WITHDRAWN: Analysis, design and implementation of hardware neural network based rotor oscillations compensator

Abstract

In this paper, an Artificial Neural Network (ANN) based Rotor Oscillations Compensator (ROC) is proposed and its equivalent Hardware Realizable Neural Network (HARNN) has been developed on the Field Programmable Gate Array (FPGA) board. The basic function of a ROC is to minimize the low frequency electro mechanical oscillations created in the Electric power network. Whenever the fault is occurred outside the operating range of a fixed gain ROC, the system damping is insufficient and the machine is still under unstable mode. Hence a hardware having sufficient “intelligence” that is capable of producing necessary and sufficient damping to the system is proposed in this paper. A detailed study has been done with the plant response without any controller, with a fixed gain controller and a FPGA based ANN controller. The studies revealed that when an intelligent controller has been included in the Exciter loop, the oscillations have been eliminated more effectively. But it is also observed that the realization of an ANN absolutely on an FPGA is having its own design limitations. The ARTIX 7 FPGA development board has been used for the complete analysis and the results and outcomes discussed at the end.