A LSTM-MSGNET Model for Remaining Useful Life of Lubricating Oil Prediction

Lubricating oil plays a significant role in the operation of mechanical equipment, as it reduces wear and tear and maintains the effective lifespan of the equipment. In actual production processes, the frequency of oil changes is very important. If the interval between changes is too long, it can cause equipment damage; if it is too short, it leads to wastage of lubricating oil. However, due to the complexity of the working environment of mechanical equipment, the degradation of lubricating oil is irregular, making it difficult to determine an appropriate replacement interval. The degradation of lubricating oil is a long-term process accompanied by nonlinear changes in various parameter indicators. This study employs the LSTM-MSGNet time series prediction method to predict the remaining useful life based on various parameters of lubricating oil collected in the laboratory and test results. LSTM can effectively capture long-term dependencies, and MSGNet can analyze the changes in multi-variable, multi-scale data. Experiments have shown that the LSTM-MSGNet model can achieve good results in predicting the remaining useful life of lubricating oil and can be applied to subsequent online monitoring of lubricating oil quality.