Method of identifying bearing fault based on serial EEMD and LSTM
As the core component of the train running part,the working state of the rolling bearing directly determines the safety of the whole train.The abnormal vibration signals of vehicle bearings are mostly non-stationary nonlinear signals.Aiming at the limitation of traditional time-frequency analysis methods,a fault identification model based on ensemble empirical mode de-composition(EEMD)and long short-term memory neural network(LSTM)is proposed in this paper.An improved ensemble em-pirical mode decomposition(EEMD)method is proposed to solve the mode aliasing problem of EMD in traditional HHT.It can decompose the original vibration data effectively and remove the trend component by correlation coefficient method,so as to i-dentify bearing faults better and predict bearing faults effectively.The wavelet threshold method is used to denoise the high fre-quency and low frequency components,and the weighted reconstruction of the high frequency and low frequency information components after denoising is carried out.The Hilbert-Huang transform is used to optimize the processing flow,and the rela-tionship between time,instantaneous frequency and instantaneous energy is calculated.Hilbert spectrum is input into long short-term memory neural network(LSTM)to extract features and judge the fault mode of vehicle bearings.The experimental results show that the model can effectively extract the characteristics of the vibration of rail vehicle bearings and diagnose the fault forms with high confidence.
NETL
NSTL
万方数据
