Research On Multi-point Fault Diagnosis Of Rolling Bearing

Rolling bearing is one of the important parts of rotary machinery,it plays a role in supporting the rotation shaft and reducing friction.The working state of a rolling bearing is directly related to the normal operation of the rotating machinery.Therefore,it is of great significance to perform fault diagnosis or condition monitoring on rolling bearings.From the previous studies at home and abroad,most of them focused on the single-point fault diagnosis of inner ring or outer ring of rolling bearing.These methods can only judge whether the bearing fails,but not whether there are multiple fault points.Aiming at this problem,the theoretical model of bearing multi-point fault is established by taking double-point fault of bearing as an example.Through theoretical analysis,the difference of bearing vibration signal between single-point fault and double-point(or multi-point)fault in time domain is found.At the same time,the formula of impluse interval ratio in time domain of bearing double-point fault signal is proposed.In this paper,the principles of three de-noising methods for bearing vibration signals,namely SVD de-noising method,wavelet threshold de-noising method and ITD de-noising method,are studied in detail.And the validity of the three methods is verified by numerical simulation and experimental data analysis.Through theoretical analysis of resonance demodulation technology,it is found that the key factor determining the result of resonance demodulation is filtering.There are two problems in the filtering process: how to select the filter and how to determine the filter parameters.In order to solve these two problems,the theory of filter and spectral kurtosis is studied and introduced in detail.Based on the resonance demodulation technique and spectral kurtosis method,a bearing fault feature extraction method is proposed in this paper.Through the example analysis,the fault feature extraction method proposed in this paper is combined with the ITD de-noising method to successfully extract the fault characteristics from the bearing vibration signal,and verify the effectiveness of this fault feature extraction method.This paper proposes a fast and efficient method for bearing fault diagnosis.The basic idea of this method is: Taking time domain waveform analysis as the time domain analysis method,spectral kurtosis method and resonance demodulation technology as the frequency domain analysis method,the vibration signal of the bearing is analyzed in both time domain and frequency domain.Based on the frequency domain fault characteristics of the bearing,the envelope spectrum can be used to determine whether the bearing has a fault and locate the fault position.Based on the time-domain difference between two-point(or multi-point)fault and single-point fault of the bearing,the time-domain waveform can be used to judge whether there is a double-point(or multi-point)fault of the bearing and even calculate the included Angle of the fault point.In addition,different de-noising methods can be selected for pre-processing(or no de-noising processing)according to actual conditions.In order to verify the validity and correctness of this method,the bearings of five failure modes were tested and the detailed data analysis was carried out.Based on resonance demodulation technology and time domain feature extraction method,a rolling bearing condition monitoring system was developed on Labview platform.This system has real-time status monitoring function,alarm function and historical data query function.