Accelerated Development of Power Response Control for Proton Exchange Membrane Fuel Cells in High-Altitude Environments Based on Domain Attention Adaptive Transfer Learning

Abstract The power response performance of proton exchange membrane fuel cells (PEMFCs) is crucial for vehicles to quickly and accurately meet their power demands. In high-altitude environments, reduced oxygen partial pressure can inhibit redox reactions, thereby affecting the power response performance of PEMFC. Therefore, this paper proposes an accelerated development method for PEMFC power response control in high-altitude environments. First, a PEMFC power response control architecture based on heterogeneous surrogate nearest neighbor policy optimization is constructed in a low-altitude environment. Second, this control architecture is transferred to the high-altitude environment using domain attention–based adaptive transfer learning. Finally, the hardware-in-the-loop simulation platform is used to verify the results in the low-altitude environment and two high-altitude environments. Under the two high-altitude operating conditions, compared with the non-transfer cases, the PEMFC power response control performance after transfer learning was improved by 29.6% and 35.9%, respectively. This demonstrates that the proposed method is significantly effective for accelerating cross-environment control development under current conditions.