Decentralized optimal neuro-controllers for generation and transmission devices in an electric power network

Abstract: In this paper, the dual heuristic programming (DHP) optimization algorithm is applied for the design of two LOCAL nonlinear optimal neuro-controllers on a practical multi-machine power system. One neuro-controller is designed to replace the conventional linear controllers, which are the automatic voltage regulator (AVR) and speed-governor (GOV), for a synchronous generator. The other is a new external neuro-controller for the series capacitive reactance compensator (SCRC), flexible ac transmission systems (FACTS) device. The PSCAD/EMTDC® simulation results show that interactions of two DHP neuro-controllers with different control objectives improve the system performance more effectively compared to when each one operates without the presence of the other one. [Copyright &y& Elsevier]