Dr. Don Reising is a Guerry and UC Foundation associate professor of electrical engineering at UTC. Photo by Angela Foster.
Customers do not have electricity in their homes or businesses, so—among other issues—they do not have lights, heat or air conditioning.
Power companies must scramble to find the problem, fix it and prevent it from happening again.
A professor at the University of Tennessee at Chattanooga, TVA and Grid Protection Alliance are collaborating on a way to prevent all those headaches.
“This was something TVA talked about early on,” said Dr. Don Reising, a Guerry and UC Foundation associate professor of electrical engineering at UTC. “It was, ‘Man, it’d be really great if we could predict these things and see that they’re going to fail before they actually catastrophically fail.’”
Reising and his graduate students have researched the issue for the past three years using $300,000 in TVA grants. They devised a solution that TVA is testing and hopes to implement entirely in 2024.
“The whole goal of it is to prevent failures on the system,” said Tony Murphy, senior program manager for power quality at TVA. “If we have a piece of equipment that is trending towards failure, now we can be alerted to it proactively and preemptively take it out of service, repair it or replace it before the catastrophic failure occurs.
“When we can do that, we improve reliability. We prevent an unplanned outage, so there’s no power outages to the customers or the system.”
The failure of some equipment can be so disastrous that it causes a chain reaction through the system, wrecking other pieces of equipment. Sometimes, equipment may even explode, sending dangerous shrapnel flying.
“With people nearby, it could result in injury or death, so it’s a safety-performance improvement,” Murphy said.
The problem is that predicting equipment failure is more challenging than it might seem at first glance.
Dr. Don Reising
TVA uses digital fault recorders spread throughout the system’s seven-state region to record minute-by-minute data from its operation. The recorders collect data 24 hours a day, seven days a week, 365 days a year—an amount so voluminous that “it was impossible to go through it all” to look for potential problems, Reising said.
“You don’t know what abnormal is if you don’t know what normal is,” he explained. “If you’re only looking at a small percentage of the events that are happening in your system, you have no idea if those are legitimate and supposed to be happening. You couldn’t tell the difference.”
Murphy said that system-wide data was scarce before the fault recorders were created.
“Years ago, the problem was we didn’t have data,” he said. “If some bad thing would happen in the system, going into an investigation to do post-analysis was very difficult.
“The problem has now gone from ‘no data’ to ‘we are drowning in data.’ We need some means of mining the nuggets of gold out from among all the rocks, so to speak.”
Reising and his team conducted research that developed software that can take reams of data, compress it and analyze it more quickly than a human could ever do.
Murphy said it is now possible to take a massive amount of data and see patterns in the electrical grid that may indicate problems.
“It makes it manageable for us to retrieve and process,” he said. “It can flag for variances of ‘off normal’ as another thing to alert us to trending health issues on the system.”
Reising said that one of the goals was to make the new software available to all power companies. Around two years ago, he began working with Grid Protection Alliance, a Chattanooga not-for-profit organization specializing in developing and supporting open-source software solutions for the electric industry.
“We think that there is a huge value in using open-source software for utilities like TVA,” said Dr. Chris Lackner, GPA’s operating officer of grid solutions. “We are a not-for-profit that focuses on enabling innovation and technological improvements by using open-source software.”
Lackner’s research background includes stints at Rensselaer Polytechnic Institute in New York, Sandia National Laboratories in New Mexico, and ISO New England. He said it has been exciting to see the new technology and algorithms “that universities and professors like Dr. Reising come up with.”
“In this case, we have worked with TVA and took some of what he has been developing in the lab,” Lackner said, “and we are able to implement that into a production environment where it’s actually being used to keep the power on for pretty much all of Tennessee.
“It’s been great working with UTC. We’re really excited to bring some of the research Dr. Reising has done into the operational environment.”
Being open source, it is available for utilities for free, Lackner explained.
Reising said it’s possible for the technology to be used by power companies both large and small, especially those in rural areas or regional-type utilities.
“When you look at some of these local companies,” he said, “they don’t have a lot of people to do the work. They might not have the resources—whether that’s in money or that’s people or both—to do it.”
Lackner agreed.
“We’re trying to facilitate the use of open-source software, so this is an example of that,” he said. “We are trying to go to utilities and say, ‘Hey, there’s this cool tool; it’s open source and it’s available to you. Let’s see if we can get that into production at your utility.’
“Whether that’s San Diego Gas and Electric over on the West Coast or the Independent System Operator of New England up in the Northeast, we’re not just limited to putting this into production here.”
Learn More
UTC College of Engineering and Computer Science
