A Novel Temporal Convolutional Network Based on Position Encoding for Remaining Useful Life Prediction

The technology of prognostics and health management (PHM) has developed rapidly. As one of the important tasks in PHM field, remaining useful life (RUL) prediction can effectively predict the remaining service time before machine failure, so that enterprises can make decisions in advance and avoid safety accidents. In this article, a new data-driven method is proposed, which adopts a position encoding scheme to extract more time sequence information from the original data, and then uses a novel temporal convolutional network (TCN) and attention mechanism to predict RUL. In order to evaluate the effect of the model, C-MAPSS dataset is used for testing the performance, and the results are compared with other methods, which shows that the proposed method is more effective.