Reset control systems: Stability, performance and application
Linear time invariant (LTI) control design is the most widely applied control design technique. But it has inherent limitation. In LTI feedback control systems, high-frequency loop gain is linked to both low-frequency loop gain and stability margins through Bode's gain-phase relationship. This linkage results in trade-offs among competing performance specifications. Reset control design is a great candidate of improving this limitation. The basic idea of reset control is to reset the state of a linear controller to zero whenever its input meets a threshold. Such a reset controller is introduced in feedback control systems with the aim of providing better trade-offs between competing specifications than could be achieved using linear controllers.
There have been some successful experimental applications of reset control technique which demonstrate its benefit. However, during the past three decades, there is a lack of theoretical results for reset control systems. For example, none of the experimental reset control applications can provide formal proof of stability and theoretic analysis of performance. Dedicated to solving this problem, this dissertation performs theoretic analysis of the stability and performance of reset control systems. A complete set of theoretic results on the stability and time-domain performance are developed. These results build a solid theoretic base for the further and wider application of reset control system. Also, reset control design is successfully applied to the speed control of a rotational flexible mechanical system with supporting theoretic analysis. This experiment demonstrates the benefit of reset control design over LTI control design and the effectiveness of the theoretic results.
0544: Electrical engineering