Water content estimation in polymer electrolyte fuel cells using synchronous electrochemical impedance spectroscopy and neutron imaging

Abstract

Polymer electrolyte fuel cells are a crucial piece of approaching net zero due to their high power density, rapid refueling, and eco-friendly operation. However, stable performance and durability rely on subtle water balance. Existing water management strategies, including humidification, drainage, and cold starts, primarily depend on indirect feedback or calibration through the output voltage. The direct, real-time measurement of the overall water content inside a fuel cell remains challenging, hindering the implementation of efficient feedback water control. To address this issue, synchronous measurement of neutron imaging and electrochemical impedance spectroscopy are carried out at various water contents. Machine learning is used to establish a non-linear correlation between the two characterizations. This enables the development of a more cost-effective and attainable real-time water-content estimation technique—inferred from a universal electrochemical impedance spectroscopy tool rather than relying solely on the limited availability of neutron imaging, which will facilitate the optimization and advancement of polymer electrolyte fuel cells.