基于改进CEEMDAN方法的PEMFC氢气泄漏故障诊断
Hydrogen Leakage Fault Diagnosis of PEMFC Based on the Improved CEEMDAN Method

本文提出了一种针对质子交换膜燃料电池(PEMFC)氢气泄漏故障的诊断方法，通过结合自适应噪声全集经验模态分解(CEEMDAN)和自适应阈值去噪(ATD)技术，有效解决了信号噪声干扰导致的故障特征提取困难问题。具体步骤如下：首先，利用CEEMDAN对采集的振动信号进行分解；接着，采用灰色关联分析(GRA)筛选出噪声主导和信号主导的分量；随后，对噪声主导分量进行ATD处理，并将其与信号主导分量进行重构；最后，通过Teager能谱分析实现氢气泄漏故障的准确识别。为验证该方法的有效性，采用机理模型生成的PEMFC数据进行测试，并与CEEMD-ATD和CEEMDAN结合小波阈值去噪的方法进行对比。实验结果表明，所提方法在自适应性和去噪性能上具有显著优势，能够有效提升PEMFC氢气泄漏故障的诊断精度。
This paper proposes a diagnostic method for hydrogen leak faults in Proton Exchange Membrane Fuel Cells (PEMFC) by combining Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Adaptive Threshold Denoising (ATD) techniques, effectively addressing the challenge of extracting fault features hindered by signal noise interference. The specific steps are as follows: Firstly, CEEMDAN is used to decompose the collected vibration signals; then, Grey Relation Analysis (GRA) is employed to screen out noise-dominant and signal-dominant components; subsequently, ATD is applied to the noise-dominant components, which are then reconstructed with the signal-dominant components; finally, accurate identification of hydrogen leak faults is achieved through Teager Energy Spectrum Analysis. To verify the effectiveness of the proposed method, PEMFC data generated by a mechanistic model is used for testing, and the results are compared with those obtained using CEEMD-ATD and CEEMDAN combined with wavelet threshold denoising. Experimental results demonstrate that the proposed method has significant advantages in adaptability and denoising performance, effectively improving the diagnostic accuracy of hydrogen leak faults in PEMFC.