Type-2 Fuzzy Q-learning-based Energy Management Strategy for Fuel Cell Hybrid Power with Uncertain Parameter

Fuel cell hybrid systems (FCHS) hold substantial promise for advancing sustainable transportation solutions. However, the maximum performance point of proton exchange membrane fuel cell exhibits significant variability under parameter uncertainties, posing challenges for energy management. To address this issue, this study proposes a type-2 fuzzy Q-learning-based energy management strategy (EMS) integrated with uncertain parameters for PEMFC-battery hybrid systems. First, a high-fidelity mathematical model of the hybrid system is developed, and the correlation between PEMFC efficiency dynamics and uncertain parameters is rigorously analyzed. Subsequently, a type-2 fuzzy logic system is embedded into the Q-learning framework to optimally allocate power between the fuel cell and battery, thereby enhancing system overall performance. To improve generalization across diverse operating conditions, the multi-scenario design methodology is incorporated into the Markov decision process (MDP). The efficacy of the proposed strategy is validated through hardware-in-the-loop (HIL) experiments.