Diagnosis of Voltage Losses of the Proton Exchange Membrane Water Electrolyzer with Nonlinear Frequency Response Method

Abstract

Proton exchange membrane water electrolyzer (PEMWE) takes a central place in the hydrogen economy as a crucial technology for green hydrogen production. However, this technology still faces challenges, such as the cost of the materials, performance losses at high current densities, and durability. To overcome these obstacles, better comprehension of the phenomena occurring in the PEMWE is necessary. Typical electrochemical methods, such as polarization curves and high-frequency response measurements, lack detailed information about the individual contributions to the overall voltage losses and cannot point out the causes of the performance shortcomings. Therefore, the development of advanced diagnosis techniques for PEMWE is of importance both for the state of health analysis during the operation, as well as for the testing of new materials.In this work, a nonlinear frequency response (NFR) method was applied for the diagnosis of a lab-scale PEMWE. The NFR method represents an extension of the electrochemical impedance spectroscopy to the nonlinear domain by investigating the second-order frequency response function (FRF) in addition to the first-order FRF (equivalent to the impedance) [1]. Typical features corresponding to processes occurring in the PEMWE were observed in both the first- and the second-order FRFs. The experimental observations were interpreted based on a simple nonlinear model. The features observed at frequencies higher than 0.1 Hz were attributed to the electrochemical half-reactions, while a high current density feature observed at lower frequencies, could be attributed to the mass transport. During the analysis, the first-order FRF (impedance) was found not to be sensitive enough for the discrimination of the different sets of parameters and the second-order FRF had to be included for better parameter identification. The obtained parametrized model is dynamic and nonlinear and, thus, it can be utilized for model predictive control of the PEMWE and optimization of the overall green hydrogen production system.[1] Vidaković-Koch, T., Miličić, T., Živković, L.A., Chan, H.S., Krewer, U., Petkovska, M., Cur. Opin. Electrochem., 2021, 30, 100851