AI-IoT Based Smart Energy System for Multi-Unit Residential Buildings

Abstract

The growing electricity demand, coupled with challenges such as energy wastage, biased billing in multi-unit buildings, and the absence of adequate predictive energy management, necessitates intelligent solutions.This paper presented the development of a smart energy system tailored for multi-unit residential buildings.By integrating IoT technology with a trained LSTM machine learning model, the system enabled real-time energy monitoring, control, and hourly prediction of energy consumption.Core components include dual PZEM004T sensors, an ESP32 microcontroller, a keypad, an LCD, and relays, all managed via the Blynk IoT platform.The system performed key functions such as threshold-based relay switching, overvoltage and overcurrent protection, and AI-powered forecasting.Results demonstrated high accuracy in monitoring, responsive control through local and remote interfaces, and effective prediction with a low Mean Squared Error (MSE) of 0.0229.The solution ensured fair energy billing, reduced waste, and supported sustainable energy practices.