Evaluation Of Wet Clutch Wear State And Prediction Of Remaining Life

The stable operation of the wet clutch for the normal operation of the vehicle or heavy machinery plays a very important role,in advance to state evaluation,wear trend prediction and residual life prediction,can be targeted for maintenance,not only can reduce the maintenance cost and also to efficient operation of the protection device,to avoid heavy losses caused by sudden fault is of great significance.In this paper,according to the research objectives of wet clutch wear state evaluation and residual life prediction,appropriate equipment is selected,a wet clutch wear test bench is built,the test scheme is formulated,the accelerated degradation experiment of wet clutch wear is completed,and the oil sample containing the clutch wear data is obtained.By comparing several oil analysis methods,such as spectral analysis,ferrography analysis and particle size analysis,atomic spectroscopic technology was selected to analyze clutch oil samples and obtain the spectral data of worn oil.Then,the Cu and Pb elements with the greatest correlation with clutch wear were selected as the optimal characterization parameters by using correlation analysis method.After that,data were preprocessed,including oil filling correction and error point elimination and correction,to improve the accuracy of data.Secondly,the least square method is used to predict the wear trend of wet clutch.A total of 50 groups of data in five stages are predicted.The maximum deviation rate between the predicted results and the actual results is 14.45%,and the prediction accuracy rate is above85%.Evaluation process,then the fuzzy comprehensive evaluation method based on the main structure,the working properties of the wet clutch,select the macro（sound）with incomplete,insufficient power transfer,separation and micro elements（Cu,Pb element）to establish wear conditions of a two-stage evaluation structure,determine the clutch wear evaluation sets running-in wear,normal wear and tear,excessive wear,Fuzzy statistical method was used to analyze the weight of macroscopic factors,entropy method was used to analyze the weight of microscopic factors,and expert experience method and assignment method were used to obtain the membership function of each wear state grade,establish the evaluation matrix,and finally establish the state evaluation model.A total of 30 groups of test data of four groups of oil samples were evaluated and verified,and the accuracy reached 86% when compared with the actual stateFinally,three random processes,namely Wiener process,gamma process and Inverse Gauss process,are analyzed.Through comparison,it is found that the wear degradation data of wet clutch conforms to Wiener process,so binary Wiener is used to predict its remaining life;The optimal characterization parameters Cu and Pb are selected as the indicator elements of the remaining life of the wet clutch;Then,according to Akaike information criterion,the appropriate copula function is selected to obtain the correlation function of Cu and Pb elements;Then the parameters of Copula function are obtained by maximum likelihood method θ,The joint probability function of residual life of wet clutch is established.The binary Wiener model is used to predict the residual life of each process with running time of 30,60,90,120,150,180,210 and 240 h.Compared with the real residual life,it is found that the deviation of the residual life prediction of binary Wiener process is less than 25% in the range of 150h-240 h.The results show that the prediction model established by the binary Wiener process has the advantages of high prediction accuracy and real-time monitoring,and can be extended to the aspects of equipment health management.