Distributed Average Observation in Inverter Dominated Dynamic Microgrids

Consensus-based distributed average observation has been widely used in decentralized coordination among distributed generators (DGs) in inverter-based autonomous microgrids (MGs). However, its observation accuracy suffers from communication delays. Such observation errors are hard to detect and could deviate the system operation states (e.g., DG operating voltage) from their desired references. This work quantifies the relationship between communication delays and the steady-state observation errors of the delayed dynamic consensus algorithm. It is worth mentioning that system under study operates autonomously in the context of dynamic MGs. Compared to conventional MGs, dynamic MGs operate with reconfigurable cyber-physical networks. The observation errors are minimized by adopting the optimal system topology through cyber-physical network reconfiguration. Detailed control diagrams are designed to observe the average DG operating voltage in the physical networks, while measurements and control references are exchanged with communication delays through the cyber network. The proposed reconfiguration approach along with detailed controller designs are validated on a 12-bus test system in Simulink/MATLAB.