In this paper, a new approach for load profile segmentation is investigated for residential energy consumption.
The proposed approach considers the daily level granularity and identifies dominant patterns of energy consumption for individual participants.
The analysis uses adaptive k-means clustering to determine the number of clusters that improve the distances between data points and cluster centroids.
The proposed method is applied to Ausgrid Solar Home Electricity Dataset for energy consumption data of 300 houses over 1 year.
The results demonstrate distinctive features including peak energy consumption, time of peak energy use, as well as seasonal variations.
The findings can help utilities to optimise demand response and pricing strategies.