Research of proton exchange membrane fuel cell decay prediction based on voltage characterization

Proton exchange membrane fuel cells (PEMFCs) represent a pivotal technology in the realm of new energy vehicles and energy storage, yet they continue to grapple with the formidable challenge of predicting performance degradation. This underscores the critical need for research focusing on life prediction. In this paper, we put forward a novel prediction method employing radial basis function (RBF) theory to address the issue of battery degradation prediction. The approach involves the utilisation of voltage characteristics, recognised as a pivotal indicator of battery life, complemented by a comprehensive analysis of the model's prediction capability in terms of error, correlation and stability. The resulting RMSE, MAE and MAPE values for the predicted degradation voltages are 0.00351, 0.0028, and 0.083%, respectively. The RBF model has been shown to possess both strong generalization ability and adaptability to data fluctuation, thus making it a suitable candidate for predicting the degradation voltage of the PEMFCs power stack.