Research On Asynchronous Motor Rotor Fault Diagnosis Technology Based On Stator Current Characteristic Analysis

With the continuous deepening of industrialization,asynchronous motors are widely used in industrial production industries,especially in the coal industry.In the coal industry,the normal operation of the motor can ensure high efficiency,safety and speed in the production process.Therefore,the research of asynchronous motor fault diagnosis technology is of great significance.In this paper,the common rotor broken bar and eccentric faults in asynchronous motor faults are taken as examples to analyze the mechanism of asynchronous motor broken bar faults and eccentric faults.The motor models under normal and fault conditions are established through finite element simulation,and then the corresponding stator current signals are collected.By analyzing and comparing the current spectra under different conditions,on the one hand,it verifies the correctness of the stator current signal data extraction,on the other hand,it provides reliable and effective data support for the subsequent feature extraction methods.Due to the non-linear and non-stationary characteristics of the current signal during the operation of the mine asynchronous motor,the emergence of fractal geometry opens up new ideas for feature signal extraction.Since wavelet transform and fractal geometry have self-similarity,the tunable Q wavelet fractal dimension is used to extract the characteristics of current signal,which can characterize the characteristics of the signal better than other fractal dimension extraction methods.The empirical mode decomposition is also suitable for processing nonlinear signals,but when the signal is decomposed,there will be phenomena such as modal aliasing,which may greatly reduce the diagnostic recognition rate.Based on this,the ensemble empirical mode decomposition is used to extract the current signal,and the entropy value of different ensemble empirical mode decomposition is compared through simulation,and it is found that energy entropy is more obvious as a feature set to distinguish faults.Then the support vector machine classifier was used to identify the fault,but the traditional support vector machine classifier had parameter optimization problems,so the particle swarm optimization and gray wolf swarm optimization methods were respectively applied to the support vector machine parameter optimization,and the particle swarm parameters were established Optimize the support vector machine and gray wolf pack parameter optimization support vector machine fault diagnosis model.In the simulation experiment,the tunable Q wavelet fractal dimension feature set and ensemble empirical mode decomposition energy entropy feature set were input into the support vector machine classifier and the support vector machine classifier after iterative optimization,respectively,to conduct a comprehensive comparative analysis.Comprehensively obtain the tunable Q wavelet fractal dimension feature,combined with the gray wolf pack optimization support vector fault diagnosis model,the diagnosis recognition rate is the highest among these methods,and the diagnosis time is the shortest.Finally,a new rotor fault diagnosis technology based on the tunable Q wavelet fractal dimension and GWO-SVM is established,and it is applied to engineering experiments to verify it,and a high recognition rate is obtained.