Predictive fuel cell thermal management for fuel cell electric tractors

This study addresses the significant challenges associated with adopting fuel cell powertrains for agricultural machinery, particularly concerning efficiency and durability due to the demanding operational environment. A critical factor is the fuel cell’s operational temperature, which can lead to degradation, higher auxiliary consumption, and larger radiator volumes. To mitigate these problems, the present study introduces a predictive control approach for thermal management. Specifically, the notable advantages of the non-linear model predictive controller over classical control approaches can be attributed to the combination of a control-oriented model and predictions into a real-time optimization problem. This approach stands as an innovative addition aimed at compensating the inertia of the cooling system while deploying predictions to improve the control accuracy and concurrently optimize the utilization of actuators. This work is organized into two principal contributions: the extensive modeling of a fuel cell system and its validation, and the comprehensive investigation of a model predictive control strategy. The results demonstrate that a predictive thermal management strategy can significantly diminish auxiliary consumption by up to 30% compared to classical control strategies across various ambient temperatures without compromising temperature reference control. In particular, a comparison with a classical control strategy reveals the effective deployment of multiple actuators and prediction under the prescribed constraints in the proposed control concept. Additionally, the study quantifies the impact of ambient temperature on auxiliary consumption and identifies operational scenarios where model predictive control performs optimally. As part of the unique contribution of this work, the cost function weights, length, and accuracy of the prediction horizon are also analyzed, with findings showing that a balance between performance and actuator consumption can be achieved. • Fuel cell system modeling from experimental data. • Analysis of stack losses and physical characteristics of the thermal system. • Significant reduction of auxiliaries power with model predictive control. • Influence of ambient temperature and load on auxiliary consumption. • Influence of cost function weights, length, and accuracy of the prediction horizon.