PEMFC Lifetime Prediction Using TCN-BiLSTM Optimized by Osprey Optimization Algorithm

Proton Exchange Membrane Fuel Cell (PEMFC) play a crucial role in many fields. Accurate lifetime prediction has become a vital means to ensure the safe operation of key equipment. This paper proposes a prediction model based on the Osprey Optimization Algorithm (OOA) optimized Temporal Convolutional Network (TCN) and Bidirectional Long Short Term Memory Network (BiLSTM). The model captures long range dependencies in time series data through the dilated causal convolution of TCN, enhances context information representation using the bidirectional feature extraction capability of BiLSTM, and introduces the OOA optimization algorithm for global optimization of model hyperparameters. Simultaneously, the SG filter is used to preprocess the input data. The preprocessed data is then subjected to dimensionality reduction via the Spearman method to extract data strongly correlated with voltage as the optimal input, thereby enhancing model accuracy. Experimental results show that the proposed model achieves MAE, MAPE, RMSE, and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\mathbf{2}}$</tex> values of 0.00097, 0.0003, 0.00116, and 0.98921, respectively, demonstrating higher accuracy compared to TCN-BiLSTM and TCN-LSTM models.