Research On Bearing And Misalignment Compound Fault Diagnosis Of Servo Motor

One of the key technologies of intelligent motor drive is to realize the fault prediction and health management of the equipment.The health monitoring of rotating machinery is an important task to ensure the reliability of industrial process.Based on this research background,National Guideline on Medium and Long Term Program for Science and Technology Development(2006-2020),Made in China 2025 and Report on the Development Strategy of Mechanical Engineering(2011-2020)all give priority to the fault monitoring,early warning maintenance and health management technology of rotating machinery and its key parts such as motor system.Therefore,the research on bearing and misalignment fault diagnosis of servo motor is carried out.This paper proposes a compound fault diagnosis method based on the speed signal of servo motor.Firstly,the additional torque expression generated by single fault excitation is deduced based on the existing theory.Then,the change of motor speed caused by the compound fault excitation is explored.The feasibility of bearing and misalignment compound fault diagnosis is analyzed theoretically.Then,a double closed-loop speed control model of PMSM under the excitation of composite fault is built in Simulink.The simulation results verify the correctness of the compound fault diagnosis mechanism of the speed method.In order to get closer to the actual working conditions,a servo drive fault diagnosis platform based on Zynq-7020 high performance SoC chip is built in the laboratory.The experiment shows that the detection of weak fault signal such as bearing in composite fault is vulnerable to the interference of installation fault,which will make the traditional diagnosis algorithm invalid.Aiming at the problem of feature extraction failure of bearing fault,the advantages and disadvantages of spectral kurtosis method and maximum correlation kurtosis deconvolution method are compared and analyzed by taking the composite fault of bearing outer ring fault and misalignment fault as an example.For the spectral kurtosis algorithm,the linear prediction model based on the minimum mean square error criterion is used to preprocess the composite fault signal,which improves the diagnosis effect of the spectral kurtosis algorithm.However,it can only prompt the frequency band with the largest kurtosis as the best filter band,and it is easy to ignore the sub resonance frequency band,resulting in false diagnosis.On the other hand,the maximum correlation kurtosis deconvolution algorithm is used to extract bearing fault features from bearing misalignment composite fault,but the effect of the algorithm is affected by deconvolution period and filter length.In order to improve the feature extraction effect of maximum correlation kurtosis deconvolution algorithm,firstly,the composite fault signal is preprocessed based on local mean decomposition to improve its signal-to-noise ratio.In order to improve the feature extraction effect of maximum correlation kurtosis deconvolution algorithm,firstly,the composite fault signal is preprocessed based on local mean decomposition to improve its signal-tonoise ratio.Aiming at the problem that it is easily affected by deconvolution period,filter length and other filter parameters,a deconvolution filter parameter optimization scheme based on PSO algorithm is proposed.Finally,based on the above research content,the bearing misalignment composite fault diagnosis and identification scheme is set,and the bearing misalignment composite fault diagnosis system software is designed,which lays the foundation for later engineering and production.