You're invited to attend
(Advisor: Prof. Dimitri Mavris)
"A Controller Development Methodology Incorporating Unsteady, Coupled Aerodynamics and Flight Control Modeling for Atmospheric Entry Vehicles"
Thursday, July 29
Future missions to the surface of Mars will present new requirements for entry, descent, and landing, including precision landing with an uncertainty less than 100 meters. Aerodynamic and control system uncertainty during entry are significant sources of this uncertainty. The aerodynamics are fundamentally unsteady, and actuation of the control system can cause coupled dynamic behavior. Meeting the precision landing goal will require both a reduction in model uncertainty, and the modification of the controller development methodology to leverage the increased model fidelity. The objective of this dissertation is to formulate and implement an entry controller tuning methodology that directly accounts for coupled, unsteady behavior. Significant progress in modeling entry vehicle aerodynamics and flight dynamics has been made over the past decade by coupling 6 degree-of-freedom rigid body dynamics simulations with computational fluid dynamics flow solvers. The proposed work seeks to incorporate guidance and control system modeling into the coupled simulation, allowing high-fidelity evaluation of an entry vehicle control system design and accompanying entry controller.
However, the added cost of coupled simulation makes the Monte-Carlo-based state-of-the-art controller tuning methodology impractical; instead, a multifidelity optimization method is used. The coupled simulation serves as the high-fidelity model, simulating an entry vehicle with open- or closed-loop control for direct evaluation of a controller. The trajectory and aerodynamic data produced by each run of the coupled simulation is used to train an aerodynamic surrogate for fast, low-fidelity modeling. The proposed methodology will be developed and evaluated through experimental testing of the NASA Supersonic Decelerator Flight Test entry vehicle equipped with a two-axis internal moving mass actuator control system, simulating the tuning of an entry controller for said vehicle. Surrogate model and multifidelity optimization methods will be evaluated to identify a satisfactory architecture. A comparison will be made against the state-of-the-art methodology to demonstrate the performance benefits of the proposed methodology.
- Professor Dimitri N. Mavris – School of Aerospace Engineering, Georgia Institute of Technology
- Professor Mark F. Costello – School of Aerospace Engineering, Georgia Institute of Technology
- Professor Lakshmi N. Sankar – School of Aerospace Engineering, Georgia Institute of Technology
- Dr. Bradford Robertson – School of Aerospace Engineering, Georgia Institute of Technology