Power system load forecasting using smoothing techniques

This paper deals with short-term load forecasting problem for a power system, The load demand at any particular instant is assumed to follow a time-Beries model. A predictor is developed which identifies the coefficients of the time series in an on-line fashion. The loads at previous instants are considered as delayed variables, and a smoothing estimation technique is made use of in conjunction with the concept of the two-stage estimator to simultaneously estimate the load and related parameters. The results of an actual study made on a receiving station in Northern India are presented to illustrate the proposed predictor.