Active alarm method for switchgear status based on K-means clustering and multi-dimensional feature quantities

Abstract

Switchgear is the critical equipment of the distribution network, and its status evaluation is of great significance to improve the reliability of the distribution network. Addressing the issues of high subjectivity and low accuracy in current methods of status evaluation, we propose an active alarm method for switchgear status based on K-means clustering and multi-dimensional feature quantities. We configure multiple sensors at different locations inside the switchgear to monitor and obtain multi-source data based on IoT sensing technology. We establish a multi-dimensional status dataset of the switchgear and analyze its fluctuation. We acquire the relationship between the number of clusters K and the sum of squares of errors by clustering the obtained multi-dimensional dataset of switchgear through the K-means clustering status evaluation algorithm. We also calculate the sum of squares of errors under different clusters and the optimal number of clustering levels. Finally, we combine a set of switchgear operation data to obtain the switchgear health status by cluster analysis, which verifies the effectiveness of the evaluation method.