Load Profile Segmentation of Various Load Categories using Clustering

Abstract

With the increasing variability in load profiles due to intermittent disperse generation and demand, load aggregation and its analysis has become crucial. The segmentation of these load profiles may facilitate the system operators with more accurate forecasts and hence reliably operate the network. As a result, load profiles segmentation has become significant and valuable. The increasing use of smart metering infrastructure has resulted in the availability of daily energy consumption data, which may be used to analyze energy consumption patterns. This paper investigates the analysis and clustering of load profiles by segmentation of daily energy consumption load profiles of different load categories. The proposed method identifies the dominant patterns of energy consumption of different categories of loads. This approach uses adaptive k-means clustering for the clusters assignments which adapts cluster number to improve the Euclidean distance between the data points and centroid of the clusters. The results demonstrate distinctive features like peak and off-peak energy consumption. The findings can be helpful to the utility or aggregators to design and optimize the demand response, analyze energy efficiency, peak load management, pricing strategies design according to consumer groups.