A New Intelligent Agent-Based AGC Design With Real-Time Application

Abstract

Automatic generation control (AGC) is one of the important control problems in electric power system design and operation, and is becoming more significant today because of increasing renewable energy sources such as wind farms. The power fluctuation caused by a high penetration of wind farms negatively contributes to the power imbalance and frequency deviation. In this paper, a new intelligent agent-based control scheme, using Bayesian networks (BNs), is addressed to design AGC system in a multiarea power system. Model independence and flexibility in specifying the control objectives identify the proposed approach as an attractive solution for AGC design in a real-world power system. The BN also provides a robust probabilistic method of reasoning under uncertainty, and moreover, using multiagent structure in the proposed control framework realizes parallel computation and a high degree of scalability. The proposed control scheme is examined on the 10-machine New England test power system. An experimental real-time implementation is also performed on the aggregated model of West Japan power system.