An improved on-line neural network controller for multi-variable control of UPFC

Abstract

This paper proposes an improved on-line neural network (OLNN) controller by incorporating changes in neural network architecture of indirect-inverse identification neural network controller (IIINNC), which was proposed earlier for controlling power flow, AC bus and DC link voltages of unified power flow controller (UPFC). A new learning algorithm has been derived for the proposed OLNN controller. The proposed OLNN controller requires less training and validation compared to that of the IIINNC. The architecture and control algorithm of the proposed neural controller reduces the complexity and latency time in comparison to the IIINNC. Simulation studies carried out demonstrates the applicability of the proposed on-line neural controller for the multi-variable control of UPFC.