AN INTERNET OF THINGS-ENABLED SMART ENERGY MONITORING AND LOAD FORECASTING FRAMEWORK FOR POLYTECHNIC CAMPUSES IN SOUTHWEST NIGERIA: DESIGN AND SIMULATION-BASED EVALUATION

Abstract

Polytechnic campuses in Southwest Nigeria manage electricity, one of their largest controllable
operating costs, largely without data, relying on aggregate billing and manual meter reading that
conceal where and when energy is consumed. This study designed an Internet of Things (IoT)-
enabled smart energy monitoring and load forecasting framework adapted to the infrastructural
constraints of these campuses and evaluated its forecasting and economic potential through a
documented simulation experiment. Guided by the design science research methodology, the study
specified a four-layer architecture comprising non-invasive metering nodes, connectivity-tolerant
wireless networking, an edge-and-cloud analytics pipeline and administrator-facing applications.
To evaluate the forecasting module in advance of field deployment, hourly demand for four
campus zones over an 89-day academic period was simulated from documented parameters
reflecting Nigerian campus load behaviour, including academic calendars, weekend effects,
examination-period loading and grid-to-generator changeover events. Four forecasting models
were trained on a 75-day partition and tested day-ahead on a 14-day hold-out. Random Forest
achieved the best accuracy in three of four zones, with mean absolute percentage error (MAPE)
between 8.45% and 13.23%, outperforming the seasonal naive benchmark in every zone; the
artificial neural network followed closely, while SARIMA was competitive only for the regular
hostel profile. Scenario analysis indicated that curtailing 15% to 25% of out-of-hours consumption
in non-residential zones would avoid 8,779 to 14,632 kWh per academic period, worth
approximately ₦7.97 million to ₦13.28 million annually at blended supply costs. The framework
offers a validated, context-sensitive pathway to continuous energy measurement and forecasting
in resource-constrained tertiary institutions.