Proton exchange membrane fuel cells (PEMFCs), as a clean energy technology, show remarkable potential for a wide range of applications. However, high altitude regions pose significant challenges for PEMFC system operation due to thin air and low oxygen partial pressure. Existing logic judgement-based controls exhibit defects such as poor robustness and poor adaptability, which seriously restrict PEMFC system operation. In order to address this issue, this paper puts forth an intelligent control of a PEMFC system air compressor (AC) and back pressure valve (BPV) using an asynchronous advantage actor-critic (A3C) algorithm and systematically compares it with the logic judgement-based control. The application of an A3C-based control under three distinct high altitude test conditions demonstrated a notable enhancement in dynamic responsiveness, with an improvement of up to 40% compared to the results for the logic judgement-based control. Additionally, an improvement of 5.8% in electrical efficiency was observed. The results demonstrate that the A3C-based control displays significant robustness and control precision in response to altitude alterations.
Read full abstract