Research On Joint Decision Of Condition-based Maintenance And Spare Parts Inventory For Rolling Bearing Based On Degradation Analysis

Rolling bearing is the core component of the mechanical transmission system of conveyance of all kinds.Its accidental failure will directly affect the transmission efficiency of the whole transmission system,and even cause severe safety accidents.Therefore,the research on rolling bearing condition assessment and maintenance strategy is of great engineering significance and application value for improving its operation reliability and reducing production cost.Based on the degradation data of rolling bearing,this research focuses on the weak fault feature extraction method,compound fault feature extraction method,the joint decision-making method considering condition-based maintenance and spare parts inventory.The main research contents are as follows:(1)The weak fault features of rolling bearings often present a non-stationary signal and weak regulation.Aiming at the problem that the weak fault features are difficult to extract under strong background noise,an adaptive VMD parameter optimization method based on genetic algorithm is proposed(namely,GA-VMD).The proposed method takes the weighted envelope entropy constructed by the envelope entropy and kurtosis as the fitness function of genetic algorithm,and reasonably matches the crossover probability,mutation probability and the optimization range of parameters according to the individual fitness function value and local optimal results.Based on two sets of test data,it is verified that the constructed weighted envelope entropy is more sensitive to the impulsive and periodic components in the signal.This method can realize the adaptive extraction of weak fault features of rolling bearing earlier.(2)Different from the single fault feature extraction of rolling bearing,its compound fault features are affected by the mutual interference of different fault sources and transmission path,which makes it laborious to accurately evaluate its operation status.Therefore,an adaptive method of extracting the compound fault of rolling bearing based on VMD and MCKD is proposed.Firstly,the signal is decomposed by VMD for reducing the noise as preprocessing.Secondly,the optimal IMF is used as the input signal of MCKD,and the bearing fault type is diagnosed through the mean maximum correlated kurtosis value,so as to adaptively determine the two key parameters of filter length and deconvolution period.Finally,the compound fault features are identified by envelope spectrum analysis.The test signal analysis results based on single fault and compound fault of rolling bearing show that this method can effectively determine the fault type of unknown fault signal,and effectively separate the compound fault features under the strong noise.Furthermore,this method enhances the diagnosis performance of single weak fault feature of adaptive VMD.(3)The degradation feature monitoring method based on rolling bearing vibration signal can not accurately reflect its operation conditions.Firstly,an operation condition monitoring method considering both degradation feature and fault feature is proposed,and both of the weak fault feature extraction and compound fault feature extraction method in the above two chapters are used to identify the fault features of the degradation feature mutation points in the condition monitoring process,and the three-stage delay time model of rolling bearing is established by combining the identification results and delay time theory.Secondly,the proportional hazards model based on Weibull distribution is used to describe the reliability evolution process,and a joint decision-making model with reliability threshold as spare parts ordering criteria and maintenance cost rate as objective function is established.Finally,two sets of accelerated life test data of rolling bearing are used to verify the effectiveness of the operation condition assement method and joint decision-making model.Compared with the traditional joint decision-making model with constant failure risk,the replacement time of rolling bearing in the proposed method can better adapt to the ordering of spare parts,and obtain lower maintenance cost rate.(4)In order to solve the joint decision problem of different levels of imperfect preventive maintenance and differential batch consumption of spare parts under competing failure,a joint decision model considering the bi-level imperfect preventive maintenance and the improved(s,S)spare parts inventory strategy is proposed.Firstly,the virtual service age reduction method considering bi-level imperfect preventive maintenance under competing failure is established.Secondly,an improved(s,S)spare parts inventory strategy is proposed using the information of remaining service life and the accumulated consumption of spare parts.Finally,the joint optimization model is solved with the maintenance interval and the maximum inventory of spare parts as the decision variables.Compared with the traditional(s,S)spare parts inventory strategy,the proposed improved strategy can effectively reduce the redundancy of spare parts inventory,and keep the equipment at a higher reliability level while obtaining lower maintenance cost.The technology routes of this research gradually extend from fault diagnosis and degradation monitoring to condition-based maintenance and spare parts management,forming a systematic decision-making framework for condition-based maintenance and spare parts inventory of rolling bearing.The research contents realize the effective integration of fault diagnosis technology and reliability analysis technology,promotes their application in the actual process of the operation and maintenance,and can provide some theoretical reference and technical support for improving the level of enterprise equipment health management.