In this article, an improved robust control strategy for voltage stabilization and desired performance satisfaction of islanded inverter-interfaced microgrids consisting of several distributed generations (DGs) with general topology is presented. The main advantages of the proposed control manner are as follows: 1) its structure is fully decentralized, 2) the design process is scalable, 3) it does not impose any limitations on the microgrid parameters and line dynamics, 4) it provides stability and the desired performance of nominal microgrid system, 5) it maintains robust stability as well as robust performance of the closed-loop system against microgrid topology changes, plug-and-play (PnP) operation of DG units, and load unmodeled dynamics, and 6) each local controller is obtained from a unique convex optimization problem which results in optimal performance of the system and also robustness to several sequential changes. To achieve these objectives, first, each DG subsystem is modeled as a linear time-invariant (LTI) system affected by disturbances caused by the local load current and the load voltages of its neighboring DGs. Next, the PnP functionality of DGs and the microgrid topology changes are modeled as a new polytopic-type uncertainty. Thereafter, the design problem is transformed into a dynamic output feedback controller for an LTI system subject to polytopic-type uncertainty with ${\bm{H}_\infty }$ performance criteria. Finally, a convex linear-matrix-inequality-based optimization problem with a noniterative direct synthesis process is proposed to solve the controller design problem. The performance of the presented controller is appraised via several simulation studies accomplished in MATLAB/SimPowerSystems Toolbox.
