-
Paulo Victor Lopes Pires
Departamento de Engenharia Elétrica, Universidade Federal de Juiz de Fora, MG
-
Eder Barboza Kapisch
Departamento de Engenharia Elétrica, Universidade Federal de Juiz de Fora, MG
-
Leandro Rodrigues Manso Silva
Departamento de Engenharia Elétrica, Universidade Federal de Juiz de Fora, MG
-
Carlos Augusto Duque
Departamento de Engenharia Elétrica, Universidade Federal de Juiz de Fora, MG
-
Paulo Fernando Ribeiro
Departamento de Engenharia Elétrica, Universidade Federal de Itajubá, MG
Keywords:
Novelty Detection, Electric Power Quality, Electric Power Systems, Dynamic Time Warping, Edit Distance on Real signal, Time Warp Edit Distance
Abstract
This article presents the use of three dynamic similarity metrics: Dynamic Time Warping (DTW), Edit Distance on Real signal (EDR) and Time Warp Edit Distance (TWED) for detection of disturbances (novelties) in Electric Power Quality (EPQ) signals. DTW, EDR and TWED are mathematical tools for comparing two vectors, generating a measure of similarity between them. The detection of novelties happens by comparing a reference frame (vector) with the frame being analyzed. If the distance between the two frames exceeds a threshold value, a novelty is detected. To assess the quality of the proposed methods, the detection probability and the false alarm probability are determined, using the ROC (Receiver Operating Characteristic) curve. The tests are performed in a Matlab® environment. The three metrics showed a good performance in detecting novelty of signals, mainly TWED for signals with SNR = 60 dB and EDR for signals with SNR = 30 dB, attesting to the importance of applicability in detecting disturbances present in Electric Power Systems (EPS).
