Abstract:In the context of predicting the remaining useful life (RUL) of rolling bearings, many models often encounter challenges in identifying the starting point of the degradation stage, and the accuracy of predictions is not high. Accordingly, this paper proposes a technique that utilizes particle swarm optimization (PSO) in combination with the fusing of a one-dimensional convolutional neural network (CNN) and a multihead self-attention (MHSA) bidirectional long short-term memory (BiLSTM) network called PSO-CNN-BiLSTM-MHSA. Initially, the original signals undergo correlation signal processing to calculate the features, such as standard deviation, variance, and kurtosis, to help identify the beginning location of the rolling bearing degradation stage. A new dataset is constructed with similar degradation trend features. Subsequently, the particle swarm optimization (PSO) algorithm is employed to find the optimal values of important hyperparameters in the model. Then, a convolutional neural network (CNN) is utilized to extract the deterioration features of rolling bearings in order to predict their remaining lifespan. The degradation features are inputted into the BiLSTM-MHSA network to facilitate the learning process and estimate the remaining lifespan of rolling bearings. Finally, the degradation features are converted to the remaining usable life (RUL) via the fully connected layer. The XJTU-SY rolling bearing accelerated life experimental dataset was used to verify the effectiveness of the proposed method by k-fold cross-validation. After comparing our model to the CNN-LSTM network model and other models, we found that our model can achieve reductions in mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE) of 9.27%, 6.76%, and 2.35%, respectively. Therefore, the experimental results demonstrate the model's accuracy in forecasting remaining lifetime and support its ability to forecast breakdowns.

What problem does this paper attempt to address?

The paper aims to address the challenges in predicting the Remaining Useful Life (RUL) of rolling bearings, particularly the issue of identifying the onset of the degradation phase of rolling bearings and improving prediction accuracy. Specifically, the paper proposes a method called the PSO-CNN-BiLSTM-MHSA model, which combines a one-dimensional Convolutional Neural Network (CNN) optimized by the Particle Swarm Optimization (PSO) algorithm with a Bidirectional Long Short-Term Memory network with Multihead Self-Attention (BiLSTM-MHSA). Through this method, the paper aims to achieve the following goals: 1. **Identify the onset of degradation**: Use signal processing to calculate features such as standard deviation, variance, and kurtosis to help identify the onset of the degradation phase of rolling bearings. 2. **Optimize hyperparameters**: Use the Particle Swarm Optimization algorithm to find the optimal values of important hyperparameters in the model. 3. **Feature extraction and prediction**: Use a one-dimensional Convolutional Neural Network (1D-CNN) to extract degradation features of rolling bearings and input them into the BiLSTM-MHSA network for learning and remaining life prediction. 4. **Improve prediction accuracy**: Experimental results show that the model reduces the Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE) by 9.27%, 6.76%, and 2.35%, respectively, when predicting the remaining life. Through these improvements, the proposed method in the paper demonstrates higher accuracy in predicting the remaining life of rolling bearings and helps prevent the occurrence of failures.

Rolling Bearing Residual Useful Life Prediction Model Based on the Particle Swarm Optimization-Optimized Fusion of Convolutional Neural Network and Bidirectional Long–Short-Term Memory–Multihead Self-Attention

A Hybrid Model-based Prognostics Approach for Estimating Remaining Useful Life of Rolling Bearings

A Hybrid ST-Bilstm Based Prediction Method for Estimating the Life of Ball Mill Rolling Bearings

Bearing Remaining Useful Life Prediction by Combining CNN with PSO_LSSVM

A RUL prediction method of rolling bearings based on degradation detection and deep BiLSTM

A Remaining Useful Life Prediction Method of Mechanical Equipment Based on Particle Swarm Optimization-Convolutional Neural Network-Bidirectional Long Short-Term Memory

Remaining useful life prediction of rolling bearings based on CNN-GRU-MSA with multi-channel feature fusion

A New Method of Bearing Remaining Useful Life Based on Life Evolution and SE-ConvLSTM Neural Network

A CNN‐BiLSTM‐Bootstrap integrated method for remaining useful life prediction of rolling bearings

A Hybrid Deep Learning Prediction Method of Remaining Useful Life for Rolling Bearings Using Multiscale Stacking Deep Residual Shrinkage Network

Remaining Useful Life Prediction of Rolling Bearings Based on Recurrent Neural Network

Prediction of Remaining Service Life of Rolling Bearings Based on Convolutional and Bidirectional Long- and Short-Term Memory Neural Networks

Remaining Useful Life Prediction of Rolling Bearings Based on Multi-scale Permutation Entropy and ISSA-LSTM

Data and Model Synergy-Driven Rolling Bearings Remaining Useful Life Prediction Approach Based on Deep Neural Network and Wiener Process

A Method for Predicting the Remaining Life of Rolling Bearings Based on Multi-Scale Feature Extraction and Attention Mechanism

Residual Life Prediction of Rolling Bearings Based on a CEEMDAN Algorithm Fused with CNN–Attention-Based Bidirectional LSTM Modeling

Remaining useful life prediction of rolling bearings based on Bayesian neural network and uncertainty quantification

A Remaining Useful Life Prediction Method Based on Time–frequency Images of the Mechanical Vibration Signals

Remaining Useful Life Prediction of Rolling Bearings Based on TCN-MSA

An entire life-cycle rolling bearing remaining useful life prediction method using new degradation feature evaluation indicators

Multi-feature Based LSTM Rolling Bearing Remaining Useful Life Prediction