Improving the resilience of the electric distribution control infrastructure is an urgent issue not only to reduce the vulnerability of the power grid to extreme events but also to facilitate the use of green or renewable energy. Among different options, microgrids are an emerging and promising paradigm. A microgrid is a small-scale network designed to supply electrical and heat load of a small community (e.g., a shopping center, an industrial park, or a university campus). It typically contains a set of distributed generators, load, storage and protection devices that are controlled by a central controller. To the main power grid, a microgrid is a single entity that may either produce power for the larger grid or consume power. This dramatically simplifies the interaction between a microgrid and the main grid, allowing for the easy introduction and rapid evolution of new forms of computer control of heterogeneous technologies within the micro-grid domain. However, microgrids with a significant renewable energy component may experience rapid changes in power generation. It is extremely important to achieve fast voltage and frequency control during these swings otherwise the system may lose balance between load and generation and may eventually collapse. Control of microgrids relies on fast, reliable communications; in the most stringent cases messages must be delivered within a few milliseconds. This project will integrate Software Defined Networking (SDN) technology with microgrid control to stabilize and optimize system operation. The project will demonstrate the techniques on a microgrid at the University of Connecticut Depot Campus.
For this project SDN technology will have three general uses: route reconfiguration to avoid network performance issues and improve network reliability, packet prioritization to provide Quality of Service (QoS) in support of low latency delivery, and monitoring of data flows associated with microgrid control. The project will use a two phase approach to demonstration and evaluation both in close collaboration with UCONN Facilities Operations and Schneider Electric, the microgrid contractor. In the first phase it will use a virtual hardware-in-the-loop environment that receives real-time measurements from the microgrid and closely mimics the actual microgrid status. Specifically, the environment will contain real communication infrastructure and (small-scale) energy sources and load, while the rest of the system will be through trace-driven simulation using real-time measurements fed from the operational microgrid. Once the first phase has demonstrated and evaluated the techniques the PIs will work with Schneider Electric and UCONN Facilities Operations to incorporate the developed techniques into a larger staging facility for further evaluation and refinement and possible adoption by the Depot microgrid.
