Argonne Research Projects Will Help Optimize Grid

The nation’s electric power grid is becoming more complex. As the system incorporates more sources of renewable energy, such as solar and wind power, utilities need new ways to manage it more efficiently. The U.S. Department of Energy’s (DOE’s) Argonne National Laboratory is responding to that need. Argonne will participate in three projects that recently received multiyear, multimillion-dollar awards from DOE’s Advanced Research Projects Agency – Energy (ARPA-E) to develop new computer algorithms to optimize the grid through the Network Optimized Distributed Energy Systems (NODES) program and to develop accurate models on which to test those algorithms through the Generating Realistic Information for the Development of Distribution and Transmission Algorithms (GRID DATA) program. Until recently, the grid has been built mainly of centralized fossil-fuel-powered or nuclear plants that generate power and send it over long transmission lines to where it is needed. Newer distributed generation sources like solar or wind, however, are not centralized, are usually smaller than traditional sources and have grown significantly in recent years.
“With distributed generation booming, the power system operator is facing a problem: How to utilize those distributed resources as efficiently as you can,” says Jianhui Wang, Section Manager of Advanced Power Grid Modeling in Argonne’s Energy Systems Division.
What’s needed is something like an orchestra conductor to interweave all the power sources and get the different instruments playing their parts, when needed, and resting when not.

Argonne Research Projects

Wang is on two teams — one with Northwestern University and one with Eaton Corporation — that have received two related ARPA-E NODES awards to develop control algorithms that will play the role of the conductor. It is a delicate task. If there is a mismatch between generation and demand, the frequency changes, causing problems for grid stability. But demand and generation fluctuate constantly. And power output from solar and wind plants cannot be fully controlled due to the variability of the wind and solar resources. The whole system must be balanced in real time – from second to second. Wang’s projects would help achieve the balance. One algorithm would control the generation end. Once power has been transmitted, another algorithm would manage the decentralized distributed generation resources that send power out to consumers and meet the demand locally. “We want to come up with a coordinated, hybrid architecture. So it has both the components,” said Wang. Without good models of the grid to test their algorithms, developers can’t know if the algorithms will work in the real world. But the data available to build such models is out of date, and current data from power companies is proprietary and often unavailable to researchers. “We wanted new data sets that were realistic and large-scale and would be able to convince industry that, yes, these algorithms work on real systems,” said Daniel Molzahn, a computational engineer in the Center for Energy, Environmental, and Economic Systems Analysis at Argonne. Molzahn, Wang and a team from the University of Wisconsin at Madison, ComEd and two software vendors recently received an ARPA-E GRID DATA award to construct data sets to model electric grids representing several regions. They will use information about population density, land usage and industrial and commercial energy consumption patterns to estimate demand for electricity in a city or region. Also included will be information about where transmission lines are connected, the electrical properties of those lines, where generators are located and what their capabilities are. “It’s everything you would need to know to be able to simulate the power system,” said Molzahn. The researchers will also use the information to figure out how best to expand the grid. Combining the two approaches will allow them to “grow a grid,” as Molzahn put it. The group will start small, with a model of Wisconsin, and gradually scale up, eventually building data sets that represent the entire Eastern or Western part of the grid. This article was originally posted here.