Daniel Bienstock | Short-Circuiting Blackouts

Daniel Bienstock
Professor of Industrial Engineering and Operations Research and of Applied Physics and Applied Mathematics
This profile is included in the publication Excellentia, which features current research of Columbia Engineering faculty members.
                                    Photo by Eileen Barroso

On August 14, 2003, an unusual combination of events shut down electrical power for 55 million people in the United States and Canada. The event was what statisticians call a black swan, something so rare that no one plans for it. Yet major blackouts also occurred—with significant consequences—in 1965 and 1977.

Daniel Bienstock believes that by studying these black swans, he can help utilities prepare for and even prevent the next major blackout.

“We’re borrowing ideas from other engineering disciplines,” said Bienstock. “If you design an aircraft wing, you test it by strapping it to a fixture and vibrating it to see what breaks. We do the same thing by using our model of the grid. Our objective is to stress the grid and see where it breaks.”

The 2003 crisis involved a number of unusual events that occurred in different parts of the Eastern United States grid. These included human errors, control room computer bugs, and a plant shutdown. These events put too much strain on several major power lines, causing them to overheat and eventually shut down. This started a process that snowballed until it knocked down much of the Eastern grid.

“One or two things like this can happen frequently and the grid can handle it,” said Bienstock. “But here, the particular combination proved catastrophic. After gathering momentum for several hours, the snowball caused hundreds of lines to fail within 15 minutes.”

His goal is to create software that will let utilities analyze cascading events and react to them before the grid comes down. First, though, he needs to anticipate what combination of events could trigger a blackout.

“The traditional way is to enumerate every possible combination of individual lines coming down and determine the consequences,” he said. “There are not enough computers in the world to do this.”

Instead, Bienstock created a mathematical model of the grid and stresses it in different ways.

“We can use the model to show us where the grid will break. Then we see what we can do to address those vulnerabilities,” he said.

He also runs what-if scenarios.

“It’s a chess game. We look at cascading events and test different strategies to find the best way to react. These become templates that can guide utility responses before these events turn into a major outage,” Bienstock said.

“Imagine,” he concluded. “We would have had at least one hour to do our computations in 2003. We could have calculated the right moves to prevent a blackout.”

B.S., Brandeis University, 1982; Ph.D., Massachusetts Institute of Technology, 1985

500 W. 120th St., Mudd 510, New York, NY 10027    212-854-2993