LSTM-Based Multi-Task Method for Remaining Useful Life Prediction under Corrupted Sensor Data

Data-driven remaining useful life (RUL) prediction plays a vital role in modern industries. However, unpredictable corruption may occur the collected sensor data due to various disturbances real industrial conditions. To achieve better RUL performance under this situation, we propose novel multi-task method for prediction, which is named deep long short-term memory (MTD-LSTM). In MTD-LSTM, convolutional neural network (CNN) and (LSTM) are first employed feature extraction fusion. Then, extracted features fed into learning module, contains missing value imputation module. The values task performed simultaneously. purpose of obtain integral degradation information by recovering complete data; thus, performs corrupted data. addition, loss term proposed smooth results without any manual post-processing. effectiveness verified on simulated dataset based C-MAPSS dataset.