Estimation and Forecasting Electricity Load in Benin: Using Econometric Model ARIMA/GARCH.

Abstract

In order to help governments in energy development programming and also public service operators and network managers to have better planning for managing electricity demand and design better operational planning on production units and distribution networks, it is necessary to make the long-term, prediction, estimation and evaluation of the electrical load. The aim of this work is to propose the econometric model to estimate and forecast the electricity load in Benin for a long term, until 2030. It is important to notice that due to the complexity and multiple parameters considered for the forecasting, the use of single model will lack of accuracy and the results will not be conform to the reality. In this paper we propose an hybrid model ARIMA/GARCH, a non-linear model that combines a linear model of autoregressive integrated moving average (ARIMA) and a non-linear model, generalized autoregressive conditional heteroscedasticity (GARCH). This model is applied to obtain a non-linear relationship between load variation and determinants such as demographic change, gross domestic product GDP and weather parameters for an accurate demand forecasting.