Power Sharing of MultiMicrogrids: Experimental Validation of Droop Transients
Ph.D. Student: Gurupraanesh Raman  Advisor: J. C. H. Peng  Project Duration: 20172018
The smallsignal stability of distribution networks with multiple droopconnected distributed generation depends on the droop constants, network parameters as well as the number of interconnected sources. The damping of the critical system modes can change frequently in realtime and the Distribution Management System (DMS) is expected to maintain the grid stability through supervisory control.
Figure. Threeinverter microgrid setup at PEng Lab. Each inverter is controller by FPGA, and the supervisory system is implemented in LabView.
Figure 2. Control schematic of distribution system with multiple droopconnected interfaces.
Figure 1. Conventional droop control law for each inverter or conventional generator.
Communication is essential to maintaining the realtime stability such distribution systems. Therefore, the communication links between the supervisory controller and the various sources in the network are vital, with some nodes more critical than others from the perspective of smallsignal stability. This work describes a method to grade the impact of supervisory control inputs in a multimicrogrid system using a novel sensitivity measure. The central idea behind the proposed sensitivity criterion is to make it scalable to the number of nodes and to reduce the computational complexity as compared to the conventional numerical sensitivity analysis methods.
The sensitivity measure is obtained by deriving a simplified secondorder equation for the angular dynamics of each source by assuming that the voltage dynamics are much faster. This is justified in practical systems where the value of the voltage droop coefficient is much larger than that of the frequency droop coefficient. Once the explicit angular dynamics equation is obtained, the sensitivity can be calculated. The sensitivity in this work pertains only to the frequency droop coefficient, whose effect on the damping of the critical mode is considered more significant than that of the voltage droop coefficient. For deriving the generalized sensitivity criterion, any multisource system is approximated as a singlesource system with suitable approximations. This makes the evaluation of the sensitivity scalable unlike conventional numerical methods.
Simulation and hardware experiments performed on the IEEE 123 distribution test feeder validate the proposed sensitivity measure. The IEEE 123 bus system is viewed as consisting of 5 microgrids, with the interfaces located at buses 18, 151, 197, 94 and 152. Labeling them as Nodes 15, the sensitivity criterion indicates that the Node1 is the most critical node, and that supervisory control inputs to it will elicit the most change in the damping, which is verified by simulation experiments shown below.
Figure 3. Application of the proposed sensitivity criterion to the IEEE 123 node test feeder.
Figure 4. Demonstration that supervisory control inputs to most (least) sensitive nodes causes the highest (lowest) improvements in damping from the same initial condition (Ϛ=0.0380).
Hardware experiments were then conducted on a 21 kW microgrid testbench, and it was verified that the impact of supervisory inputs could be effectively assessed using the proposed supervisory criterion.
Applications
The sensitivity measure can be used for:

assessing the criticality of the communication links

evaluating resource placement options

determining optimal contingency response sequences
The evaluation of sensitivity does not entail computationally expensive eigenvalue calculations and can therefore be augmented into the existing DMS system at no significant additional burden.