A Review of Artificial Neural Network Applications in Proton Exchange Membrane Fuel Cells: Performance Prediction, Fault Diagnosis, and Degradation Modeling

Proton exchange membrane fuel cells (PEMFCs) are promising clean energy technology for vehicles. However, their performance and durability depend on complex physicochemical processes, which complicates accurate modeling. Artificial neural network (ANN) provides a powerful data‐driven approach to capture the nonlinear and multidimensional behavior of PEMFCs. This review systematically analyzes and quantitatively compares over 100 studies, assessing the predictive accuracy, measured by metrics such as root mean square error and coefficient of determination ( R 2 ), along with the limitations of various ANN architectures, feedforward neural network, deep neural network, convolutional neural network, recurrent neural network, and hybrid models. Key application areas covered are performance prediction, fault diagnosis, and degradation modeling. Finally, future research directions are highlighted, focusing on the development of more comprehensive models, advanced optimization techniques, and improved interpretability to further enhance the role of ANNs in advancing PEMFCs technology.