AN IMPROVED CWT-BASED ISLANDING DETECTION METHOD FOR A DEVELOPED MICROGRID IN REAL-TIME

Abstract

The island mode operation problem is a significant event of deterioration in a power system, and this fault must be detected in the fastest and most accurate way for the reliable operation of the microgrid structure. Recently, numerous islanding detection methods based on signal processing have been proposed in the literature. In this study, an improved, continuous wavelet transform (CWT)-based islanding detection method is proposed for microgrids. Island mode conditions are investigated in the developed PV-based microgrid connected to a low voltage grid. The proposed method uses only the voltage signal on the point of common coupling (PCC). A series of discrete values are selected for scales and shifts of continuous wavelets and then CWT is applied for PCC voltage. In this way, the computational load is minimized. This method has many advantages comparing to conventional methods and has been tested in real-time for a PV-based microgrid prototype. The results show that the developed CWT-based islanding detection method can detect different types of island modes in the developed microgrid. Besides, the islanding detection time of the proposed method varies between 105-110 ms in any island mode operations, and it is faster than the conventional detection methods. None detection zone (NDZ) is also almost zero in the proposed method. Thus, the CWT-based islanding detection method provides both a reliable NDZ and a short detection time for microgrid applications.