Research On Compression Sensing Method For Fault Diagnosis Of Subway Auxiliary Inverter

The auxiliary inverter is a device with a high probability of failure of the subway train,so it is especially important to carry out fault diagnosis.However,China has not yet formed a complete auxiliary inverter fault analysis and diagnosis method.Most of the fault diagnosis methods for inverters are based on the Nyquist sampling theorem.It is necessary to collect a large number of signals to ensure the accuracy of fault diagnosis,which will cause problems such as huge data processing,data storage and data transportation difficulties.For the above reasons,it is important to establish an intelligent fault diagnosis method for fault diagnosis of the auxiliary inverter.Firstly,based on the analysis of Compressed Sensing(CS)theory,the sparse representation method based on discrete cosine transform is used to illustrate the sparseness of the voltage signal of the auxiliary inverter,and the data sampling and data compression are combined into one.feature.Secondly,making full use of CS theory to compress the projection to obtain a small amount of data can preserve the characteristics of a large number of key information in the original signal.In the process of signal compression measurement,multi-frequency signal fusion,redundant dictionary and Gram matrix are adopted.The specific optimization method is used to optimize and improve the sparse basis of the sparse representation process and the measurement matrix in the measurement process.Simulation experiments show that after the compressed signal is optimized,not only the correlation between the sparse base and the measurement matrix is reduced,but also the compression measurement performance of the signal is significantly improved.Finally,the CS signal is processed by the CS method.On the basis of no reconstruction,the measured values obtained by the compressed projection are directly extracted,and combined with the support vector machine(SVM)to realize the auxiliary inverter.Fault diagnosis and identification.The simulation experiment further shows that the auxiliary inverter fault diagnosis and compression sensing method needs to process less signal data,which can ensure the accuracy of fault diagnosis and speed up the fault diagnosis.