The life Prediction of PEMFC based on Group Method of Data handling with Savitzky–Golay Smoothing. (December 2022)

• Record Type: Journal Article
• Authors: Liu, Jiawei; Li, Ting; Tang, Quan; Wang, Yunling; Su, Yunche; Gou, Jing; Zhang, Qiao; Du, Xinwei; Yuan, Chuan; Li, Bo
• Is Part Of: Energy reports. Volume 8(2022)Supplement 16
• Page Start: 565
• Page End: 573
• Publication Date: 2022-12
• Subjects: Proton exchange membrane fuel cells -- Aging prediction -- Savitzky–Golay Smoothing -- Group Method of Data; Power resources -- Periodicals; Energy industries -- Periodicals; Power resources; Periodicals; Electronic journals; 621.04205
• DOI: 10.1016/j.egyr.2022.10.256
• Languages: English
• ISSNs: 2352-4847

Abstract

Abstract: To solve the problem of inaccurate prediction of the stack life of proton exchange membrane fuel cells, this paper first proposed a fuel-cell aging prediction method based on method Savitzky–Golay Smoothing and Group Method of Data, which was based on the data drive. Savitzky–Golay Smoothing is an optimal piecewise fitting method based on polynomial in the time domain and using the least square method through moving window, which is widely used in data flow smoothing and denoising. Group Method of Data is a modeling method of the complex nonlinear dynamic system. The inner criterion and outer criterion are used in the training set and test set respectively, and the optimal solution is finally solved by iterative screening. The method presented in this paper was verified by 1020 h fuel cell aging experiment. The experimental results show that: the MSE, RMSE, R of the test data are respectively 6.3935e−05, 0.0079959, and 0.99616. The data-driven prediction method proposed in this paper can be used for fuel cell aging prediction and fault warning.