Model predictive control for aircraft engine thermal management under nonlinear heat transfer and time delay

Growing thermal loads and frequent engine shifts in advanced aircraft pose critical thermal management challenges. Fuel thermal management systems (FTMS) enable heat dissipation from distributed sources but exhibit nonlinearity and strong coupling that challenge conventional control methods. Although model predictive control (MPC) has been applied, its dependence on linearized models limits effectiveness across wide operating conditions. This paper applies the heat current method to derive a linearized equivalent model from the original nonlinear heat exchange constraints, explicitly incorporating time-delay effects. This model enables the design of a constrained MPC strategy, which seamlessly integrates real-world system characteristics to achieve robust and fully automated thermal management across wide operating conditions. Experimental validation on a full-scale three-medium (fuel/water/air) ground-test rig with an intermediate circuit demonstrates the effectiveness of the proposed strategy under realistic flight conditions. The system maintains air outlet temperature within strict operational limits while ensuring complete prevention of water boiling, achieving a maximum temperature deviation of 3.43 K and a time-averaged deviation of 0.96 K. These quantitative results confirm the controller's robust performance in managing complex thermal dynamics under strong time-varying operating conditions.