This article presents a k-means clustering and weighted k-nearest neighbor (k-NN) regression-based algorithm for the protection of transmission line. Three-phase current signals of both the terminals are synchronized and sampled with a sampling frequency of 3.84 kHz. Cumulative differential sum (CDS) is computed by subtracting the samples of current cycle from the previous cycle at both the terminals of transmission line. k-means clustering is applied on CDS to compute two centroids using moving window of width, equal to one cycle. Difference between the absolute values of centroids is computed at both the terminals and represented by the centroid difference (CD). The CD of both the terminals is added to compute the fault index. The computed fault index is used to detect and classify the types of faults. The location of the fault is estimated by the weighted k-NN regression method. Various case studies are performed to validate the robustness of the algorithm for different fault parameters such as fault impedance and fault location. The effect of noise is also considered to check the accuracy of the proposed algorithm in the noisy environment.

A Novel k-Means Clustering and Weighted k-NN-Regression-Based Fast Transmission Line Protection

Siano P.
2021-01-01

Abstract

This article presents a k-means clustering and weighted k-nearest neighbor (k-NN) regression-based algorithm for the protection of transmission line. Three-phase current signals of both the terminals are synchronized and sampled with a sampling frequency of 3.84 kHz. Cumulative differential sum (CDS) is computed by subtracting the samples of current cycle from the previous cycle at both the terminals of transmission line. k-means clustering is applied on CDS to compute two centroids using moving window of width, equal to one cycle. Difference between the absolute values of centroids is computed at both the terminals and represented by the centroid difference (CD). The CD of both the terminals is added to compute the fault index. The computed fault index is used to detect and classify the types of faults. The location of the fault is estimated by the weighted k-NN regression method. Various case studies are performed to validate the robustness of the algorithm for different fault parameters such as fault impedance and fault location. The effect of noise is also considered to check the accuracy of the proposed algorithm in the noisy environment.
2021
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11386/4774681
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 16
  • ???jsp.display-item.citation.isi??? 13
social impact