A Machine Learning-Driven Residential Energy Management System with Smart Web Application Development

Domestic energy management self-discipline is necessary to address climate change and energy crises on a global scale. In order to study the issue of energy management, this paper proposes a smart Residential Energy Management System (REMS). By adjusting appliance loads to locally stored renewable energy according to grid availability and renewable resource energy levels, REMS reduces dependency on the grid. Decisions produced by machine learning systems automating this transition can be just as good as those made by humans. In the proposed work, XGBoost (XGB) machine learning-based approach is used. The REMS decision-making process is driven by a machine learning-based algorithm that takes into account a number of input parameters, such as battery state of charge, load, and the availability of energy sources. The XGBoost-based machine learning-based approach is compared with K-Nearest Neighbor (KNN), Support Vector Machine (SVM) and Decision Tree (DT). Algorithms are compared using several statistical measures and one machine learning algorithm which outperforms the others is selected and is used for further development. This machine learning model developed is further integrated in a web application which is based on Flask, HTML, and Python-based framework. The results obtained in this work recommend a suitable machine learning model for energy management in real-time with the help of web application.