Optimal ANFIS Based Automatic Generation Control for Steam Power Plant

Abstract

The growing complexity, nonlinear characteristics, and continuous load variations in modern power systems require advanced control techniques to ensure reliable operation and maintain frequency stability. Automatic Generation Control (AGC) plays a crucial role in balancing power generation with load demand and in maintaining system frequency within acceptable limits. Although conventional controllers such as PID and FLC have been widely used in AGC applications, their performance often decreases under nonlinear operating conditions and load disturbances. This can result in higher overshoot, longer settling time, and reduced robustness. This thesis proposes an Adaptive Neuro-Fuzzy Inference System (ANFIS)-based AGC scheme for a steam power plant. The proposed approach integrates the learning capability of neural networks with the reasoning ability of fuzzy logic for effectively handle system nonlinearities and uncertainties. A detailed mathematical model of single-area steam power plant, including the governor, turbine, and generator-load dynamics, is developed and implemented in the MATLAB/Simulink. The performance of the proposed controller evaluated under both steady-state conditions and different load disturbance scenarios, including load addition and load rejection. The results compared with those obtained PID and FLC controllers.