Lifetime Prediction of a Proton Exchange Membrane Fuel Cell via an Improved TCN‐iTransformer Model

ABSTRACT Proton exchange membrane fuel cells (PEMFC) have been widely utilized in transportation and power generation due to their high efficiency and low pollution. However, their durability remains insufficient, and their output power decreases over time during operation. Therefore, it is important to predict the remaining useful life (RUL) of the PEMFC to ensure its efficient operation. In this paper, an improved TCN‐iTransformer model is proposed for predicting the RUL of PEMFC, which integrates temporal convolutional network (TCN), iTransformer, discrete cosine transform (DCT), and the channel attention mechanism. The reliability of the model was validated using both static and dynamic datasets of different lengths. And the results showed that the improved TCN‐iTransformer achieved a significant improvement over the Transformer prototype in long sequence time‐series forecasting. Furthermore, smaller mean absolute percentage error (MAPE) and root mean square error (RMSE) were obtained compared to other improved models, such as long short‐term memory (LSTM) and gated recurrent unit (GRU). In addition, the RUL prediction error of the model was found to not exceed 1 h.