Michigan oil spill echoes Minnesota leak in 2002

By Todd A. Heywood
Monday, August 09, 2010 at 8:08 am

Photo of the ruptured seam in line 6B in Michigan's Calhoun County. Photo: NTSB

Remarkable similarities between a 2002 leak in an Enbridge oil pipeline known as Line 4 in Minnesota and the recent leak in Enbridge pipeline 6B in Michigan’s Calhoun County — and marked differences in the way the company responded to the two incidents — are raising new questions for the company and for investigators looking in to the cause of the spill.

The National Transportation Safety Board’s report on the 2002 spill in Line 4 in Minnesota shows many similar elements between the pipelines themselves and the events leading up to the leak.

Both lines were put in the ground in a similar time frame, with Line 4 being buried in 1967 and 6B being buried in 1969. They both have carried a combination of crude oils from Alberta, Canada to points in the Midwest of the United States as well as refineries in Sarnia, Ontario. Both were likely transported by rail and truck line to their final resting places. And both were buried in marshy wetlands.

In addition, both lines caused the supervisory control and data acquisition system, or SCADA, to trigger pressure and suction alarms in the Enbridge Edmonton control center. SCADA is a system of complicated computer monitoring sensors up and down thousands of miles of Enbridge pipeline. Those sensors are constantly measuring various values, including pressure and suction in the line, sending thousands of readings into that control center an hour.

The din has been described by Richard Kuprewicz, president of Accufacts, Inc. in Redmond, Washington, and an expert with 40 years experience in pipeline engineering, as “noise.” Employees, he said, can see as many as 1000 pressure or suction alarms in one eight hour shift. Employees in the control room must listen to the noise and filter out pressure readings indicating a real problem versus readings that can be caused by a variety of fluctuation issues inherent in sending crude oils thousands of miles through pipelines.

There were also warning signs of possible trouble before each incident. Prior to both spills, inline testing of the pipes indicated weak spots in the line. In Minnesota, the testing indicated a crack that was not substantial enough to set off red flags. In Michigan’s case, the EPA had warned Enbridge of corrosion issues in the same pipeline and the company was asking for more time to correct them.

But Kuprewicz says there is a “95 percent probability” the line did not rupture as a result of corrosion, at least based on the initial information available, particular the location of the rupture. Officials have said the rupture was located at the three o’clock position on the 30 inch pipe, and Kuprewicz says corrosion-induced ruptures are almost always at the bottom of the pipe.

In both instances, the leaks dumped hundreds of thousands of gallons of crude oil into sensitive wetlands. In Minnesota, officials decided to do a controlled burn of the oil, creating a plume of thick black smoke a mile high and five miles wide in the July sky. In Michigan, more traditional booms and skimmer operations contained the spill.

In both cases the goal was preventing the oil from reaching significant bodies of water — the Mississippi River in Minnesota and in Michigan an 80-mile stretch of the Kalamazoo River which is a designated Superfund site by the EPA due to previous contamination with PCBs.

Photo of seam rupture in line 4 in Minnesota in 2002. Photo: NTS

Finally, each rupture produced what Kuprewicz identified as “fish mouth” rupture holes. Those rupture images are a telltale sign of a pressure fracture or seam blow out, he said. But defining the cause will require detailed metallurgical analysis of the pipelines. In the case of the 2002 rupture, the determination was made that the rupture happened as the result of a stress fracture that began when the steel pipeline was loaded for its trip to Minnesota in 1967.

“[Fish mouth] ruptures are typical, but the initiating failure could be different,” Kuprewicz said.

There are also key differences in the two incidents. The most obvious is that Line 4 was in operation at the time of the incident, while Line 6B had been shut down for a scheduled bit of upkeep when the rupture occurred. But potentially the most important difference is how employees at Enbridge’s Edmonton, Alberta control center responded.

In 2002, employees had shut the line down completely within 15 minutes of the rupture, closing the valves throughout that section. Within 20 minutes they had notified emergency first responders in Minnesota. Within three hours, the leak had been identified by Enbridge employees on the ground.

In the case of Marshall, however, officials noticed alarms at 5:58 p.m. on Sunday, July 25 as they shut the line down for scheduled maintenance. Edmonton employees chalked the alarms up to normal shut down operations. At 4:04 a.m. Edmonton control crews attempted to restart the line, and after nearly 30 minutes of trouble doing so, called in specialists. For the next few hours, those specialists and control room employees turned the pipeline on and off, as well as increased the pressure in the line.

They didn’t isolate the area until shortly before 8 a.m. and didn’t dispatch ground crews until 9:49 a.m. The leak itself was not visually confirmed until 11:16 a.m. by a Consumers Energy employee.

Ultimately the cause of the rupture will be determined by investigations by NTSB, the U.S. Environmental Protection Agency, the Michigan Department of Natural Resources and the Environment, the U.S. Department of Transportation Pipelines Hazardous Materials Safety Administration and Enbridge officials themselves.

Kuprewics says it’s simply too early to tell whether the similarities between the two incidents point to a similar origin or whether they are merely coincidental — and whether the actions of the Enbridge control room in the more recent incident may have contributed to making Michigan’s spill four times greater than the Minnesota spill.

“Seam cracks and resulting failure can occur for a variety of reasons,” he said, “and right now, if this is a seam crack initiated failure, there just isn’t enough information yet to determine the initiating cause.”

Categories & Tags: Top Stories|


1 Comment

Comment posted August 9, 2010 @ 10:33 pm

This is good. We need more of this kind of reporting.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.