Modeling and control-oriented thermal characteristics under variable load of the solid oxide fuel cell

The performance of the solid oxide fuel cell (SOFC) is greatly influenced by its operating temperature. Therefore, making a profound study of the thermal characteristics inside the SOFC stack and maintaining the operating temperature within a reasonable range are the difficulties. This paper proposes a novel modeling and control strategy. Firstly, for studying the thermal characteristics of SOFC under variable load current and designing a more effective control schedule, the problem of SOFC temperature dynamic modeling is resolved by the beetle antennae search (BAS) optimizing–based back propagation (BP) neural network model (BAS-BP) innovatively. The simulation indicated that the mean absolute error (MAE), mean absolute percentage error (MAPE), and mean square error (MSE) of the designed BAS-BP algorithm are 19.14%, 24.51%, and 27.67% lower than those of the traditional BP model. The designed modeling method is armed with higher accuracy and faster convergence speed. Then, a sliding mode controller (SMC) based on the disturbance observer is represented in this paper, wherein the disturbance observer is first presented to predict disturbance functions, while the SMC realizes the tracking control of the SOFC node temperature by compensating for the observation error instead of ignoring it. It can be seen from the simulation that the proposed control strategy compared to the conventional proportional-integral-derivative (PID) controller achieves a 62.57%, 61.11%, and 6.27% reduction in MAE, MAPE, and MSE. The simulated results demonstrate that the presented strategy has better robustness to disturbances in addition to its validity in compensating for observation error, for the SOFC node temperature control.