A life prediction method of rolling bearing based on signal reconstruction and fusion dual channel network

Abstract

In addressing the problem of low prediction accuracy in the Remaining Useful Life (RUL) prediction of rolling bearings, caused by noise interference and insufficient extraction of sensitive features by deep learning models, this paper presents a life prediction method based on signal reconstruction and dual-channel network fusion. First, addressing the issue of extracting weak features from rolling bearing vibration signals, an optimized combination of Variational Mode Decomposition (VMD) and Teager-Kaiser Energy Operator (TKEO) for signal reconstruction is proposed. TKEO is used to track the energy of high-frequency impulses in the original signal. The Sparrow Search Algorithm (SSA) is employed for optimizing VMD to perform high-frequency filtering, resulting in an optimized impulse energy signal. This signal is then multiplied with the original signal to enhance weak impulse features under noisy background. Next, a dual-channel network model for Remaining Useful Life (RUL) prediction is constructed based on Temporal Convolutional Network (TCN) and Convolutional Neural Networks (CNN). The one-dimensional time-series signal obtained after reconstruction and denoising serves as the input for the TCN network channel. Simultaneously, the signal obtained after reconstruction and denoising undergoes Synchrosqueezed Wavelet Transforms (SWT) to generate a two-dimensional time-frequency representation as input for the CNN network channel. This setup allows TCN and CNN to respectively extract temporal features from the vibration signal and time-frequency features from the spectrogram. By incorporating self-attention mechanisms, internal correlations between different features are fully explored, thereby enhancing prediction accuracy.

Ultra-long-term spatiotemporal feature fusion network for bearing remaining life prediction under strong interference

“…However, methods based on deep CNN struggle to balance local and global considerations during feature extraction process and lack the ability to model long-term dependencies, resulting in features that inadequately capture the long-term degradation trends of bearings. Li et al [14] constructed a dualchannel RUL prediction network based on TCN and CNN, the one-dimensional time series signal is input to TCN, and the two-dimensional time-frequency diagram is input to CNN. The time and frequency characteristics of the vibration signal are extracted through TCN and CNN respectively, thereby improving the prediction accuracy.…”

Life Prediction of Rolling Bearing Based on Optimal Time–Frequency Spectrum and DenseNet-ALSTM

“…The ratio of the duration between the moment corresponding to the i th spectrogram and the bearing failure time to the overall duration from the beginning of the bearing to the bearing failure time is referred to as the remaining useful life of the bearing ( y i ). The formula for y i is given by [ 28 , 29 , 30 ] where n represents the total number of time–frequency spectra in this dataset, and i is the label or index of the current moment’s time–frequency spectrum.…”

“…The ratio of the duration between the moment corresponding to the ith spectrogram and the bearing failure time to the overall duration from the beginning of the bearing to the bearing failure time is referred to as the remaining useful life of the bearing (y i ). The formula for y i is given by [28][29][30]…”

Rolling bearing degradation trend prediction based on composite multiscale grey entropy and dynamic particle filter

“…better comfortableness and good stabilization. The achievement of these demands is highly dependent on the operating condition of rolling bearing [2]. Consequently, rolling bearing prognostics and health management (PHM) research have received increasing attention in recent years.…”