NMSU researcher tackles power grid blackouts
August 25, 2015
Sukumar Brahma, a professor in the Klipsch School of Electrical & Computer Engineering at New Mexico State University, is developing methods to produce more reliable and resilient delivery of electric power to end users.
Electric power grids often experience disturbances that are managed, controlled or eliminated by built-in system protection measures that have been developed through engineering studies and years of operational experience. Although these measures are in place, power grid systems are still vulnerable to certain disturbances that will cause the power grid to black out.
“Twenty years ago, we did not have the technology to determine what was causing the blackouts,” Brahma said. “Now, as technology has advanced, we have the ability to supervise the protection system, detect malfunctions and nullify their effect.”
Within power grid systems, relays provide protection against short-circuits. Relays, however, may cause blackouts if their operating contacts are stuck or if the relay itself has an operation malfunction.
“Traditionally, we have been unable to tell if a relay has gotten stuck,” Brahma said. “Now, we have digital relays that can self-diagnose for stuck relays. In addition, digital relays allow us to verify if the relay is operating correctly.”
Brahma said the problem with digital relays was they were susceptible to hacking, which created more danger. Hackers can isolate the most critical relays, shut them and make them operate incorrectly. To prevent hacking and other issues in the relays, Brahma and his team will use an extra set of precise measurement devices – phasor-measurement units (PMUs).
PMUs can be mined for data to add another layer of intelligence to determine whether the relays are operating appropriately. If the relay operation malfunctions, the PMUs can be used to disable and reverse the incorrect operation in a timely manner to prevent any potential blackouts.
“The data we can obtain from the PMUs are substantial,” Brahma said. “Currently, there are around 1700 PMUs in the USA, but that number will increase over time.”
Electric grids are stable most of the time, as destabilising disturbances are not very frequent. PMUs allow researchers to mine data and detect the disturbances. If a disturbance is detected by a PMU, Brahma said it was easy to identify the cause based on the PMU data, which can then reveal whether the disturbance was caused by a short circuit, the loss of a generator or the loss of a big load. If no disturbance was detected, but the relay still operated, the problem is a relay operation malfunction.
“Basically, we are trying to develop resilient power systems, which essentially makes them more reliable,” Brahma said. “The power grid is our nation’s most critical infrastructure, so curtailing the size, scale and progression of blackouts while improving the robustness and resilience is vital.”
Brahma’s research project focuses on two approaches to the problem. The first is data-driven using pattern recognition and classification of data streams. Brahma is collaborating with Huiping Cao, a professor of computer science at NMSU, whose expertise is in data mining and classification. This research is funded by the Crest Interdisciplinary Center of Research Excellence in Design of Intelligent Technologies for Smart Credits, a smart-grid centre established in 2014 with a grant from the National Science Foundation. The primary goal of the iCredits centre is to serve as a new epicentre for research and training in smart grids.
The second approach is to analyse energy functions of power systems from the measured data. Brahma is collaborating with professors at North Dakota University on this research.
Brahma has been teaching at NMSU since 2007. He is currently the William Kersting endowed chair professor in the Klipsch School. In addition, he is the associate director of the electric utility management programme at NMSU. His research focuses on power system protection, modelling, simulation and analysis, with a focus on smart grids.