Stabilization control for multi-machine power system by nonlinear state feedback control using neural network

Abstract

Previously, the authors have reported on nonlinear state feedback control for synchronous generators in power systems. However, the nonlinear controller has not been implemented in real systems, only in simulations, because it takes the load angle information as the input of this controller. This paper presents nonlinear excitation control for improving electric power system transient stability using a neural network (NN). The NN models the nonlinear excitation controller using only measurable state variables from the synchronous generators in power systems. Therefore, the proposed method can realize nonlinear excitation control which does not require load angle information of the synchronous generator. On the other hand, gains of the nonlinear excitation controllers modeled by NN are optimized by using a genetic algorithm (GA). This gain tuning method using a GA can decide the gains of each nonlinear excitation controller in a large power system at once.