MAS-Based Distributed Coordinated Control and Optimization in Microgrid and Microgrid Clusters: A Comprehensive Overview

The increasing integration of the distributed renewable energy sources highlights the requirement to design various control strategies for microgrids (MGs) and microgrid clusters (MGCs). The multiagent system (MAS)-based distributed coordinated control strategies show the benefits to balance the power and energy, stabilize voltage and frequency, achieve economic and coordinated operation among the MGs and MGCs. However, the complex and diverse combinations of distributed generations (DGs) in MAS increase the complexity of system control and operation. In order to design the optimized configuration and control strategy using MAS, the topology models and mathematic models such as the graph topology model, noncooperative game model, the genetic algorithm, and particle swarm optimization algorithm are summarized. The merits and drawbacks of these control methods are compared. Moreover, since the consensus is a vital problem in the complex dynamical systems, the distributed MAS-based consensus protocols are systematically reviewed. On the other hand, the communication delay issue, which is inevitable no matter in the low- or high-bandwidth communication networks, is crucial to maintain the stability of the MGs and MGCs with fixed and random delays. Various control strategies to compensate the effect of communication delays have been reviewed, such as the neural network-based predictive control, the weighted average predictive control, the gain scheduling scheme, and synchronization schemes based on the multitimer model for the case of fixed communication delay, and the generalized predictive control, networked predictive control, model predictive control, Smith predictor,
$H_{\infty}$
-based control, sliding mode control for the random communication delay scenarios. Furthermore, various control methods have been summarized to describe switching topologies in MAS with different objectives, such as the plug-in or plug-out of DGs in an MG, and the plug-in or plug-out of MGs in an MGC, and multiagent-based energy coordination and the economic dispatch of the MGC. Finally, the future research directions of the multiagent-based distributed coordinated control and optimization in MGs and MGCs are also presented.