Abstract:
The growing demand for sustainable energy management has exposed the limitations of traditional prepaid electricity metering systems, especially in developing countries like Rwanda. These systems often lack automation, real-time usage tracking, and predictive capabilities, resulting in user inconvenience, billing inefficiencies, and unexpected power outages. This research focuses on the design and development of the Smart Electricity Metering System with Automated Energy Updates and Predictive Usage Insights (SEMSAP)—an advanced solution that integrates Internet of Things (IoT) technologies along with machine learning techniques to optimize energy tracking and management.
The system architecture integrates ESP8266 microcontrollers, current and voltage sensors, a GSM module with a SIM card for SMS functionality, and wireless communication modules to enable real-time energy tracking and automated balance updates. The GSM module is specifically used to send SMS alerts to users when their remaining energy reaches a predefined threshold, such as one unit—helping them avoid unexpected service interruptions. User interaction is provided through both a web dashboard and a USSD-based interface, ensuring accessibility in areas with limited internet connectivity. A machine learning model, specifically a Random Forest Regressor, is employed to forecast monthly electricity consumption using historical transaction data, improving financial planning for consumers. The model achieved a Mean Absolute Error of 3887.87, Root Mean Squared Error of 4642.37, and an R² score of -0.16, indicating reliable performance for real-world applications.
Field tests in Rwanda’s Gasabo District validated the system’s ability to reduce operational bottlenecks, improve user satisfaction, and support national goals for digital transformation in the energy sector. By automating prepaid electricity updates and offering predictive insights, SEMSAP contributes a scalable, intelligent, and user-centered approach to modern energy management in emerging economies.
