Decentralized Robust Control of a Network of Inverter-Based Distributed Generation Systems

Abstract

This paper presents the design of decentralized robust controllers for a network of inverter-based distributed generation systems with LC filters in the scale of a nanogrid. Using overlapping decomposition, the network of inverters is clustered into several subnetworks such that all inverters within a subnetwork are strongly coupled and there is no or a weak coupling effect between any two inverters from different subnetworks. For the inverters within the same subnetwork, decentralized robust controllers are designed sequentially in the μ-synthesis framework. In addition, all controllers are designed to be robust against ±10% variations in the LC filter parameters. To assess the performance of the proposed sequentially-designed controllers and compare it to that of the benchmark independently-designed ones, the distances between two neighboring inverters from the same and different subnetworks are considered to be 200 (m) and 800 (m), respectively. In this case, time-response and robustness analysis results illustrate the superiority of the proposed sequentially-designed controllers in the overlapping decomposition framework over the benchmark independently-designed ones. Moreover, transient overload and nonlinear load analyses demonstrate that the proposed sequentially-designed decentralized controllers are able to keep the load voltage within ±10% of the nominal value and the harmonic voltage distortions to less than 4%.

Keywords

• networked system
• overlapping decomposition
• distributed generation
• <i>μ</i>-synthesis
• robust controller
• sequential design