Rolling Bearing Fault Diagnosis Under Various Working Conditions Based On Different Resonance Methods In Nonlinear System

Rolling bearing is the key mechanical component of rotating equipment,which is widely used in modern industrial fields such as manufacturing,medical care,transportation,aerospace and so on.However,the early faults of bearings(such as cracks,wear)will gradually deteriorate with the passage of time,affecting the operation of equipment,and even causing severe economic losses.Therefore,the fault diagnosis of rolling bearing is very important.At present,the bearing fault diagnosis technology based on vibration method is widely concerned,but these technologies are still limited to integer order system,and the research on vibration technology under fractional order system is relatively lack.In recent years,due to the development and progress of fractional calculus theory and computer technology,fractional system has shown more prominent advantages by virtue of its time memory and spatial correlation.In addition,these vibration techniques mainly focus on the problem of bearing fault feature extraction under constant conditions,ignoring the equally importance of variable conditions,so the bearing fault diagnosis technology of variable conditions based on vibration method needs to be studied urgently.Based on the fractional calculus theory and nonlinear vibration theory,this paper studies the problem of early weak fault feature extraction of rolling bearing under strong noise and constant(variable)working conditions.(1)Based on the fractional calculus theory,the fractional order nonlinear system resonance method is proposed.The fractional order system resonance is studied in detail from the perspective of theoretical analysis and numerical analysis.The fractional order system resonance excited by high frequency signal is realized based on the common variable scale theory,and it is applied to the feature extraction of bearing early weak fault under constant working condition.Simulation and experimental results show that the method is sensitive to periodic signals,which can not only greatly amplify the amplitude of fault frequency,but also suppress other frequency components.Compared with stochastic resonance method,the proposed method can directly enhance the weak characteristic signal only by the system itself,and will not increase the complexity of the system,so it has obvious advantages in enhancing the weak fault characteristic signal.(2)Considering the memory property of fractional derivative,the stochastic resonance model of fractional bistable system is established,and the fractional aperiodic stochastic resonance and fractional rescaled aperiodic stochastic resonance methods are proposed.The numerical simulation results show that the stochastic resonance effect of fractional system is better because of the optimization effect of fractional damping on stochastic resonance.In addition,based on the adaptive particle swarm optimization algorithm,an adaptive fractional stochastic resonance method is proposed to realize the early weak fault feature extraction of rolling bearing under certain conditions.The experimental results show that the fractional stochastic resonance can transform the noise energy to enhance the weak fault characteristics.In addition,the proposed method can adjust the parameters adaptively without relying on personal experience,which improves the processing efficiency of fractional stochastic resonance.(3)Based on the idea of segmentation and the common variable scale method,a piecewise rescaled stochastic resonance(PRSR)method is proposed,which makes the stochastic resonance method can be extended to the processing of frequency conversion signals(such as linear frequency modulation signals),and solves the problem that the existing stochastic resonance methods are difficult to process frequency modulation signals.The theoretical results show that this method can realize the stochastic resonance of the frequency conversion signal well,remove the noise interference and enhance the amplitude of the signal effectively.In addition,with the help of the advantages of fractional Fourier transform filtering algorithm in LFM signal extraction,combined with the piecewise rescaled stochastic resonance method,a FRFT-PRSR algorithm for bearing feature extraction under variable speed condition is proposed,which solves the problem of stochastic resonance enhancement and extraction of bearing early weak fault extraction under variable speed condition.By comparing the spectral kurtosis method,the FRFT-PRSR algorithm can remove noise interference,and can accurately extract and enhance the bearing fault features.To sum up,this paper makes an in-depth study on the fault feature extraction of rolling bearing under the fixed conditions with weak noise and strong noise,and the variable condition with strong noise.The simulation and experimental results show that the proposed method can effectively extract and enhance the fault feature components,and remove noise interference.This paper has 58 pictures,2 tables and 101 references.