152. A QUANTIZED-INPUT CONTROL LYAPUNOV APPROACH FOR MOTOR DRIVES

Department: Mechanical & Aerospace Engineering
Research Institute Affiliation: Center for Control Systems and Dynamics (CCSD)
Faculty Advisor(s): Miroslav Krstic | Massimo Franceschetti
Award(s): Honorable Mention

Primary Student
Name: Gideon Andrew Prior
Email: gprior@ucsd.edu
Phone: 310-259-4666
Grad Year: 2013

Abstract
We present a new method for the generation of input switching sequences in a motor control system that minimizes switching events while providing stability guarantees. Typical reference input realization methods such as space vector modulation rely on high frequency state space averaging which can yield unnecessary switching events and increased switching losses. While in many cases non-minimal switching loss may be acceptable as other performance objectives are selected for optimization, power electronic devices are frequency limited according to the amount of current they are regulating and in applications requiring control of large currents, high frequency switching ceases to be an option. In this work we adopt a quantized-input view of the motor-inverter system by using a nonlinear dynamic model to evaluate a control Lyapunov function for each member of the small discrete set of realizable inputs, yielding an energy related input selection process that provides stability guarantees with minimal switching. We provide a theoretical analysis of a motor-inverter system leading to a stability proof for a quantized input control law. The method is shown provide excellent dynamic response and robustness to parametric uncertainty while operating at low switching frequencies. The controller performance is verified through computer simulations and experimental results.

« Back to Posters or Search Results