E3S Web Conf.
Volume 351, 202210th International Conference on Innovation, Modern Applied Science & Environmental Studies (ICIES’2022)
|Number of page(s)||5|
|Published online||24 May 2022|
Low frequency-based energy disaggregation using sliding windows and deep learning
1 Faculty of Sciences & Technology, University of Algarve, 8005-294 Faro, Portugal
2 SIGER, Faculty of Sciences and Technology, Sidi Mohamed Ben Abdellah, Fez, Morocco
3 IDMEC, Instituto Superior Técnico, Universidade de Lisboa, Lisboa, Portugal
4 CISUC, University of Coimbra, Coimbra, Portugal
5 LIPI, ENS, Sidi Mohamed Ben Abdellah, Fez, Morocco
* Corresponding author: firstname.lastname@example.org
The issue of controlling energy use is becoming extremely important. People’s behavior is one of the most important elements influencing electric energy usage in the residential sector, one of the most significant energy consumers globally. The building’s energy usage could be reduced by using feedback programs. Non-Intrusive Load Monitoring (NILM) approaches have emerged as one of the most viable options for energy disaggregation. This paper presents a deep learning algorithm using Long Short-Term Memory (LSTM) models for energy disaggregation. It employs low-frequency sampling power data collected in a private house. The aggregated active and reactive powers are used as inputs in a sliding window. The obtained results show that the proposed approach gives high performances in term of recognizing the devices' operating states and predicting the energy consumed by each device.
© The Authors, published by EDP Sciences, 2022
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.