Efficient model selection for real-time adaptive cold start strategy of a fuel cell system on vehicular applications. (31st July 2020)

• Record Type: Journal Article
• Authors: Amamou, A.; Kandidayeni, M.; Macias, A.; Boulon, L.; Kelouwani, S.
• Is Part Of: International journal of hydrogen energy. Volume 45:Number 38(2020)
• Page Start: 19664
• Page End: 19675
• Publication Date: 2020-07-31
• Subjects: Cold start -- Experimental approach -- Online identification -- Proton exchange membrane fuel cell -- Semi-empirical model; Hydrogen as fuel -- Periodicals; Hydrogène (Combustible) -- Périodiques; Hydrogen as fuel; Periodicals; 665.81
• DOI: 10.1016/j.ijhydene.2020.04.253
• Languages: English
• ISSNs: 0360-3199

Abstract

Abstract: The PEMFC maximum power is greatly influenced by subfreezing temperature and degradation phenomena. Therefore, a dependable model is required to estimate the power with respect to the variation of the operating conditions and state of health. Semi-empirical models are potent tools in this regard. Nonetheless, there is not much information about their cold environment reliability. This paper comprehensively compares the performance of some models (already tested in normal ambient temperature) in subfreezing condition to introduce the most reliable one for PEMFC cold start-up application. Firstly, seven models are compared regarding voltage losses and precision. Subsequently, the three most dependable ones are selected and experimentally compared at sub-zero temperature in terms of polarization curve estimation for three PEMFCs with different degradation levels. The results of this study indicate that the model introduced by Amphlett et al. has a superior performance compared to other ones regarding the characteristic's estimation in below-zero temperature. Highlights: A PEMFC model review for cold start-up application is conducted. A semi-empirical voltage model for sub-zero temperature condition is introduced. Online parameters estimation in sub-zero temperature is performed. Experimental validation for online characteristics extraction is indicated.