Research On The Fault Diagnosis Of Compressed Sensing Asynchronous Motor Based On SPIN

At present,the communication technology of state monitoring and fault diagnosis of asynchronous motor in the process,which often need to increase the amount of collected information to ensure the reliability of diagnosis results,resulting in a large amount of data that needs to be processed,and resulting in a problem that the data storage requires a lot of space and transmission bandwidth.Since it is considered that the normal signal does not have the meaning of analysis in the fault diagnosis system,it can save more time and space to transmit and process the fault characteristic signal individually.Therefore,this paper proposes the application of Successive Projections to ontocoherent manifolds(SPIN)algorithm to fault feature extraction in the fault diagnosis of rotor breakage in induction motors,using motor current signal analysis(MCSA)method to analyze the asynchronous motor stator current signal of asynchronous motors.It is modeled as the sum of the fault component and the normal component signal,iterative projection reconstruction of the original signal,the component signals are recovered respectively,and then the purpose of extracting the fault feature signal is achieved.and then to extract the fault feature signal.The Support Vector Machine(SVM)is used to diagnose the fault part of the fault feature signal.Firstly,the basic theoretical framework of compressed sensing is studied,and the basic principle and the derivation process of SPIN algorithm are elaborated in detail.Based on the MCSA method,the stator current during asynchronous motor operation is mathematically modeled,to analyze the correlation between the normal frequency and fault frequency,and its feasibility and measurement matrix design was performed according to the requirements of the SPIN algorithm.Secondly,the mechanism of rotor broken bar fault in asynchronous motor is analyzed,and the simulation model of rotor broken bar fault in asynchronous motor is built by MATLAB/SIMULINK,which provides the required data for subsequent progress.The stator current signal and the corresponding rotational speed of the asynchronous motor during normal operation and rotor broken bar fault operation are obtained through simulation,and the stator current time domain signal is converted into frequency domain by using FFT.Then,SPIN algorithm is programmed through MATLAB to realize the separation and reconstruction of normal signal components and fault components in the stator current signal of asynchronous motor when rotor broken bar fault is running.Experiments show that the SPIN algorithm can completely reconstruct components from the low sampling rate of M / N = 1.3%,and successfully separate the fault components,so that the data processed by the fault detection system will be further reduced,and the operation efficiency of the system will be greatly improved.Finally,the SVM is applied to the fault diagnosis system of the rotor broken-bar in the asynchronous motor,and the fault diagnosis results of the faulty signal sampled through compressed sensing are obtained.The training samples of different parameters are obtained by using the simulation model,and then the feature quantity is extracted according to the characteristics of the rotor broken bar of the asynchronous motor.After testing,the diagnosis effect can meet the practical application.