IMPROVING THE ECONOMETRIC-BASED LOAD FORECASTING IN NIGERIAN POWER SYSTEMS USING EXPONENTIAL SMOOTHING MODELING TECHNIQUE

Abstract

Econometrics load forecasting is a blend of economics, statistics and mathematics in the forecast of electricity power load demand. It is gaining recognition in areas such as generation, network expansion planning, and formulation of the tariff structure. The use of Load Forecasting Models in Power System Networks has continued to inspire researchers to seek ways of realizing accurate models for Load Forecasting. In this work, a procedure for improving econometric-based load forecasting in Nigeria power systems using exponential smoothing modeling technique is proposed. In this work, three types of exponential smoothing namely, Single Exponential Smoothing, Double Exponential Smoothing and Triple Exponential Smoothing were examined. Mathematical modeling of the exponential smoothing technique is presented. A Simulink model is then developed in MATLAB based on the mathematical model derived to show real time behaviour when simulated for use in the power load forecasting process. The Simulink model so developed was simulated to yield forecast results in the categories of years as Residential Load Demand (RLD), Commercial Load Demand (CLD), Industrial Load Demand (ILD) and Total Load Demand (TLD), all in Megawatts of electricity. The forecast covered a period of 22 years. Computations of the forecast improvement percentage were done. The results were tabulated and MAPE was used as a metric to validate the result of the simulation. A MAPE value of 2.109% was realized thus validating the veracity of the technique applied.