Pinning-based Distributed Predictive Control of Secondary Voltage for an Islanded Microgrid Yi Yu * Guo-Ping Liu *,** Wenshan Hu * Hong Zhou * * School of Electrical Engineering and Automation, Wuhan University, Wuhan, 430072, China, (e-mail: yiyu.py@whu.edu.cn) ** School of Engineering, University of South Wales, Pontypridd CF37 1DL, U.K., (e-mail: guoping.liu@whu.edu.cn) Abstract: In this study, a pinning-based distributed predictive control of secondary voltage is proposed for an autonomous microgrid (MG) with communication constraints. The proposed predictive control is fully distributed, which requires the information of each distributed generator (DG) on an islanded microgrid and that of its neighbors. In particular, the predictive control scheme can compensate for the communication constraints actively rather than passively. Moreover, it could reduce the computation burden of the controller owing to the developed pinning-based control scheme. It is worthwhile to mention that the purpose of voltage restoration based on distributed predictive control is achieved by minimization of the utility function, which not only ensures the feasibility of the control input but also restores the voltage of an autonomous microgrid to a prescribed level simultaneously through an introduction of tracking and coordination cost. Finally, simulation results are presented to validate the effectiveness of the proposed control methodology. Keywords: Distributed predictive control, islanded microgrid, pinning control, communication delays. 1. INTRODUCTION With the growing concerns on the environment, more renewable energy resources (RERs) have been adopted in industry and living supplies. Another main reason for the introduction of RERs is that it is inexhaustible and economical for supplying electric power to the remote district. The power systems that integrate varieties of RERs are called microgrids. A microgrid can operate in grid-connected and islanded modes. Generally speaking, a grid-connected MG means that it is connected with the power grid, whose dynamic characteristics obey the assigned control command of the big grid and without being dominant by any other control. Therefore a MG operating in this mode is similar to an energy storage device or a load, which imposes no challenges of grid control. However, a MG operating in the islanded mode needs to be controlled artificially for stable operation and reliable power outputs. Moreover, the inertia of an islanded MG is too small inertia to consider, comparing with that of the big grid. In this case, the small power system is very sensitive to disturbances and load changes. Hence it is crucial to develop valid control strategies for the islanded MG system. To form a microgrid for the grid connection of local distributed generators, the hierarchical control scheme is a commonly adopted framework (Guerrero et al., 2011), which generally consists of primary control, secondary ? This work was supported in part by the National Natural Science Foundation of China under Grants 61773144 and 61690212. control, and tertiary control. Droop control algorithms that respond quickly without relying on communication plays the role of primary control, which mimics the par- allel operation characteristics of synchronous generators (Zhong and Weiss, 2011). However, droop control can result in voltage and frequency deviations between the output and the reference values, which may not generally be within the permitted range. Therefore, secondary con- trol is demanded to eliminate the deviations. At present, a large body of literature (Bidram et al., 2013, Lu et al., 2018, Aryani and Song, 2019) studies the voltage and frequency secondary control of MG based on the theory of multi-agent consensus, which realizes the interconnection of multiple DGs in this stage. In the cooperative secondary control framework, distributed controllers usually collect the operation and status information of neighboring DGs through the network to adjust their own power generation behaviors. This control scheme can be efficient and robust. Obviously, the communication system plays a crucial role in the cooperative control of microgrid outputs. However, once the communication system is introduced, no mat- ter wired communication or wireless communication, or even with the currently most advanced 5G communication techniques, communication constraints are ubiquitously caused by the limits of bandwidth, traffic congestions et.al. Therefore, considering the reliability of MG, it is necessary to study the voltage and frequency synchronization of microgrid with communication constraints using cooper- ative control theory. Many efforts have been devoted to this topic (Lai et al., 2019, Zhang and Hredzak, 2019). Preprints of the 21st IFAC World Congress (Virtual) Berlin, Germany, July 12-17, 2020 Copyright lies with the authors 13071