modal model of multiple unit and multivariable are established.The critical range of damage and integrity is determined.(4)Through the method of data back calculation,the evaluation of the model is studied,and the evaluation system of structural damage location and damage degree is established by single variable and multivariable. |
PDF Full Text Request