ELECTRICITY DEMAND FORECASTING TOWARDS VISION 2016 FOR BOTSWANA USING DECOMPOSITION (MULTIPLICATIVE) TIME SERIAL MODEL

A. Obok Opok, G.O. Anderson, and K.M. Yanev

Keywords

Electricity demand forecast, decomposition time series model

Abstract

Demand forecasting for electricity falls under short-term, mediumterm or long-term forecast timeframe regimes. Depending on the type of forecasting, autonomous models and conditional models may be applied, taking into account the rate of growth of the utility, which can be, slow or fast. This paper is about the application of autonomous models, which embraces both time series model and stochastic models (smoothing and decomposition) to forecast for average net hourly demand and system maximum net hourly demand for electricity in Botswana in the year 2016. The growth of the electricity demand in Botswana is considered fast. The year 2016 is selected to coincide with the national vision target year, for Vision 2016 national goals. Historical data for average net hourly demand and system maximum net hourly demand covering 5 years, from 1997 to 2002 were used. The results indicate that the margin of error achieved in the forecasts is ±0.09%.

Important Links:



Go Back