Simple temperature modeling of proton exchange membrane fuel cell using load current and ambient temperature variations

ABSTRACT

This paper proposes a simplified proton-exchange membrane fuel cell (PEMFC) temperature model for the purpose of estimating PEMFC temperatures with high accuracy using air-cooling systems. Besides knowing that most of the existing models were designed for specific systems, the proposed model also focuses on generalizing the conventional temperature model for easy adoption by other PEMFCs. The proposed model is developed based on the first-order exponential equation to avoid the limitations of complex mechanistic temperature models. The model uses only the information available from typical commercial PEMFCs, the main inputs of which are the current, elapsed time, and ambient temperature. In addition, the PEMFC area, number of cells in the stack, and high/low operating currents were incorporated in the proposed model to ensure its generalizability and applicability to different PEMFC technologies with air-cooling systems under various ambient conditions. The required model parameters were optimized using the Harris hawks optimization method. The proposed model was validated using experiments conducted on the Horizon-500 W and NEXA-1.2 kW PEMFC systems equipped with air-cooling mechanisms under different ambient temperatures and load currents. The root mean square error of all the examined cases was less than 0.5. The proposed model is helpful for simulations, dynamic real-time controllers, and emulators because of its fast response and high accuracy.

KEYWORDS:

• Current-based model
• dynamic temperature model
• Harris hawks’ optimization
• pemfc
• algorithm

Acknowledgment

The authors would like to thank UAEU library and Mr. Shehab Majud for the spell check and valuable proofreading. This research was co-founded by the United Arab Emirates University (Fund code no. 31R067) and the Natural Sciences and Engineering Research Council of Canada (Grant no. 2018-06527).

Additional information

Funding

This work was supported by the Natural Sciences and Engineering Research Council of Canada [2018-06527]; United Arab Emirates University [31R067].