ANFIS based automatic voltage regulator with hybrid learning algorithm

Abstract

In this paper, the authors propose a design methodology of adaptive neuro-fuzzy inference system (ANFIS) based automatic voltage regulator (AVR) using hybrid learning algorithm to improve the small-signal performance of power system. Here, a zero order Sugeno fuzzy model is considered, whose parameters are tuned off-line through hybrid learning algorithm. This algorithm is a combination of least square estimation and error backpropagation method, where the least square method is applied for the tuning of linear output membership function parameters and the backpropagation method is used to tune the nonlinear input membership function parameters. The proposed method is verified through digital simulation with a single machine infinite bus system. It is found that the AVR is performing well in restoring the terminal voltage instantaneously and the damping characteristics of the rotor angle are also improved. The simulation results establish that the design of ANFIS based AVR employing hybrid learning algorithm can be very useful in small signal stability of power system.