Finite time multi-agent coordination of distributed generation for grid reactive support

Abstract

Inverter Interfaced Distributed Generation (DG) in the Smart Grid has the potential to contribute to the reactive power support of the overall power system. However, because of the large numbers and distributed nature of the DG units, using a fully centralized communication structure to achieve coordination can be prohibitive. A decentralized coordination approach is a good candidate solution to address this problem. In the literature, asymptotic consensus based algorithms have been proposed, in order to coordinate a set of DG units so that they collectively provide a certain reference reactive power. This paper presents an alternative decentralized coordination methodology that achieves the same objective in finite time. The protocol is based on linear iterative updates and known observability results from graph structured linear systems. In this paper, the methodology is customized to solve the reactive support coordination problem from distribution-connected inverters and it is modified to reduce the number of operations per step, ensuring applicability to a large distribution network. The IEEE 37-node test feeder is used as a test system, with added inverter interfaced generation in each node. For this sample system, the proposed approach is shown to coordinate the nodes faster than the asymptotic consensus approach.