Robust model design for evaluation of power characteristics of the cleaner energy system

Hydrogen based fuel cell such as solid oxide fuel cell (SOFC) combines compressed hydrogen and oxygen from the air to produce electricity. Since, this technology does not emit emissions and therefore also known as cleaner energy systems. For improving the performance of the fuel cell, it is highly important to understand the effect of operating conditions on its performance. In this context, experimental studies are conducted to understand the fundamentals of the fuel cell mechanism. In view of limited resources, hence numerical studies also become crucial for design of robust models for determining and optimizing the power density based on the dynamic operating conditions. In real scenario, there exist uncertainties in precise measurement of operating conditions such as the temperature and the flow rate of hydrogen, nitrogen and oxygen. In this work, the automated neural search (ANS) approach is proposed to formulate the relationships between power density and the operating conditions. Two types of uncertainties, namely the settings of the ANS approach and in the operating conditions are considered to formulate the robust models. Optimization performed on the robust model reveals that the operating temperature of 778 °C, hydrogen and oxygen flow rate of 1 L/min are the optimum settings for achieving maximum power density of 574.2 mW/cm².