An Optimal Distributed MPC Scheme for Automatic Generation Control Under Network Constraints

Abstract

Traditional Automatic Generation Control (AGC) regulates an area control error (ACE) signal to control deviations in both the tie-line flows and system frequency. This approach can be suboptimal when the generation mix incorporates a high penetration of Renewable Energy Sources (RESs), and can lead to a biased control response under large supply and/or demand disturbance events. This paper proposes an alternative distributed Model Predictive Control (MPC) scheme which regulates the frequency deviations, and enforces thermal limit constraints on the tie-line flow deviations. The proposed State Constraint Distributed Model Predictive Control (SCDMPC) scheme achieves bias-free control and operation at the economically optimal operating point. The methodology is developed by considering system dynamics accounting for the primary generation sources and the tie-line power flow constraints. The new SCDMPC methodology reduces the regulating reserve requirement, enables a self-smoothing response to the supply/demand fluctuations between control areas, and improves the economics of AGC under high RESs penetration conditions.