Unconventional frequency response analysis of PEM fuel cell based on high-order frequency response function and total harmonic distortion.

Electrochemical impedance spectroscopy (EIS) is a commonly used technique for studying internal dynamics of the proton exchange membrane (PEM) fuel cell. However, due to its linear frequency response nature and reliance on the linear target hypothesis, EIS has limitations in analyzing the nonlinear behavior of the PEM fuel cell, which is the inherently nonlinear system. To overcome this issue, this paper employs an unconventional frequency analysis technique that utilizes the high-order frequency response function (FRF) and total harmonic distortion (THD) in a broad frequency range, to conduct the nonlinear frequency response analysis (NFRA) of the PEM fuel cell. In addition to the traditional electrochemical impedance based on the first-order harmonic response, the high-order harmonic response is also included. Based on this, the appropriate excitation amplitude is determined through a comprehensive examination of the impact of varying excitation amplitudes on the impedance, second-order FRF, and THD. Then, the impacts of operating conditions, such as reactant stoichiometry, humidity, temperature, and pressure, on the FRF and THD were investigated incorporating the dynamics losses computed using EIS and equivalent circuit model. Furthermore, the nonlinear frequency response is investigated under various fault conditions, such as flooding, membrane drying, and air starvation. The results demonstrate the heightened sensitivity of the NFRA method towards detecting changes in internal processes of the PEM fuel cell, which offers a multidimensional frequency-domain perspective, facilitating the study of the advanced diagnosis, system control, material optimization, and even other electrochemical energy sources. • FRF and THD spectrum are used for nonlinear frequency response analysis of fuel cell. • Effect of various operating conditions on nonlinear frequency response is analyzed. • FRF and THD spectrum provide new perspicacity for fault feature extraction. [ABSTRACT FROM AUTHOR]