Prediction of Remaining Bearing Life Based on 1D-MSCNN
Bearings are the basic components of rotating machinery and equipment,and their status is closely related to the safe operation of the equipment.Accurately predicting the remaining service life of the bearing can improve the reliability of the equipment,and at the same time provide a practical reference for equipment maintenance.Convolutional Neural Network(CNN)has achieved certain results in RUL prediction due to its strong learning ability.However,the fixed-size convolution kernel of the traditional convolutional neural network is difficult to learn the local and global features of complex signals.Therefore,this paper proposes a one-dimensional multiscale convolutional neural network(1D-Multi Scale Convolutional Neural Network,1D-MSCNN)prediction model to achieve accurate estimation of bearing degradation state.First,a degenerate function is used to establish an ef-fective health index(HI)of the bearing.Secondly,the multi-scale convolution structure is used to fully extract the deep representa-tive features of the original data.Finally,the effectiveness of the proposed 1D-MSCNN model is verified through a case study on the PRONOSTIA data set,and compared with other prediction methods to verify the effectiveness and superiority of this method.
国家科技期刊平台
NSTL
万方数据
