Home energy monitoring and alert system using ESP32 and CT sensors for real-time power tracking

Ling, Ivan Xin Yu (2026) Home energy monitoring and alert system using ESP32 and CT sensors for real-time power tracking. Final Year Project, UTAR.

PDF Download (6Mb)

Abstract

The increasing cost of electricity and a growing global emphasis on sustainable energy consumption have highlighted a critical need for real-time energy visibility within residential households. This project addresses this need by developing a cost-effective, non-invasive and Internet of Things (IoT) system for home energy monitoring and alerting. The system architecture leverages an ESP32 microcontroller, a high-precision ADS1115 Analog-to-Digital Converter (ADC), and a non-invasive SCT-013-000 Current Transformer (CT) sensor to accurately capture and process real-time electricity usage data. This data is transmitted securely over WiFi via the Message Queuing Telemetry Transport (MQTT) protocol to a robust, Docker-containerized backend. The backend consists of a Mosquitto MQTT broker, a Telegraf data agent, an InfluxDB Time-Series Database (TSDB), and a Grafana visualization platform. The resulting system provides users with an intuitive, dynamic dashboard displaying live power metrics, historical trends, and precise cost estimations based on Malaysia’s tiered tariff structure. Furthermore, the system incorporates an intelligent alerting mechanism, sending notifications via a Telegram bot upon detection of anomalous power consumption. The bot also facilitates user interaction by serving on-demand dashboard panel snapshots, providing unparalleled access to energy insights. To move beyond simple alerts, the system now incorporates a Random Forest (RF) Machine Learning model. This allows for context-aware anomaly detection by learning cyclical household habits and setting dynamic power thresholds. Furthermore, an interactive Telegram bot facilitates bidirectional communication that allows users to fetch real-time Grafana snapshots and conduct smart energy audits via an inline calendar user interface (UI). This report details the design and successful validation of this intelligent system and proving its effectiveness as a robust tool for proactive energy management and machine learning-based energy analytics. The complete source code for this project are open-source and available at https://github.com/IvanLXY04/Home-Energy-Monitoring-and-Alert-System.

Actions (login required)