Adaptive RTRL based hybrid controller for series connected FACTS devices for damping power system oscillations

This paper presents a novel design of a co-ordinated controller for series connected FACTS devices like Thyristor Controlled Series Capacitor(TCSC) and Thyristor controlled Power Angle Regulator (TCPAR). The scheme can be used for non-linear system control, in which the exact linearized mathematical model of the system is not required, can be used to control many FACTS devices with a single controller. The basis of the proposed design is the Real Time Recurrent Learning (RTRL) algorithm in which the Neural Network (NN) is trained in real time. This requires two sets of neural networks. The first set is a fully connected Recurrent Neural Network (RNN) which acts as a neuro-identifier that provides the dynamic model of the system. The second set of neural network is the neuro-controller, used to generate the required control signals for the thyristors. Simulations results of the system using MATLAB/SIMULINK show that the performance of the system with the proposed controller is better than the conventional PI controllers and GA-based PI controllers.