Power System State Forecasting UsingA rti cial Neural Networks

Abstract

This paper presents a new method for power system state forecasting using artificial neural networks (ANN). The state forecasting problem has been solvedin two steps: the filtering step and the forecasting step in an open loop configuration. Because under normal operating conditions the power system behavesin a quasi-static manner, a simplified model of the dynamic behavior of thepower system states is considered. Two different ANN models have been usedfor these two steps of power system state forecasting problem. For the altering step, a functional link network (FLN), and for the forecasting step, a time de-lay neural network (TDNN) have been used to simulate the dynamic behaviorof the power system states. The proposed method has been tested on two IEEEtest systems, and a practical Indian system and results have been comparedwith an extended Kalman filter (EKF) based technique [Leite da Silva et al.,1983].