The Research On Structural Damage Identification Based On Wavelet Genetic Overrun Learning Machine Algorithm

The building structure will face various complex environments in the process of construction and use,which may lead to structural damage.With the accumulation of damage,the building structure may fail to reach the expected life span,and more serious,it may collapse in the service process,causing harm to people’s life and property safety.Therefore,the research on structural damage identification has important theoretical significance and engineering application value.In this paper,the problem of structural damage identification is studied,and the random selection of weights in the over-limit learning machine algorithm is optimized to improve the recognition accuracy without losing the recognition speed.A structural damage identification method based on wavelet genetic overrun learning machine(GA-ELM)is proposed.Through numerical simulation and experimental study,it is found that this method can accurately identify the damage location and damage degree of the structure,and solve the shortcomings of wavelet transform in damage degree identification.Firstly,genetic algorithm(GA)and particle swarm optimization(PSO)are used to optimize the weights of the overrun learning machine,respectively.By comparing the results,it is found that the optimization of the overrun learning machine by genetic algorithm is more suitable for the field of structural damage identification.The improved algorithm significantly improves the accuracy of damage degree identification.Then based on the principle of wavelet singularity,the improved algorithm is combined with wavelet analysis theory,and the wavelet GA-ELM algorithm is proposed.Secondly,in order to verify the effectiveness of the proposed method,this paper verifies the proposed method through numerical simulation and experimental research.Firstly,the three-dimensional finite element model of the beam structure with damage is constructed by using the large-scale finite element software ANSYS,and the modal parameters are obtained.Then,the strain modal parameters of the structure are obtained by differential treatment,and the preset damage location is identified by using the wavelet singularity principle.Then,the three-dimensional model is sliced into space,and the strain modal data of each layer are brought into the GA-ELM algorithm program written by MATLAB for damage identification.Then,based on the numerical simulation model,the damage identification test of the fixed beam structure with multiple damage conditions was designed.After the modal parameters of each layer of the beam structure were measured by the instrument,the difference was converted to the strain mode,and then the data were denoised and the above method was used for damage identification.Numerical simulation and experimental results show the effectiveness of the proposed method.Finally,the proposed method is applied to the continuous beam structure,and the threedimensional finite element model of the continuous beam structure is established by ANSYS.Using this method,the identification of the damage location and damage degree of the continuous beam structure is successfully completed.It further shows that the method has certain reference value for practical engineering.