Since at least 2006, ExxonMobil has known that its 1940s-era Pegasus pipeline had many manufacturing defects like the faulty welds that recently sent crude oil spewing into an Arkansas neighborhood. The company also knew that the seams of the pipe have been identified by the industry as having another dangerous flaw: They are especially brittle, and therefore more prone to cracking.
“Having a crack or flaw in a pipeline is a whammy,” said Patrick Pizzo, a professor emeritus in materials engineering at San Jose State University. “But having a crack embedded in brittle material, such as the heat-affected zones of the pipeline seams—that’s a double whammy.”
Despite those inherent risks, Exxon added new stresses to the Pegasus by fundamentally changing how it used the line. It began carrying a heavier type of oil, reversed the direction of the flow and increased the amount of oil that surged through it.
Old pipelines that are well maintained and carefully monitored can withstand such changes, experts say. But doing so is significantly harder for lines that—like the Pegasus—were built from pre-1970 pipe that is predisposed to cracking and seam corrosion problems.
In early 2006, before the 858-mile Pegasus was reconfigured, Exxon tested its strength by conducting a hydrostatic pressure test. Hydrotests are expensive and disruptive but they are also considered the most reliable way to detect and eliminate perilous cracks on problematic pipelines. Operators must suspend business and empty the pipeline, then pump water through it at higher-than-normal pressure to force vulnerable parts of the pipe to fail—a process that’s repeated until the pipe and its repairs hold the water without leaks or splits.
After a series of ruptures and repairs, the Pegasus cleared the hydrotest. Soon after, Exxon started using it to ship a Canadian heavy crude oil-like substance known as diluted bitumen, or dilbit, from Patoka, Ill., to refineries on the Texas Gulf Coast. Two years later, Exxon expanded the pipeline’s capacity by 50 percent to 90,000 barrels per day.
Exxon declined to say whether the expansion also boosted the line’s operating pressure, and if so, by how much.
“All I can say at this time is that prior to the incident, the pipeline was operating under normal pressure and flow conditions,” said Exxon spokesman Aaron Stryk.
Many key details about the March 29 spill in Mayflower, Ark. have yet to be made public, and the pipeline failure is still under investigation by the federal Pipeline and Hazardous Materials Safety Administration (PHMSA). The agency recently released the metallurgical report on the broken section of pipe as well as the results of the 2006 hydrotest—information PHMSA had previously declined to make public. Other information, including data and analysis from inspections in 2010 and 2013, has not been circulated outside a small group of politicians and agencies.
What has emerged so far suggests that, at a minimum, Exxon took calculated risks given the known condition of the 20-inch pipeline and either used a flawed integrity management plan, or had a good plan and didn’t adhere to it.
“I’m looking at this from a distance … but they were not very careful,” said Mohammad Najafi, a pipeline construction expert and professor at University of Texas at Arlington, who has reviewed the metallurgical report and hydrostatic test results. “They should have been more concerned about the pipe.”
Stryk, the Exxon spokesman, said the company “established a system for determining seam-failure susceptibility and for initial and subsequent seam integrity assessments.” That system, he noted, “has been reviewed by PHMSA, is modeled after [the agency’s] recommendations and guidelines for seam assessments, and is consistent will all applicable pipeline regulations.”
PHMSA last reviewed the Pegasus integrity management plan during a special inspection in March and April 2011, Stryk said.
The section of the Pegasus that burst open in March was operating at a pressure of 708 pounds per square inch, well below its maximum operating pressure of 820 psi, according to PHMSA, which regulates most of the nation’s liquid fuel pipelines.
The line split apart, opening a 22-foot gash and sending 5,000 barrels—or 210,000 gallons—of dilbit into a nearby neighborhood and lake. The noxious fumes and sticky flood forced 22 families to flee their homes.
Testing the Pipe
The type of operational changes Exxon made after the Pegasus was hydrotested in 2006 would have changed the dynamics inside the pipeline, industry experts agree. Switching to the more viscous Canadian dilbit, for example, could have subjected the pipeline to more pronounced pressure swings and “pressure-cycle-induced fatigue,” a condition known to activate defects and expand cracks, sometimes to the point of rupturing the pipe.
“You don’t have much safety margin with these kinds of crack threats, especially if they grow,” said Richard Kuprewicz, a pipeline consultant who has been hired to advise an Arkansas water agency affected by the spill. To be safe, “you may have to do another hydrotest in a few years.”
Two reports commissioned by PHMSA and released in 2012 confirmed the value of hydrotests. One examined 280 defect-related seam failures and concluded that more than 20 percent of them “could have been prevented by a timely hydrostatic test.” The other discussed the drawbacks of high-pressure hydrotesting, but still endorsed it as an important safety measure for pipes with seam weld defects.
Not all hydrotests are sufficiently robust to do the job, however. To eliminate the most dangerous cracks in vulnerable pipelines like the Pegasus, operators must use extra high pressure—pressures higher than Exxon used, according to Kuprewicz. The 2006 Pegasus test met the minimum pressure required to establish the line’s maximum operating pressure, but it wasn’t high enough to clear the line of cracks that could later split the pipe.
Exxon did not conduct another hydrotest after 2006. Instead it checked the pipe section that later failed with in-line inspection tools that didn’t entail shutting down the line. Neither of those tests—conducted in 2010 and early 2013—could be counted on to discover or eliminate the type of cracks that would threaten the Pegasus, Kuprewicz said. The specialized “smart pig” technology Exxon used in 2010 was designed to find general corrosion, not cracks. Three years later, the company employed a new crack-detection technology that hasn’t proven itself in the field.
About a third of the nation’s hazardous liquids pipelines are prone to the types of cracks found on Pegasus, according to Carl Weimer, executive director of the Pipeline Safety Trust, a nonprofit watchdog group based in Bellingham, Wash. They were built from pipe made with low-frequency electric resistance welds, a process that was widely used by steel mills before 1970.
But not all the lines built that way have the flaws. And those that do can still be used safely with appropriate caution and vigilance, said Kuprewicz, the pipeline consultant.
A metallurgical analysis of the ruptured section of the Pegasus concluded that the primary cause of the rupture was a seam defect that was accompanied by J-shaped “hook” cracks that grew while under operational stress. The report also said the steel’s brittleness limited the pipe’s ability to flex with changes in pressure or flow inside the line, and it found damage to the pipe’s outer protective coating and limited corrosion in some areas.
Pizzo, the materials engineer, examined the metallurgical report and noted that the Pegasus hook cracks extended about halfway into the 0.312-inch steel. The most important question now, Pizzo said, is “What drove these cracks over time to become deeper and sharp, such that normal pressurized conditions—those that the pipeline was subject to for many years—was finally too much for this pipeline to bear?”
What Exxon Knew
The safety of the liquid fuel and natural gas pipelines that pass through neighborhoods, cities and fragile waterways rests largely with the companies that own and operate them. Critics say PHMSA’s testing and other requirements aren’t strong enough to protect old pipes against crack risks, and the agency lacks the funding and staffing to fully police the nation’s vast network of pipes.
PHMSA requires pipeline operators to re-assess their systems at least once every five years. If a pipeline is known to have defects, and is located in heavily populated or environmentally sensitive areas, operators must create an “integrity management” plan showing how they will prevent leaks and ruptures. PHMSA rules let operators decide what tests to use and whether to inspect those lines more frequently than required, according to Weimer of the Pipeline Safety Foundation.
Based on industry studies and its own records for the Pegasus, Exxon knew or should have known:
► That about 650 miles of the 858-mile long Pegasus was made of pipe manufactured using a technique known to create hook cracks and other defects that can cause the seam welds to fail if operators don’t take proper preventative measures.
► That the same stretch of line was built in 1947 and 1948 with pipe made by Youngstown Sheet & Tube Co., whose pipe from that era is known to be prone to brittleness and fractures.
► That the 2006 hydrotest it performed on that stretch of pipe was conducted at stress pressures appropriate for calibrating maximum operating pressures, but not at levels experts believe is necessary to rid a pipeline of seam crack threats. The stress pressure Exxon used in 2006 also was lower than the stress pressure it used in 1991 to test a newer segment of the Pegasus.
► That 11 seam welds failed during the 2006 hydrotest, even though that test was conducted at relatively low pressure levels.
► That 12 seam cracks were found during a 2010 in-line inspection of a segment of the pipeline outside the failure area.
► That the new crack-detection technology used in the February 2013 probe of the Pegasus hasn’t been proven effective in the field and anecdotal evidence suggests it has a spotty success rate.
What adjustments Exxon made to compensate for those risks—and for the additional operating stresses it placed on the line after the 2006 hydrotest—is not yet clear.
The company has labeled the results of its 2010 and 2013 in-line inspections proprietary and confidential, and has urged PHMSA to deny requests to make them public. The 2013 test produced so much data that Exxon is still analyzing and interpreting them five months later, although the company has provided some preliminary data to PHMSA.
Both PHMSA and Exxon have assured the public that the Pegasus will not be re-opened until it has been tested and found safe. Kuprewicz said the question then will be one of trust: “How can they assure the public that what they’re going to do from now on, or what they’ve done or what they’re about to do, will prevent these hook cracks from going to failure?”