Designing a self-constructing fuzzy neural network controller for damping power system oscillations.

Abstract This study presents a self-constructing fuzzy neural network (SCFNN) based static synchronous series control (SSSC) to mitigate the inter-area oscillations in interconnected power systems. The proposed intelligent system includes an on-line trained fuzzy neural network (FNN) controller with adaptive learning rates (ALRs) and self-constructing mechanism. The Lyapunov scheme is employed to obtain the adaptive learning rates. Therefore, convergence of the suggested controller can be ensured. The proposed approach is such that, at first, originally, no neurons exist in the structure of FNN. In fact, they are automatically created and if it is required, they will be created. Therefore, the total time for training algorithm is significantly reduced and the speed of controller is considerably increased. In addition, the Prony technique is utilized to guesstimate the damping ratio of oscillations. The results confirm the usefulness of the suggested controller. [ABSTRACT FROM AUTHOR]