Fuzzy logic power system stabiliser in multi-machine stability studies.

Abstract

Conventional power system stabilisers (PSS) are designed to eliminate poorly damped, low frequency power oscillations that occur between remote generating pools or power stations, due to different types and settings of the automatic voltage regulators at different power stations. The supplementary control of the PSS is exerted on the power system through a generator's excitation system to which the PSS is attached. In order to design these conventional power system stabilisers , requires accurate system data and an in-depth knowledge of classical control theory. This thesis investigates the use of an intelligent, non-linear PSS that utilises fuzzy logic techniques. Others have proposed the concept of such a PSS, since it does not require accurate system data. This thesis describes the basic aspects of power system stability . Thereafter the methods of modelling synchronous machines in a multi-machine power system are presented. The sample power system being studied and the simulation packages used in the investigations are introduced and the methods involved to design and tune a conventional power system stabiliser using classical control theory and design methods proposed by others, are discussed. The general concept of fuzzy logic is introduced and the application of fuzzy logic techniques to controller design is explained. Using the principles of fuzzy logic controller design, a fuzzy logic power system stabiliser utilising 9 rules is designed and tuned for the multi-machine power system under investigation. The fuzzy logic stabiliser is then applied to a synchronous motor in a pump storage scheme. Previous work has applied fuzzy logic stabilisers only to synchronous generators . To further compare the performance of the 9 rule fuzzy stabiliser, a 49 rule stabiliser developed by other researchers, and adapted to operate on the synchronous motor, is evaluated. Computer simulated results of each of the stabiliser's performances are presented. The results of the 9 rule fuzzy stabiliser are compared with the performance of a conventional linear stabiliser as well as with a 49 rule fuzzy stabiliser. The robustness properties of the fuzzy stabilisers are evaluated. The results further prove that with proper membership function selection, a simple fuzzy stabiliser that demands very little computational overheads can be achieved to provide adequate system damping.