Research On Fault Diagnosis Method Of Explosion-proof Squirrel Cage Motor Under Variable Frequency Power Supply

With the development of frequency conversion technology,more and more explosion-proof squirrel cage motors in coal mines use frequency conversion power supply to obtain better starting and speed control performance.The diagnostic methods for typical faults of squirrel cage motors under power supply conditions are mature,but under the condition of variable frequency power supply,some of these methods reduce the diagnostic accuracy,and some methods even failed.Therefore,it is still of great significance to study the diagnosis method of squirrel cage motor under the condition of variable frequency power supply.Firstly,Analysing the three-phase output voltage's harmonics of SPWM controlled inverter.On this basis,the fault characteristic mechanism of motor is analyzed,including inter-turn short circuit,rotor broken bar,air gap eccentricity and rotor broken bar and air gap eccentricity.And then obtain the fault characteristic frequency of stator current under each fault.Secondly,taking YB3-160M-4 explosion-proof squirrel-cage motor as the research object,the finite element model of the motor is established in Ansoft Maxwell and the SPWM control circuit is established in Simpler,then the complete module of squirrel cage motor fed by variable frequency power supply is established by joint simulation under two software.And the normal operation of the moor,turn-to-turn short circuit,rotor broken bar,air gap eccentricity,broken bar and eccentricity compound are built On this basis.In the model,the frequency spectrum analysis of stator current in power-frequency power supply and variable-frequency power supply is carried out.It is found that the effect of stator current spectrum analysis on fault diagnosis of motor supplied by variable-frequency power supply is not satisfactory,and the fault characteristic component of high-frequency part is difficult to identify.Then,notch filter method is used to filter the fundamental component of stator current,EMD method is used to extract the fault feature component,and EEMD decomposition method based on K-L divergence is used to improve the mode aliasing and false component in EMD,which is the core content of HHT.The normalized energy of IMF signal after eliminating the false component is used as the fault feature component to input the neural network.Finally,the artificial bee colony algorithm is used to optimize BP neural network in motor fault pattern recognition,and an improved searching artificial bee colony algorithm is given to optimize the weight and threshold of BP neural network.The recognition results show that the fault recognition rate of the improved artificial bee colony algorithm optimized neural network is 5.16% higher than that of BP neural network,which verifies the feasibility of the improved method.