Non-Intrusive Load Monitoring Based Smart Energy Management System For Smart Buildings

This paper presents a non-intrusive load identification and monitoring system using the aggregated power consumption measurement from residential homes. The main feature of the proposed system is to identify all the connected loads inside the house without individual sensors, smart meters, or appliance detection devices. Different neural network-based algorithms are implemented and compared on the real-time aggregated data of power consumption, and K-Nearest Neighbors is found as the best suitable algorithm for appliance detection. The Raspberry Pi-4 based hardware setup is used for a non-intrusive load monitoring (NILM) system. In addition, a mobile application is also developed for illustration of the results and control of home appliances. This article presents a quantitative research technique to show how the provision of disaggregated data to a user can lead to more responsible and improved electricity consumption. The research proposes a scalable and sustainable approach to load monitoring and effective power consumption.