Skip to Main Content
Capacitor switching transient is one of the most commonly encountered power quality disturbances. Measuring and identifying these transients remains a challenge as the characteristics depend on the system parameters, capacitor ratings and makeup of the connected loads. This paper describes a systematic approach to design an automated system for identifying such transients. It relies on the belief that some of the system and component parameters are either known or can be determined, and thus can be used to develop the necessary pattern for identification and discrimination. However, identification also needs to be robust against uncertainties in these parameters. The proposed method combines the techniques of wavelet transform, rank correlation and fuzzy logic to account for the uncertainties. Dynamic simulations were undertaken to verify the method and to illustrate its robustness when dealing with various uncertain transient characteristics. It is shown to be capable of identifying transients caused by switching of both isolated and back-to-back capacitors.