Research On Vibration Source Localization Based On Compressed Sensing

With the rapid development of industrial technology,plates are widely used in various fields,and their quality requirements are getting higher and higher.Plates are prone to cracks,erosion and other damage during use.Damaged plates are prone to damage to property and personal safety.Therefore,health monitoring is one of the necessary tasks.For the structural damage on the plate,vibration will occur during the process of damage,that is,damage excitation.In this paper,the health monitoring of plate structure is realized by receiving the vibration signal generated by damage excitation.In this process,a large amount of data will be generated,resulting in an increase in the cost of vibration signal sampling and serious time-consuming.The introduction of compressed sensing technology can reduce the signal sampling rate,reduce the cost of sampling,data storage and transmission while ensuring the signal quality,thereby improving the efficiency of non-destructive testing.In this paper,the vibration source localization based on compressed sensing is studied.The contents are as follows :1)Compressed sensing theory and wave superposition theory are the theoretical basis of this paper.The compressed sensing theory and signal reconstruction algorithm are mainly discussed,and the performance of various algorithms is verified by data experiments.This paper introduces several typical algorithms in detail,and calculates the relationship between the number of tests and the probability of reconstruction.In addition,the propagation equation of lamb wave and the method of locating vibration source by wave superposition principle are briefly introduced.2)Damage excitation and vibration signal receiving are simulated by simulation experiments.Aiming at the problem of signal distortion caused by sparse sampling of vibration signals,the signals are sparsely sampled by equal interval and bicubic interpolation respectively,and the sampled signals are compared.The measurement matrix is constructed by using the sampling signal,and the optimization matrix is obtained by orthogonalizing the measurement matrix.Combined with relevant theoretical knowledge,the maximum value of the absolute value of the column correlation coefficient before and after the optimization of the measurement matrix is compared.Different algorithms are used to test the reconstruction effect of the measurement matrix and the optimization matrix.3)The feasibility of the vibration source localization method based on compressed sensing is verified by experimental simulation.The propagation of lamb wave in infinite healthy aluminum plate is simulated by dynamic finite element software,and the experimental simulation process of lamb wave propagation is studied and analyzed.According to the theory of compressed sensing and using MATLAB software to locate the vibration source of aluminum plate,the experimental results are analyzed.In the aluminum plate vibration source localization experiment,different algorithms and measurement matrices are used for signal reconstruction.Combining the principle of wave superposition and the principle of compressed sensing signal separation,the vibration signal is reconstructed and the vibration source is located.