Pearson's Correlation Coefficient for Islanding Detection Using Micro-PMU Measurements

Abstract

An islanding detection scheme (IDS) based on using frequency and phase angle information obtained from the micro-phasor measurement units (μPMU) is proposed in this article. From the μPMU data of reference substation and distributed generator terminals, two parameters are formulated; a cumulative sum of frequency difference and a phase angle difference. In the event of islanding, a linear correlation exists between the two parameters. The linear correlation between the two parameters is evaluated in real time using Pearson's correlation coefficient formula. A time bound threshold is considered to differentiate islanding events from other non-islanding events like load switching and faults. The advantage of the proposed IDS is it does not require the establishment of any frequency and phase angle based thresholds. Also the islanding detection time is independent of power mismatch conditions. Real-time performance of the algorithm is accessed on a test bench comprised of dSPACE DS1104 real-time controller and a real-time digital simulator (RTDS).