Isolation Forest Based Method for Low-Quality Synchrophasor Measurements and Early Events Detection

Abstract

This paper proposes an online data-driven approach that utilizes phasor measurement unit (PMU) data for early-event detection and low-quality data monitoring based on isolation forest (iForest). By skillfully selecting the feature subspaces, we design three levels of detectors that are capable of distinguishing early events from low-quality data measurements. The proposed online detection algorithm is practical in the sense that it does not require any prior knowledge of the grid topology or communication among buses. Besides, it is fast responding with low computational complexity, and thus is suitable for online applications. Numerical simulations with synthetic PMU data validate the effectiveness of the proposed method.