New model for peak demand forecasting applied to highly complex load characteristics of a fast developing area

Models have been developed for analysing and forecasting the monthly peak demands of a rapidly growing power system with highly complex load characteristics. The complexity of the load characteristics arises from the interaction of the static weather-sensitive load component and the cyclically moving load component associated with special events which occur annually. In the models, historical demand data are smoothed, using an exponential seasonal smoothing technique with optimal parameter values, to reduce random fluctuations. The smoothed load data are then decomposed into deterministic and stochastic load components. The deterministic load component consists of a long term trend component and a seasonal temperature component. Comparison of forecasts with actual demands has shown the adequacy of the model for analysing and predicting the demands of the fast growing power system considered in this paper.