CONTROL STRATEGIES FOR COMMUNICATION INDEPENDENT ISLANDING DETECTION IMPLEMENTATION

Abstract

This thesis project focuses on the investigation and integration of two islanding detection methods: the fuzzy inferencing system (FIS) and the wavelet packet transform (WPT). The primary objective is to address the limitations of the FIS method through the incorporation of the WPT method, while also ensuring independence from a communication system. The FIS method offers a robust framework for modelling complex relationships and uncertainties in power systems. However, it may exhibit limitations in accurately analysing transient events and disturbances, potentially leading to false detection outcomes. To overcome these limitations, the WPT method is introduced, which decomposes power system signals into frequency sub-bands, enabling the capture of unique patterns associated with islanding events. Through extensive simulation experiments, the effectiveness of both the FIS and WPT approach are evaluated. The simulation results consistently support the superiority of the WPT method compared to the FIS method. Moreover, a significant advantage of the proposed methods is their independence from a communication system. This aspect eliminates the need for additional infrastructure and reduces the potential for communication failures or latency issues that can affect the reliability of the islanding detection process. The findings of this thesis project have important practical implications. The research outcomes contribute to the development of advanced protection mechanisms that enhance the resilience and stability of power systems, catering to the needs of the increasing DG penetration into the grid. Looking ahead, the future scope of this research includes further optimization of the combined FIS and WPT method, integration with advanced machine learning techniques, real-time implementation, testing in diverse power system scenarios, validation against alternative islanding detection methods, and integration with smart grid technologies.