Study On Fault Prediction Technology Of Rolling Bearing Based On Extreme Learning Machine

According to the statistics of oil field data,nearly 80% of the motor faults were caused by rolling bearing.As an important component in rotating machinery,rolling bearing's conditions will directly decide mechanical equipment's normal operation or not,even cause serious consequences.The propane compressor working in Hade operation district of Tarim oilfield is an important equipment to the process of production.Bearing malfunctions occurring in the side of motor's fan frequently had produced a huge effect on oil field production.The traditional diagnosis methods cannot provide an accurate evaluation for the motor' future operating state,which is bad for raising the level of oilfield equipment maintenance and management.Based on the problems above,this article focus on several aspects researching as following:(1)The vibration index prediction model based on Extreme Learning Machine is proposed.This article analyzes the advantages and disadvantages of the commonly used prediction models and builds a prediction model based on ELM to forecast the variation tendency of the RMS values of the rolling bearing.The forecast error between the actual and estimated values of the results is smaller than the predicting outcomes of BP and SVM,which explains that the prediction model has outstanding ability of fitting and prediction rolling bearing vibration parameters.(2)Vibration index evaluation criteria of rolling bearing is established.Totally,16 indexes are selected as the rolling bearing evaluation indexes for vibration including time-domain,frequency-domain and wavelet packet decomposition energy.Each index' evaluation criteria is formulated by referring to its historical vibration data.The diagnosis result by using the vibration indexes evaluation standard is the same as the actual fault results,which proved that the vibration index evaluation standard established above is practical.(3)The application study on fault prediction technology of rolling bearing based on ELM.The fault prediction result of the prediction model based on ELM and the vibration indexes evaluation standard is basically consistent with the actual fault result.Therefore,the fault prediction technology based on ELM can forecast and evaluate the future operating state to a certain extent and provide effective support for oilfield equipment maintenance and management.