Development of PSS tuning rules using multi-objective optimization

Abstract

This paper explores the process of tuning controllers for conditions that better reflect realistic operating conditions, such as input signals of different shapes. In this study, controllers are tuned simultaneously for input signals of different shapes, using the tools of double objective optimization. For this purpose, a novel performance index is proposed to evaluate a system’s ramp response. At the same time, a novel method is proposed for faster and more efficient generation of Pareto plots and Pareto frontiers. The proposed method retrieves the dominant Pareto solutions only, and produces the Pareto frontier directly. The study is applied on the tuning of a Power System Stabilizer (PSS) of a simple power system, at several operating points of the synchronous machine. Finally, a relation between the optimum PSS parameters and the operating points is sought. A linear regression model is used for this purpose, and the tuning rules are tested in an application study. Finally, the optimization technique is applied on a multiple machine system. The results show significant improvement, while maintaining fault resilience.