Recursive short-term load-forecasting algorithm

Abstract

The problem of forecasting the integrated-load demand of an interconnected power system is studied. It is proposed that the harmonic-series model for the weekly load pattern be expressed in the state-variable form. This then permits the use of Kalman-filtering techniques, for load-prediction purposes. It is shown that this approach leads to an improvement in the forecasting reliability, without increasing the online computational requirements. The results of a particular study are also presented.