Research On Life Prediction Method Of Shearer Rocker Arm Based On Data Drive

As the main mechanical equipment for the safe and efficient production of modern mines,the shearer is a complex system that integrates machinery,electronic and electrical systems and hydraulic transmission systems into one.Nowadays,the research on shearers has attracted more and more attention.Nowadays,more and more attention is paid to the research of shearer.Among them,the rocker arm,as a direct responsibility to bear the heavy responsibility of cutting coal wall and power transmission,is also a fault-prone area of the shearer,so the life prediction research on the shearer rocker is of great practical value and engineering significance.This paper analyzes the drivetrain of the shearer swinging arm and its actual operating condition characteristics,and on this basis carries on the sensor point deployment,completes the feature extraction according to the collected signal,and finally an adaptive combined life prediction model of the rocker arm is established.First of all,the operation data of coal mine equipment is analyzed,the abnormal data cleaning model is obtained by improving K-means clustering algorithm,and the parallel design of the abnormal data cleaning model is realized by using the MapReduce computing framework of Hadoop platform.Finally,the data cleaning is completed to lay the foundation for the later feature extraction and prediction.Secondly,considering the complexity and nonlinearity of vibration signal,wavelet packet decomposition,EMD empirical mode decomposition,VMD variable mode decomposition and LMD local mean decomposition are carried out.The decomposed energy fraction is taken as a preliminary characteristic index,and then correlation analysis is carried out in combination with Time-domain and Frequency-domain characteristic indexes to achieve the preliminary index screening.Finally,PCA technology is used to complete the analysis The main metadata is selected as the characteristic index of vibration signal.Then,in view of the low reliability of single prediction model,ARIMA time series model and Elman neural network model are established respectively,and an adaptive combination prediction model method is proposed by analyzing the characteristics of fixed weight combination model and variable weight combination model,and the prediction accuracy of the combination prediction model is verified by experiments.Finally,combined with the characteristics of the transmission system of the shearer rocker arm and the actual working conditions,the structural model of the shearer rocker arm is obtained,and the weight of each signal of the structural model of the shearer rocker arm is determined by the principal component analysis method.On this basis,the life prediction of the rocker arm is realized by the adaptive combination prediction model.Finally,the reliability of the prediction model is verified by the case analysis.