Research On Early Fault Diagnosis And Prediction Method Of Rolling Bearing

With the progress of science and technology,rotary machinery is developing towards the direction of heavy load,high speed and precision.Rolling bearing is an important part of rotating machinery,and its health directly affects the performance of the equipment,the entire unit and the production line,and even affects personal safety.If bearing early failure occurs,it can be found and intervened,which can reduce the occurrence of accidents and prolong the service life of equipment.Therefore,the research on early fault diagnosis and prediction method of rolling bearings is of great practical significance.The rolling bearing is taken as the research objective,the feature extraction technology of vibration signal is analyzed,and novel methods for fault diagnosis and prediction are presented respectively in this thesis.The main work contents are as follows:(1)Aiming at the problem that it is difficult to detect early weak fault of rolling bearing,an improved Maximum Correlation Kurtosis Decon volution(MCKD)algorithm is proposed to diagnose bearing faults.Firstly,the influence of MCKD parameters on the denoised effect is studied,and considering the lack of unified measurement index in the selection of three parameters of MCKD,a new index-Envelope Spectrum Multi-Pulse(ESMPF)is defined.Secondly,the ESMPF is taken as the evaluation index of parameters in MCKD algorithm,and the optimal parameters of the deconvolution algorithm are selected.Then,the original vibration signal is denoised by using the optimized MCKD algorithm,and the denoised signal is analyzed by envelope spectrum.Finally,the proposed method is verified by actual engineering signals,and the results show that the proposed method can enhance the bearing early weak fault signal and extract the bearing fault feature frequency effectively for the early weak fault diagnosis of bearing.(2)Due to the strong correlation among the three parameters of MCKD method,and low efficiency and error-prone of parameters selection by trial method,To solve these problems,an adaptive parallel optimization of MCKD parameters using artificial fish swarm algorithm is proposed,and the optimized algorithm is used to diagnose early bearing faults.Firstly,the ESMPF is introduced into the artificial fish swarm algorithm,which is taken as the adaptive function of the artificial fish swarm algorithm,and the artificial fish swarm algorithm is adopted to optimize adaptively the parameters of MCKD in parallel.Secondly,the optimized MCKD algorithm is used to denoise the original signals,and the envelope spectrum of the denoised signals is obtained.Finally,the experimental signals are used for comparative analysis,and the results show that the method can realize the adaptive and accurate diagnosis of early weak fault of rolling bearing.(3)A combined model of sample entropy and Variational Mode Decomposition(VMD)is put forward to predict the early faults of rolling bearings.Firstly,the sample entropy value of rolling bearing vibration signal is calculated to form the time series of the bearing health state.Secondly,the decomposition layers K of VMD is taken as an integer value from 2 to 10 respectively,and the time series of bearing health state is decomposed by VMD,and different IMF components are obtained under different K value decompositions.Then,according to the correlation,kurtosis variance and Euclidean distance criteria,the optimal trend term of bearing vibration signal is found.Finally,the proposed model is verified by engineering signal and compared with other indicators,and the results show that the prediction model can not only monitor the running state of bearings,extract the trend t erm of bearing vibration signal,but also predict the early weak faults of bearing earlier.