Emre Yilmaz
(Advisor: Prof. Brian J. German)

will defend a doctoral thesis entitled,

Hybrid Automaton Based Nominal and Contingency Planning

for Over-Actuated Tandem Tiltwing eVTOL Aircraft

On

 Monday, June 26th at 12:00 p.m. (EDT)

Montgomery Knight Building 317

 Teams:

Click here to join the meeting

Abstract:

The objective of this dissertation is to explore nominal and off-nominal flight planning for a tandem tiltwing electric vertical takeoff and landing (eVTOL) aircraft with over-actuated flight controls based on the concept of a hybrid automaton. The hybrid automaton, which is a finite state machine composed of feasible trim points and maneuvers, provides safe transitions between flight modes. This formulation is leveraged because it helps to reduce the problem dimensionality and the computational complexity. This framework is specifically used to understand the performance limitations of the eVTOL aircraft in the transition corridor and while switching from a specific mode to another in rotor-borne or wing-borne flight. For contingency planning purposes, motor-out scenarios are considered. 

The redundancy in actuators of many eVTOL vehicles, e.g., distributed propulsion rotors and multiple control surfaces, results in many options to produce the same set of commanded forces and moments in the body-fixed reference frame. For this reason, control allocation (CA) strategies are investigated as a component of the broader planning problem. Control allocation optimization problems that employ different objectives are formulated based on force and moment commands for specific flight conditions such as rotor-borne flight, wing-borne flight, and transition. Specifically, a nonlinear programming formulation is proposed for control allocation optimization, and the problem is solved using a sparse nonlinear optimizer.

The control mappings are constructed as local polynomial surrogate models that relate the control effectors to the aircraft's aero-propulsive forces and moments across the full flight envelope, including interactional aerodynamics effects. The data used for constructing the surrogates are generated by conducting a design of experiments (DoE) and performing the corresponding analyses with an unsteady vortex lattice method (VLM) based flow solver.

The contributions of this dissertation are an interactional aerodynamics model suitable to serve as an example for control algorithm and autonomy research for an over-actuated tandem tiltwing eVTOL aircraft, a nonlinear control allocation optimization methodology for the entire flight envelope, and a hybrid automaton-based framework for nominal and contingency planning for the eVTOL aircraft.

Committee

• Prof. Brian German – School of Aerospace Engineering (advisor)
• Prof. Eric Feron – Electrical and Computer Engineering, KAUST
• Prof. J. V. R. Prasad – School of Aerospace Engineering
• Prof. Graeme Kennedy – School of Aerospace Engineering
• Dr. Justin Gray – CTO, Toolpath Labs