Event-Triggered Distributed Generation Control Architecture for Islanded AC Microgrids
Stanton T. Cady with adviser A. Domínguez-García
This project develops and implements a distributed control architecture for islander microgrids which replicates the functionality of the centralized control utilized in bulk power systems, i.e., frequency regulation and optimal dispatch. We assume each generator is equipped with a local controller and a wireless transceiver for information exchange with neighboring generators. Relying on this communication network and simple computations executed by the local controllers, we utilize distributed, iterative algorithms that allow the generators to determine their output commands in order to regulate the system frequency and minimize total operational costs while accounting for output constraints. The distributed control architecture is event-triggered, i.e., generator commands can be updated at non-fixed time instants. This allows a faster return to operation at nominal frequency.
The 6-bus microgrid shown in Fig. 21 is implemented in the lab using three synchronous generators rated at 2-hp and 230 Vll. An Arduino microcontroller outfitted with an XBee wireless transceiver, which acts as the local controller, is connected to the prime mover of each generator. After a load change, one local generator controller measures the frequency error and determines how much torque to add or remove from the system to return the frequency to the desired value. The local controllers then execute the distributed algorithm which serves to disperse the torque error among all the system generators. After reducing the frequency error to within a specified bound, the generators are dispatched so as to minimize overall cost.
Figure 22 illustrates the torque supplied by each of the three system generators as well as their mechanical speed after a load increase. The control architecture effectively returns the frequency to the desired value of roughly 1200 RPM (equivalent of 60 Hz) and adjusts each generator output to minimize the overall costs.
This work is funded by the National Science Foundation (NSF) under grant ECCS-CPS-1135598.