Power System Contingency Classification Using KNN Algorithm

EasyChair Preprint 586

Abstract

Contingency analysis is an efficient technique in a large interconnected power system to identify the effect of post contingencies for its security. In this paper, Fast decoupled load flow method is used for each transmission line outage. The overall performance index (OPI) is calculated with the help of active power performance index (PIp) and Voltage performance index (PIv) for the static security classification of the power system. The static security is classified into five classes secure, critically secure, insecure, highly insecure and most insecure. The K nearest neighbour machine learning algorithm is proposed to classify these patterns. The proposed machine learning classifiers are applied on IEEE 14 and IEEE 30 bus test systems. Proposed KNN classifier is giving better accuracy for the classification of security assessment of the power system. Fuzzy logic approach has also been studied and implemented for the same test systems for the prediction of the above five classes.

Keyphrases: Fast decoupled load flow, KNN classifier., Overall Performance Index, contingency analysis

@booklet{EasyChair:586,
  author    = {Gongada Sandhya Rani and M Chakravarthy and B Mangu},
  title     = {Power System Contingency Classification Using KNN Algorithm},
  howpublished = {EasyChair Preprint 586},
  year      = {EasyChair, 2018}}