FORECAST: Increased chance of a blackout?
 |
n August 14, 2003, the largest electrical blackout in U.S. history knocked out power in the northeastern United States and parts of eastern Canada. For up to four days, the outage brought the lives of some 50 million people to a standstill, at a cost of $4 billion to $10 billion.
Six years later, researchers worldwide still study the event, hoping (among other things) that the evidence will help them identify the perfect balance between the risk of another blackout and the reliability of a massive, intricately interconnected power system. “It’s both a very interesting problem and an important problem to keep the lights on and to have enough reliability in our power system so that we can use electricity,” says Professor Ian Dobson.
A failure in your home electrical system might mean that a circuit breaker trips or a fuse blows to prevent an overload from melting down the whole system. Similarly, the nation’s power grid is designed with enough redundancy and robustness to handle isolated failures, such as a lightning strike or even a scheduled line maintenance.
However, at the heart of a blackout are failures that propagate, or cascade, throughout the power system. “The distinctive thing about cascading failure is that one failure happens and it maybe weakens the system a little bit,” says Dobson. “It’s more likely that the next failure can happen after that.”
And just by chance, he says, the failures continue to snowball until there’s a blackout.
While blackouts such as the August 2003 event can wreak havoc for days, they actually are relatively infrequent, occurring every couple of decades. As a result, it’s not necessarily cost-effective to take measures to eliminate them entirely. Extra transmission lines, for example, add additional reliability—yet cost more than $1 million per mile to construct. “There’s a balance here, as a society,” says Dobson. “We need reliable electrical power, but we also need inexpensive electrical power. And so how much do we all pay in our electricity bills in order to pay for reliability?”
Using computer models, Dobson, his students and colleagues are studying cascading failures at a very basic level. “Even though they’re very complicated in all the details, we can look at blackouts and try to estimate how big the initial disturbance was and how much, on average, it tended to propagate,” he says. “Looking at blackouts in this way is a very high-level description, but I hope to learn information that will help people understand the risk of these large blackouts and help us to put approximate numbers on how likely they are and what are the consequences for our society.”
The result might be that electrical system operators can “forecast” an increased chance of a blackout for a given time period. Those projections, in turn, could help utilities better determine how much to invest in adding robustness and redundancy to the nation’s transmission grid.
