Fault Prognosis Method for Solid Oxide Fuel Cells Based on Mechanism Degradation Process Model and Particle Filtering

Abstract

The research of fault prognosis for Solid Oxide Fuel Cell (SOFC) plays a critical role in increasing lifetime and improving durability. Therefore, a model-based SOFC fault prognosis method combining the mechanism degradation process model and particle filtering is proposed. The State of Health (SOH), as well as End of Life (EOL) and Remaining Useful Life (RUL) are estimated by the proposed method. Firstly, the degradation resistance is separated in the SOFC electrical balance to signify SOH. Furthermore, the feasibility of this scheme is verified by the parameter identification results of the SOFC's electrochemical performance curves. Secondly, the degradation process model of SOFC is developed by aging mechanism analysis. On this basis, an online parameter estimation method based on a particle filter is utilized to estimate SOH and its Probability Density Function (PDF). Moreover, EOL, RUL, and their PDFs are predicted by extrapolating the time when the SOH reaches the performance threshold. Finally, the application cases of actual SOFC test data demonstrate that the proposed fault prognosis method effectively predicts the SOH evolution trend, and then gives reasonable EOL, RUL and their posterior PDFs.