Research On Remaining Useful Life Prediction And Remanufacturability Of Crusher Roller Sleeve Based On State Monitoring Data Fusion

Due to the special working conditions of large mechanical equipment,it poses a severe challenge to its health management.How to accurately and effectively predict the remaining useful life of key components and how to remanufacture them are becoming a hot topic of enterprises.This paper aiming at two major problems which are urgently needed to be solved in the field of health management of large machinery and equipment—remaining useful life prediction and remanufacturability evaluation.The algorithm of predicting remaining useful life and the method of evaluating remanufacturability coefficient about large-scale components based on online monitoring data fusion under complex working conditions are systematically studied.Firstly,this paper studied the status of common remaining useful life prediction and remanufacturing method,data fusion algorithm and support vector machine model.Through analysis and comparison,the prediction method of remaining useful life for large parts is put forward,and the theoretical foundation of health management diagnosis is also laid.Secondly,the roll sleeve of the large crusher is chosen as the research object to operate accelerated life test on the PRONOSTIA test platform.The raw data information of acceleration and temperature are acquired by sensors.Through analysing the time domain feature,frequency domain feature and time-frequency feature to extract the character signal which can reflect the health status of the equipment.And the character signal is pretreated by a series of pretrements such as removing the singular value,removing the trend item,and then the wavelet technique is used to denoise the original signal.Thirdly,construct the remaining useful life prediction model for roll sleeve.The acceleration signal and temperature signal are integrated by least squares and D-S data fusion.Using the two training sets to train the prediction model,and then optimize the prediction model through grid search algorithm and cross validation.Finally the prediction results were obtained by the optimized support vector regression machine.At the same time,compared the prediction results of this paper's method with the prediction results obtained by using only a single data source and other prediction methods.According to the roller sleeve tests were carried out under different conditions,results prove the correctness and reliability of the algorithm proposed in this paper.Lastly,this paper analyzed the remanufactureability problems in roll sleeve's health management.According to the current product remanufacturability evaluation criteria,a remanufacturability coefficient evaluation model which is more suitable for this paper's research object is designed,and the specific algorithm of each index in the model is determined.The model comprehensively evaluated the remanufacturability of roll sleeve from technology,economy and environment index,formed a complete remanufacturability coefficient evaluation method,and the validity and practicability of the method were proved by case study.