This dissertation studies the modeling and control of an urban air mobility (UAM) electric vertical take-off and landing (eVTOL) aircraft with tilt-rotors, covering a brief UAM survey, analytical nonlinear flight modeling, control-oriented linear parameter varying (LPV) modeling, adaptive model predictive control (MPC) design and optimization with simulation validation. The core UAM aircraft control framework adopts the adaptive MPC with dynamic reference compensation (DRC) based on developed LPV models with certain modifications and improvements to adaptive MPC due to specific UAM aircraft characteristics. Both hovering and tilt transition control cases are studied. Furthermore, motor failure cases are also investigated under hovering and tilting flight conditions, where motor failure is introduced by limiting motor output power to the corresponding level of failure and the adaptive MPC design is conducted by modifying associated hard constraints. Note that failure recovery control is used to study safety redundancy of UAM aircrafts.UAM has attracted tremendous attention from both transportation industry and academia community. It is viewed as the promising future of next-generation transportation and is expected to be a quiet, fast, clean, efficient, and safe point-to-point transportation. The most promising and popular UAM concept is eVTOL aircrafts. Equipped with a distributed electric propulsion (DEP) system, eVTOL aircrafts use multiple electric motors and propellers to provide the required lift force for vertical takeoff and landing, and tilt-rotors (or extra push rotors) are used to cruise horizontally. The fixed-wing design allows for improving cruise efficiency during the flight. However, aeromechanics and flight dynamics of tiltrotor eVTOL aircrafts are complicated due to the increased system complexity. Thus, a comprehensive study through dynamic modeling and control of eVTOLs is necessary, which will form a foundation for future dynamic analysis, controller design and optimization. This dissertation firstly introduces an analytical flight dynamic formulation for UAM eVTOL aircrafts that feature tiltrotors for fixed-wing level flight, vertical takeoff and landing. In this analytical formulation, a nonlinear rigid-body dynamic model is developed by incorporating multiple tiltrotor dynamics, their gyroscopic and inertial coupling effects. A quasi-steady aerodynamic formulation is implemented to calculate the aerodynamic loads on all lifting surfaces. In addition to the conventional control surfaces for fixed-wing aircrafts, such as elevator, both tilt angle and individual rotor rotational speed are considered as control inputs in the formulation of nonlinear flight dynamic model. These nonlinear dynamics is then linearized with respect to a set of trimmed flight conditions of interest (or a tilt operational trajectory) to render the corresponding linear time-invariant state-space models used to create an LPV model as a function of flight condition to be used for adaptive MPC control design. Based on the above analytical model, the aircraft control strategies for hovering and tilt transition operations are investigated using adaptive MPC based on the developed LPV models. To model the aircraft dynamics variation, a family of linearized time-invariant models are obtained by trimming the nonlinear aircraft model at multiple equilibrium conditions or linearizing along a predetermined flight trajectory, and the associated LPV model is obtained by linking the linear time-invariant models using scheduling parameter(s), where the system variation is modeled by the measured scheduling parameter(s) under current flight condition. The proposed adaptive MPC strategy is developed to optimize the system output performance, including the rigid-body aircraft velocity and altitude, and solved by using quadratic programming optimization with dynamic reference compensation (DRC), subject to a set of time-varying hard constraints representing the available propeller acceleration calculated based on the available motor power. The adaptive MPC simulations are conducted based on both developed LPV and nonlinear rigid-body models, and the simulation results demonstrate that the designed adaptive MPC controllers are able to achieve desired closed-loop performance for the target UAM aircraft.
Read
