Neural network based predictive Automatic Generation Control

The NERC's Control Performance Standards (CPS) represent a great flexibility in relaxing the control of generating resources and yet assuring the stability and reliability of interconnected power systems. The design enhancement of Automatic Generation Control (AGC) plays a vital role in meeting these challenges. This paper for the first time provides a mathematical formulation for AGC in the context of meeting the NERC control performance standards and integrating renewable generating assets. In addition, this paper proposes a neural network based predictive control approach for AGC. The proposed controller is capable of handling complicated nonlinear dynamics in comparison with the conventional Proportional Integral (PI) controller. Furthermore, a coordinated control policy is proposed: the neural controller is responsible to control the system generation in the relaxed manner to achieve the desired control performance.