Hierarchical Model Predictive Control for Real-Time Energy-Optimized Operation of Aerospace Systems

This paper presents a hierarchical model predictive control (MPC) framework for hybrid power or propulsion systems that can be used in future energy-optimized aerospace systems. This control framework aims to decouple the slow energy optimization from dynamic power management in a holistic optimized manner and can be implemented on heterogeneous real-time computing hardware. The higher-level MPC minimizes the power generation costs and power delivery losses subject to dynamic load profiles, as well as maintains a dynamic level of reserved energy in energy storage to meet future changes in power demand. The lower-level MPC is to optimize the dynamic power management, i.e., controlling the currents to satisfy transient power flow needs while regulating the bus voltage. As such, the power generators should not necessarily meet the peak power demand, but just the average load demand. The proposed hierarchical MPC framework organically integrates energy optimization and dynamic power management in a decoupled manner and is suitable for integrated control of propulsion, power and thermal systems which are multi-input-multi-output control systems and very challenging with traditional PID-like approaches. This hierarchical MPC framework can perform multi-objective optimization and meet multiple constraints such as economic, operational, safety and power quality constraints at different levels. Challenges such as control objectives assignment, model selection, and constraint considerations are addressed in the paper. This framework can cover a wide bandwidth of model fidelity and enforce all necessary physical laws in the models using both MPCs together. State variable constraints are explicitly included in the controller formulation by using the equality constraints (i.e., discretized model or system state transfer equations) to transform the state constraints to the control constraints. The hierarchical MPC framework allows operational constraints to be assigned to different levels of MPC schemes. Simulation results show that the hierarchical control system operates well in optimizing for both energy flow and dynamic current/voltage regulations under dynamic conditions. By examining the interactions between two levels of MPC, it is shown that the objectives and constraints can be transferable but tradeoff should be made to achieve better coordination. This work will provide a capability of reconfigurable control paradigm on the same control platform via dynamic selection of control objectives and models.