Remaining useful life estimation for proton exchange membrane fuel cells using a hybrid method

Abstract

This paper proposes a complete hybrid prognostics method which can predict the degradation trend and estimate the remaining useful life of proton exchange membrane fuel cells (PEMFCs) under different current loads. The proposed hybrid prognostics method can be divided into two phases. In the first phase, the automatic machine learning algorithm that based on the evolutionary algorithm and the adaptive neuro-fuzzy inference system is proposed to predict the long-term degradation trend. In the second phase, based on the degradation data obtained in the first phase, the remaining useful life estimation is implemented by using a semi-empirical degradation model of PEMFCs and the proposed adaptive Unscented Kalman filter algorithm. Finally, the proposed hybrid prognostics method is validated by using the aging experimental data of PEMFCs. Test results show that the proposed hybrid prognostics method can achieve accurate long-term degradation trend prediction and remaining useful life estimation for PEMFCs.