Advanced Online Broadband Impedance Spectrum Acquisition of Fuel Cells by S-Transform

Abstract

Electrochemical impedance spectrum (EIS) applied in fuel cell characterization and diagnosis facilitates its system management in the vehicular application, but the first consideration of impedance-based management is how to obtain impedance online. The traditional laboratory method based on sine sweep is unsuitable for vehicular EIS measurement as limited by measured time and equipment price. In this article, we design an excitation device that can generate the square wave signal containing many harmonics as the excitation current. Besides, introduce S-transform (ST) for online broadband fuel cell impedance spectrum fast calculation, whose window width can be adaptively changed with frequency, and meanwhile, preserving high computational efficiency by fast Fourier transform operation. Based on this, the performance of ST and the previous method based on Morlet wavelet transform (MWT) is compared under different operating conditions. Extensive experimental results show that the ST can calculate impedance with the mean relative error of less than 3% based on appropriate excitation, while the computed time can be saved by nearly 95% as compared to MWT. Furthermore, the effectiveness and robustness of ST against different current densities, cell temperatures, fault scenarios with various fault degrees, and stable time are also verified through experimentation.