08/20/03

Researchers examining the behavior of the nation's electrical grid in the minutes before last Thursday's blackout have spotted a unique pattern that may help to explain why the massive power failure happened.

The novel theory from a respected industry research group comes as the U.S. Department of Energy dispatched hundreds of investigators to secure and collect computer data about the grid's condition leading up to the largest power failure in American history.

Knowing the cause of the blackout is essential if preventive steps are to be taken, government and industry officials say.

Recordings of electrical activity on the grid Thursday afternoon show what appear to be signs of a relatively new and poorly understood phenomenon called "fast voltage collapse" - a seconds-long event in which available power suddenly drops, then gradually recovers, but not in time to meet the demand of electricity-hungry air conditioners and industrial equipment.

What triggered the collapse - if that is what happened - is unknown. Nor do they know where it occurred.

The engineers at the electrical industry's research arm who did the nonpublished assessment the night of the blackout say the answer is deep in the piles of computer data that must still be reviewed.

"Based on the initial data on Thursday night, it has some characteristics related to fast voltage collapse," said Mike Howard, president of a technical group within the California-based Electric Power Research Institute, or EPRI, whose scientists will be working with the Energy Department. "But to definitively say it was fast voltage collapse requires much more analysis of all the events. That's going to take months."

"I think that the initial analysis by EPRI is a good working theory of what might have caused the cascade," said Robert Burns, a senior research specialist at Ohio State University's National Regulatory Research Institute. "While we don't know the cause, this theory seems to be consistent with the facts that we do know."

A power problem in the Phoenix area in 1995, which the researchers previously had identified as a "near-fast voltage collapse" has similar characteristics to last Thursday's much larger event, said Arshad Mansoor, one of the authors of the EPRI analysis.

The Phoenix incident, on a July afternoon when the temperature soared to 112 degrees, began when equipment meant to stabilize electrical flow failed, causing five high-voltage transmission lines to shut down. Voltage dropped by nearly half, but the output of the Palo Verde nuclear plant helped prevent a total collapse.

In Thursday's blackout, readings from grid monitors in New Jersey showed that voltage dropped by 60 percent from normal levels at 4:10 p.m. The attempted recovery as generators tried to make up for the lost power took seconds - "an eternity" in electrical system time, according to Mansoor.

That gradual creeping up of voltage is a classic symptom of fast voltage collapse, said Mansoor. Unlike Phoenix, however, nuclear and coal-fired generating plants in the Northeast and Midwest were unable to handle the power fluctuations and tripped offline. The transmission system, lacking enough power to satisfy millions of residential and industrial customers, shut off and the voltage fell to zero.

If customers of the nation's electrical grid used nothing but light bulbs, a sudden voltage drop would do nothing more than cause the bulbs to dim. But modern motors such as those in refrigerators, air conditioners, computers and factories are controlled by microprocessors that demand more current to keep the power levels constant.

Some of those devices will shut down, lowering the power demand and causing voltage to surge back up. The sudden imbalance between available power and the demand for it overloads the transmission lines and generating stations, leading to the collapse of both.

The bulk of the limited research into fast voltage collapse has been done in only the last several years, and Howard and Mansoor said the phenomenon still is not well understood. Computer models that would simulate the conditions are almost impossible to create, because there is no way to account for the variations in how and when millions of customers use appliances and machinery.

"It's kind of a moving target," Mansoor said.

One possible solution is being developed by Alex Huang, a Virginia Tech engineering professor. The device is a high-speed computerized switch that works to correct voltage imbalances by shifting electrical load to another transmission line. Its drawbacks are its high cost and the need to put the switching equipment in many locations.

Theories about the triggering event abound, from an explosion at a New York power plant to a lightning strike in Ontario to the short-circuiting of high-voltage transmission lines in Akron-based FirstEnergy's territory.

FirstEnergy officials continue to insist that the problem originated with an eastern connection.

Energy Secretary Spencer Abraham, addressing reporters at a Washington news conference yesterday, said it was too soon to comment.

He said there had been "a lot of rumors, a lot of speculation as to what happened, who might be at fault. There's time to deal with those issues.

"What I think the American public wants, what I'm sure the citizens of Canada want, is a comprehensive and thorough investigation before we determine precisely what happened and what we need to do about it."

Abraham announced that the Department of Energy and its Canadian counterpart will lead the inquiries being conducted by government agencies, utilities and regional transmission groups. The intent is to have "one investigation" and "one ultimate finding or one ultimate set of findings."

The North American Electric Reliability Council, a private nonprofit group that administers the electrical grid, had been doing its own investigation.

It has agreed to work with Abraham's task force "and to forgo its own investigation of the incident," the energy secretary said.

Also cooperating will be the Nuclear Regulatory Commission, the Federal Energy Regulatory Commission, electrical grid operators in the regions affected, utility companies, and national laboratories that conduct scientific research and test theories for the government, Abraham said.

The investigation will include "hundreds" of people, he said.

He said the task force will produce "one final and authoritative" report, but would not commit to a firm time frame, saying the work will be done "as effectively as we can in as short a time frame as we can."

This morning, Abraham will meet in Columbus to discuss the blackout with Gov. Bob Taft, who is cutting his Canadian vacation short to attend.

Plain Dealer reporters Stephen Koff and Sandy Theis contributed to this story.

To reach these Plain Dealer reporters:

jmangels@plaind.com, 216-999-4842

jfunk@plaind.com, 216-999-4138