Long-Term Load Forecasting Incorporating GDP and Population Dynamics

This research paper seeks to create a dependable long-term load forecasting (LTLF) model for Nigeria that includes socio-economic indicators like Gross Domestic Product (GDP) and population trends. Accurate electricity demand forecasts help policymakers plan infrastructure expansion to meet demand increases due to population growth and socio-economic development. A multi-model approach integrating Linear Regression with Support Vector Machines and Artificial Neural Networks was used to reach this objective. The performance of these models was assessed through implementation and comparison on MATLAB and Python platforms to determine predictive capabilities. The Artificial Neural Network model achieved the best results among its counterparts by having the lowest RMSE and MAPE values. By employing a multi-model approach, the test results showed that this approach gave an RMSE less than 1000GWh compared to all forecast methods giving an RMSE less than 9000GWh and greater than 1500GWh thereby establishing it as the most dependable forecast technique with superior error metrics. The findings encompass electricity load predictions from 2024 through 2035 for every model.