A FUZZY NEURAL NETWORK APPROACH TO VAR CONTROL

Abstract

The present size of power systems and prevailing constraints create strenuous circumstances for system operators to correct voltage problems at any given time. In the past system operators maintained a reliable performance based upon their experience and on the spot assessment of the system conditions, but now power systems have grown very large and are becoming complex day by day. In the present situation it is beyond the system operator's analytical ability to take decisions of his own and definitely there is a need for decision-making.aids in such a predominantly fluctuating and uncertain environment. The capabilities of ANN have spurred a surge of interest in employing AT for the online solution of different power system problems. The modem reactive power control deals not only with the regulation of system voltages, but also with the coordinated operation of the- reactive, power sources within the system. This work presents a fuzzy neural network approach to the optimal reactive power (var) control problem. The method - incorporates the reactive load uncertainty in optimizing the overall system performance. The artificial neural network (ANN) enhanced by fuzzy sets is used to determine the memberships of control variables corresponding to the given load values. Test cases and numerical results demonstrate - the applicability of the proposed approach. Simplicity, processing speed and ability to model load uncertainties make this approach-a viable. option for on-line reactive power control.