Forecasting monthly peak demand in fast growing electric utility using a composite multiregression-decomposition model

Provision of reliable electric demand forecasts in a typical fast developing utility constitutes a unique problem. The problem becomes more complex with the cyclic movement of the electric demands of religious festivals occurring in Hijra months which are related to the lunar cycle. This is in addition to drastically changeable weather conditions, in particular ambient temperatures that influence the residential air-conditioning demand. A multiregression model has been used for estimating the contribution of previous religious festivals towards the system peak demand in Saudi Arabia. The historical system peak demand time-series data are thereafter adjusted to allow for the effects of these festivals. The Census II decomposition method is then applied to the corrected time-series data. A logistic model coupled with an optimisation technique has been used to represent the trend component. The peak demand forecast obtained by the proposed method, and two other standard classical forecasting techniques are compared with actual peak demand.<>