Model based temperature controller development for water cooled PEM fuel cell systems

Saygili Y., Eroglu I., Kincal S.

INTERNATIONAL JOURNAL OF HYDROGEN ENERGY, vol.40, no.1, pp.615-622, 2015 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 40 Issue: 1
  • Publication Date: 2015
  • Doi Number: 10.1016/j.ijhydene.2014.10.047
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.615-622
  • Keywords: Fuel cell thermal management, Cooling system modeling, Cooling system characterization, Water cooled fuel cell, THERMAL-ANALYSIS, MEMBRANE, EXCHANGE, PERFORMANCE
  • Middle East Technical University Affiliated: Yes


PEM (proton exchange membrane) fuel cell operation necessitates thermal management to satisfy the requirements of safe and efficient operation by keeping the temperature within a certain range independent of varying load conditions. Heat generation within the fuel cell changes according to the power delivered from the stack, requiring a dynamic control system to remove this excess heat and maintain the desired stack temperature. In this study, a closed loop water circulation strategy is considered and evaluated for cooling a 3 kW PEM fuel cell. The cooling system consists of a water circulation pump and a radiator coupled to a fan, integrated with the fuel cell stack. A first principles based model is developed for the integrated cooling system through an energy balance containing the relevant terms. A sequence of dynamic tests is performed on the cooling system to identify the parameters appearing in the model developed. The resulting semi-empirical model is used to evaluate possible control strategies managing the cooling loop. Three specific strategies are analyzed and the performances of these controllers are evaluated in terms of stack temperature, integral time weighted absolute error (ITAE) and the parasitic energy requirements. Minimizing fan usage with an on/off controller while keeping the pump voltage as a continuously manipulated variable through a feedback PI (proportional integral) controller delivers the best results. The MATLAB-SIMULINI (R) platform is used in the development and implementation of the models and controllers. In our strategy, the characterization of the cooling loop is physically de-coupled from the development of the fuel cell stack - allowing for the evaluation of candidate equipment and algorithms prior to the fuel cell stack being available, which is often the case during prototype development. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.