Phaselet Approach for Islanding Detection in Active Distribution Networks

A novel moving window phaselet based approach has been proposed in this paper for quick islanding detection (within 5-10ms) in ADNs, even in the case of zero power mismatch/ power balance case. The Instantaneous Phase Angle (IPA) and magnitude of the PCC voltage have been extracted through a linearly moving window phaselet to generate two indices - Phaselet Angle Difference (PAD) and Phaselet Magnitude Difference (PMD) at sub-cycle rate, which are critical parameters in quickly discriminating islanding from non-islanding events in ADNs. These indices also offer the added advantage of quickly classifying the type of fault, in the events of faults in ADN. Real Time Digital Simulator (RTDS) based results prove the efficacy of the proposed phaselet based technique in accurately and dynamically estimating the phasors within sub-cycle interval and quickly detecting islanding, even in power balance case with different load quality factors as per standard.

I. Introduction

The presence of Distributed Generations (DGs) in an Active Distribution Network (ADN) has the major advantages of reduced pollution and increased power supply reliability [1]. However, few notable disadvantages of integrating DGs are the power quality issues due to the converters of DGs, unintentional islanding from the grid and bidirectional power flow. Thus, protection of ADN becomes further more challenging due to the power quality issues and bidirectional power flow. Unintentional disconnection of a DG in the active network may lead to technical issues and hazards to the system, such as reduced power supply and reliability, worker safety, quality of supply, lack of proper grounding of islanded network, synchronization issues, etc. [2]. Operation and control standard for grid integration in ADNs by IEEE 1547 [1] suggests that an unintentional disconnection of a DG should be detected within 2s of disconnection to avoid any system hazards.