The National Technical University of Athens (NTUA), one of the DEDALUS scientific partners, has completed a study on grouping residential electricity consumers, based on their historical electricity consumption, to create more targeted demand-response programs.
This grouping will be used in almost every DEDALUS service at the end of the day, making the services more targeted per group. The study was published in the journal Applied Energy.
Specifically, the paper introduces a machine learning-based framework to optimize demand response programs. Using data from nearly 5,000 households in London, four clustering algorithms—K-means, K-medoids, Hierarchical Agglomerative Clustering, and DBSCAN—were evaluated to identify groups with similar consumption patterns.
The problem was reframed as a probabilistic classification task, leveraging Explainable AI to improve model interpretability. The optimal number of clusters was found to be seven, although two clusters, comprising around 10% of the data, exhibited high internal dissimilarity and were excluded from further consideration.
This framework offers a scalable solution for utility companies to enhance the targeting and effectiveness of demand response initiatives.
“Our research aims to tackle a key challenge in energy management: efficiently identifying and classifying household energy consumption patterns to enhance the implementation of Demand Response programs”, said Vasilis Michalakopoulos—one of the paper’s authors.
“Optimizing household energy use is increasingly critical, both for promoting environmental sustainability and for enabling utility companies to deliver more targeted and effective DR solutions.
“This work aligns with the overarching objectives of the DEDALUS project, which seeks to expand residential participation in DR programs across Europe by bringing together key stakeholders and advancing smarter energy management strategies.”
More information:
Vasilis Michalakopoulos et al, A machine learning-based framework for clustering residential electricity load profiles to enhance demand response programs, Applied Energy (2024). DOI: 10.1016/j.apenergy.2024.122943
iCube Programme
Citation:
Machine learning framework boosts residential electricity clustering for demand-response (2024, October 4)
retrieved 4 October 2024
from https://techxplore.com/news/2024-10-machine-framework-boosts-residential-electricity.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.
