Control-oriented thermal management strategies for large-load fluctuation PEM fuel cell systems

Thermal management control is of great significance to the performance and durability of proton exchange membrane fuel cell (PEMFC), which is challenging under large-load fluctuations due to its strong nonlinearity and variable time delay. Therefore, we employ cascade internal model control (IMC) to achieve better tracking performance under wide-range load variation and robustness against delayed disturbances, combining with current feedforward to reduce the time delay. Additionally, a double inner-loop cascade IMC for both thermostat and fans is proposed here to further improve the robustness, and a modified Smith predictor is introduced to ameliorate time-delay disturbance rejection. Firstly, the responsiveness and robustness of these proposed control strategies are evaluated by step tests and white noise disturbance tests, respectively. The results show that the cascade IMC of thermostat with the current feedforward control of fans (CS3) has the best responsiveness under load steps due to the time-delay reduction by current feedforward, while the double-inner loop cascade IMC with modified Smith predictor (CS2) exhibits the best responsiveness under ambient-air-temperature steps as well as the best robustness under either voltage interference or ambient temperature disturbances, indicating the effectiveness of its robust improvement and delayed disturbance rejection. Moreover, these control strategies are also validated under large-load fluctuation. CS3 is found to strictly keep the temperature tracking the target within ±0.6 °C, while CS2 shows a slightly worse convergence but presents the strongest temperature tracking under large-load fluctuations with voltage decay and disturbances, which shows practical value in automotive PEMFC systems, especially for long-term operation.