7th International Conference on Probabilistic Methods Applied to Power Systems - PMAPS 2002, Naples (Italy). 22-26 September 2002
Summary:
In this paper a new methodology to forecast the day ahead electricity market behaviour is presented. This behaviour can be easily modelled by means of the so called residual demand curves (RDC's) . The pattern of these curves (as the spot market is an hourly market there is one RDC for each hour) changes greatly according with the type of the day (labour-non labour) and the hour (peak, valley, plateau,...) so this fact must be taken into account. Firstly, a classical ARIMA analysis without explanatory variables is carried out. Afterwards, adequate explanatory variables are searched in order to build a more accurate Transfer Function Model. Next a new procedure called weighted estimation is developed and the differences between these two methods are pointed out. Finally, a case study is presented in order to check the validity of the weighted estimation model.
Keywords: Spot Market, Residual Demand Curves (RDC's), ARIMA models, TF models, weighted estimation, explanatory variables, Market behaviour estimation.
Publication date: 2002-09-22.
Citation:
A. Martín, J.I. de la Fuente, New forecasting method for the residual demand curves using time series (ARIMA) models, 7th International Conference on Probabilistic Methods Applied to Power Systems - PMAPS 2002, Naples (Italy). 22-26 September 2002.
