Deep learning-based fault diagnosis of high-power PEMFCs with ammonia-based hydrogen sources

• Published in: Journal of power sources 2025-02, Vol.629, p.236018, Article 236018
• Main Authors: Chen, Zhang-Liang, Zhang, Ben-Xi, Zhang, Cong-Lei, Xu, Jiang-Hai, Zheng, Xiu-Yan, Zhu, Kai-Qi, Wang, Yu-Lin, Bo, Zheng, Yang, Yan-Ru, Wang, Xiao-Dong
• Format: Article
• Language: English
• Subjects: Ammonia-based hydrogen source; Fault diagnosis; Machine learning method; PEMFC; t-SNE-CNN-HMM model
• ISSN: 0378-7753
• DOI: 10.1016/j.jpowsour.2024.236018

The fault diagnosis of high-power proton exchange membrane fuel cells (PEMFCs) with ammonia-based hydrogen sources (AHSs) is studied by an ensemble machine learning methods as considering various operating conditions. Under various operating conditions, the results show that when the single-fault state and the multiple-faults state, appear in a single high-power PEMFC system, the overall diagnostic accuracy of 98.89 % is realized by the optimized hidden Markov model (HMM) coupling with the convolutional neural network (CNN) and the t-distributed stochastic neighbor embedding (t-SNE). When using the t-SNE-CNN-HMM fault diagnosis model, the overall diagnostic accuracy is 100 % in a single AHS system. Similarly, the overall diagnostic accuracy is 98.58 % for the single-fault and the multiple-faults states in the AHSs-PEMFCs coupled system. Based on these single-fault and multiple-faults states, the overall diagnostic accuracy of t-SNE-CNN-HMM model is larger in comparison with these of other diagnosis models, of which these diagnosis models are composed of the support vector machine (SVM), backpropagation neural network (BPNN) and only-CNN models. The t-SNE-CNN-HMM method integrates the generalization and identification ability of CNN and HMM, leading to provide an efficient and accurate fault diagnosis for the AHSs-PEMFCs coupled system. •Fault diagnosis is implemented under various operating conditions.•The fault diagnosis is realized based on an ensemble machine learning method.•The diagnostic accuracy of 98.89 % is realized for single or multiple-faults states.