Real-Time NILM-Based Smart Energy Monitoring Using AI and IoT

Abstract

In this paper, a real-time smart energy monitoring system using Non-Intrusive Load Monitoring (NILM) and Artificial Intelligence (AI) for appliance level energy estimation from aggregated electrical signals is proposed. The proposed system uses a single clamp current sensor and an ESP32 microcontroller to measure the house current, instead of using several sensors for each individual appliance. The collected data is sent to a cloud-based platform, where a lightweight AI model identifies the appliances in real-time based on the electrical features extracted from the data, including RMS current and estimated power. The system also dynamically calculates the energy consumption and electricity cost, allowing users to monitor their energy usage and make informed adjustments to optimize energy consumption. A web-based interface for real-time visualization, appliance-wise analysis and cost estimation. Experimental results show high classification accuracy up to 97% with low computational complexity that is suitable for IoT environments. The proposed system is scalable, cost-effective, and non-intrusive approach for smart energy management.