A Study on Predicting the Degradation Performance of Non-stationary SOFC Systems Based on Time Convolutional Networks

Considering the inherent temporal nature of SOFC system performance degradation and the existing research gap concerning the prediction of non-stationary systems, this study employs Time Convolution Networks (TCNs) to predict the behavior of non-stationary SOFC systems. Specifically, the focus is placed on predicting the degraded performance characterized by voltage drop. The findings demonstrate that TCNs offer superior performance in predicting the degradation of SOFC systems across various operating conditions.