This dissertation presents a hierarchical model predictive control (MPC) framework for energy management onboard vehicle systems. High performance vehicle systems such as commercial and military aircraft, on- and off-road vehicles, and ships present a unique control challenge, where maximizing performance requires optimizing the generation, storage, distribution, and utilization of energy throughout the entire system and over the duration of operation. The proposed hierarchical approach decomposes control of the vehicle among multiple controllers operating at each level of the hierarchy. Each controller has a model of a corresponding portion of the system for predicting future behavior based on current and future control decisions and known disturbances. To capture the energy storage and power flow throughout the vehicle, a graph-based modeling framework is proposed, where vertices represent capacitive elements that store energy and edges represent paths for power flow between these capacitive elements. For systems with a general nonlinear form of power flow, closed-loop stability is established through local subsystem analysis based on passivity. The ability to assess system-wide stability from local subsystem analysis follows from the particular structure of the interconnections between each subsystem, their corresponding controller, and neighboring subsystems. For systems with a linear form of power flow, robust feasibility of state and actuator constraints is achieved using a constraint tightening approach when formulating each MPC controller. Finally, the hierarchical control framework is applied to an example thermal fluid system that represents the fuel thermal management system of an aircraft. Simulation and experimental results clearly demonstrate the benefits of the proposed hierarchical control approach and the practical applicability to real physical systems with nonlinear dynamics, unknown disturbances, and actuator delays.
PDF
download
878987797
028-85220240
