Research On Rotor Broken Bar Fault Diagnosis Of Asynchronous Motor Based On Machine Learning

Asynchronous motor is widely used in national industrial production.Brkoen rotor fault of asynchronous motor will damage the motor itself,resulting in substantial economic losses and disastrous consequences.Therefore,the realization of broken rotor bar fault diagnosis is of great significance at the early stage of fault.This paper proposes a new method to diagnose the number of broken rotor bars of induction motors based on the estimation of signal parameters via rotational invariance technique(ESPRIT),pattern search algorithm(PSA),and light gradient boosting machine(LightGBM).The performance of ESPRIT-PSA is tested with the Matlab-based simulated instantaneous reactive power signal.The results show that ESPRIT-PSA can identify the broken rotor bar fault-associated components in the instantaneous reactive power signal even with the short-term sample.Recently,motor instantaneous reactive power signal analysis(MIRPSA)has been used to detect induction motors’ broken rotor bar fault due to its effectiveness in low slip situations.However,this method cannot accurately diagnose the number of broken rotor bars.Therefore,a broken rotor bar fault experiment is conducted using a3 k W Y100L-2 three-phase asynchronous motor on the experimental platform to construct the training data and label sets.Then train the data sets with random forest,support vector machine,Ada Boost,and LightGBM.LightGBM algorithm is selected for its highest accuracy and the shortest operation time.The results show that the model optimized by parameter adjustment can accurately diagnose the broken rotor bar fault.