Novelty Detection Based on Dynamic Time Warping Similarity Metric Applied to Power Quality Signals

Abstract

This paper presents an application of the Dynamic Time Warping (DTW) similarity metric in the detection of electrical disturbances in Power Quality (PQ) Signals. The DTW is a mathematical tool used to compare two vectors, generating a measure of similarity between them. The detection of disturbances is performed by comparing a reference vector (frame) with the frame under analysis. If the distance between the two frames exceeds a threshold value, a disturbance is detected. To assess the performance of the proposed application, the detection probability and the false alarm probability are evaluated with different threshold values, using the Receiver Operating Characteristic (ROC) curve. Tests are performed using Matlab® environment and the performance of the application is compared to another disturbance detection technique based on similarity present in the literature.