- The Duncan Firm
In the five months since ExxonMobil's Pegasus oil pipeline burst in Arkansas, two things have become clear. Flawed, 1940s-era welding techniques used when the Pegasus was built set the stage for the rupture, and an internal pipeline inspection failed to spot the problem just weeks before the spill.
The most critical question of all, however, has yet to be answered — What caused the pipe's long-dormant flaws — assumed to be J-shaped "hook cracks," in this case — to awaken and grow undetected until catastrophe struck?
"It is the extension of the hook cracks that is the key to this failure," said Patrick Pizzo, a professor emeritus in materials engineering at San Jose State University. "You have to get those cracks in motion in order to lead to a leak or a fracture."
Pizzo and several pipeline failure experts who reviewed the publicly available Pegasus reports say the pipe's cracks probably grew because of large swings in the pressure inside the pipe. So-called "pressure-cycle-induced fatigue" is one of the top four threats in pipelines that — like the Pegasus — were built from pre-1970 pipe that is predisposed to cracking and corrosion problems along lengthwise seams.
But there could be other factors, too, including problems associated with the type of product the Pegasus was carrying — an oil-like substance called bitumen that is mined in Canada and diluted to form diluted bitumen, or dilbit.
One theory is that dilbit made pressure swings inside the pipe larger and more frequent — and thus more harmful over time — because it's heavier and harder to push through pipe, and because its ingredients can vary more widely than conventional crude oils.
Another theory holds that an excessive amount of hydrogen accelerated or triggered the crack growth. In that scenario, the destructive hydrogen could have come from an overcharged corrosion protection system or from the sulfur-heavy dilbit being carried by the Pegasus.
A metallurgical report on the ruptured segment of the Pegasus concluded that substandard manufacturing methods left tiny cracks near the pipe's seam. Those cracks grew until the steel pipe split, sending crude oil gushing from a 22-foot-long gash. Hook cracks are typically formed at the steel mill. But the metallurgical report said that in this case the hook cracks were probably formed later, after micro-cracks from the manufacturing process grew and merged during service. The report didn't say what caused the cracks to grow.
Answering that question is vital, because 30 percent of the nation's 180,000 miles of onshore hazardous liquid pipelines could have manufacturing flaws similar to those on the Pegasus. Operators of those pipelines need to know if they should conduct new inspections, rethink assumptions about crack growth, or adjust the way they operate their lines.
If dilbit turns out to have been a factor in the Pegasus spill, it would further inflame the debate over U.S. imports of Canadian bitumen. That, in turn, would put the spotlight on other flawed pipelines that carry dilbit and would also provide fodder for opponents of the proposed Keystone XL pipeline, which would carry dilbit from Canada to oil refineries on the Gulf Coast.
The federal Pipeline and Hazardous Materials Safety Administration (PHMSA) is investigating the Pegasus failure but has not discussed its work. Exxon spokesman Aaron Stryk said the company would not comment on the matter until the probe is complete.
PHMSA investigators are analyzing the toughness of the steel, residual stress, mechanical properties, chemical analysis of deposits and resistance to environmentally assisted cracking, among other things, according to an Exxon presentation to Arkansas officials. Pressure cycling and hydrogen cracking are types of environmentally assisted cracking.
Pressure in flux
Pressure cycling is pervasive in pipelines that carry liquids because viscosity of the products inside the pipes is constantly changing. Hilly terrain can also add to the pressure swings. Operators also repeatedly stop and start a pipe's flow to unload batches of oil or fuel along the length of the pipe or, in some cases, to take advantage of lower nighttime electricity rates, according to John Stoody, spokesman for the Association of Oil Pipe Lines, an industry trade group. To limit pressure variations, operators can adjust their pumping systems and carefully monitor the viscosity of the batches of liquids moving through the line.
Some operators may change their pumping pressures and their cycles to accommodate customers or to push more crude through the pipe faster, which generates more fees. Exxon, for example, increased the amount of dilbit flowing through the Pegasus by 50 percent in 2009. To accomplish that without installing larger pipe, Exxon had to send oil through the pipe faster, either by adding pumping stations or increasing the overall operating pressure, or a mix of the two.
Three years earlier, in 2006, Exxon also reversed the direction of the pipeline's flow, a move that would automatically alter the impact of pressure cycles by changing where the highest and lowest pressures hit along the pipeline.
Big changes in the internal pressure cause pipe to repeatedly flex, and that can cause special problems in crack-prone vintage pipes like the Pegasus. Exxon's pipe was doubly challenged, however, because its pipe was known to be exceptionally brittle around the seams. Brittleness can cause pipes to fracture instead of flex, just as the way wire will break after being bent back and forth repeatedly.
"Pressure cycling has the ability to seriously affect, vary, and accelerate crack growth rates," pipeline safety consultant Richard Kuprewicz said in a report filed with Canada's National Energy Board about an Enbridge Inc. pipeline project that would carry dilbit to Montreal. Kuprewicz, who is assisting an Arkansas water district in the Pegasus case, has not aired his conclusions about the Pegasus failure, but his Aug. 5 report underscores the dangers of adding pressure cycling and dilbit to an already vulnerable pipe.
"Changing (types of) crude, especially running dilbit, can significantly increase pressure cycles that can accelerate crack growth," he said in the report. "The movement of dilbit in pipelines at risk to cracking threats presents a higher potential to cause pipeline ruptures if not adequately managed."
Cases of pressure cycle fatigue often leave telltale "striations" along the fracture surface that resemble wave lines on a beach, said Pizzo, the materials expert from San Jose State. The Pegasus metallurgical report found no evidence of striations, but Pizzo said that doesn't rule out pressure cycling as a cause of the problem.
"What that means is that either there wasn't fatigue, or there were striations there, but they were obliterated by the activity post-failure," said Pizzo. "It's very difficult to read and see what's there. So it doesn't mean there's no fatigue."
Excess hydrogen could also have played a role — or could even have been the primary factor — in the Pegasus failure, according to Pizzo and a failure analyst who did not want to be identified because of ongoing work with oil companies.
Hydrogen is found in and around oil and gas pipelines as a matter of course. The cathodic protection systems that operators use to prevent corrosion can give off hydrogen atoms if the systems are overcharged. The products that flow through pipelines also usually contain hydrogen in the form of hydrogen sulfide. "Sour" crude oil, which includes many forms of dilbit, tends to have more hydrogen sulfide than typical U.S. crude oils.
The Pegasus, for example, was carrying a diluted bitumen called Wabasca Heavy. That variety has the second-highest sulfur content of the 29 kinds of Canadian crude oil and dilbit listed in a reference guide from Crude Quality Inc., which operates a website that tracks the chemical makeup of Canadian crudes.
Hydrogen sulfide becomes a problem only if it decomposes and the hydrogen atoms move into fragile areas of the pipe.
Pizzo said it works like this — An especially brittle area, like the seam weld region of the Pegasus, will draw hydrogen atoms into the steel pipe. The atoms congregate at the tips of cracks, where the internal stress is higher. Then they weaken the steel by creating larger gaps between the iron atoms that form the pipe. As the atomic hydrogen cluster grows, the pressure builds until the tip of the crack is extended. Then the hydrogen atoms move to the new tip, and the process repeats itself. When the crack grows large enough, the pipe breaks.
This phenomenon is well documented in the world of metals and has had catastrophic consequences for bridges, ships, liquid tanks, pipelines and other metal structures.
Studies have found that cracks in steel pipelines grow when higher levels of hydrogen sulfide are involved. One study found that cracks grew up to 10 times faster, Pizzo said. "If it's 10 times faster, then the consequence is that [the pipe] fails, but instead of in years, it fails in months."
The failure analyst who asked not to be identified said that's a strong possibility with the Pegasus, based on the information that's been made public. The presence of cracks, a brittle fracture and no evidence of fatigue crack growth over time, "certainly, all the pieces are there that would support that conclusion — but I don't know which source the hydrogen was from," the analyst said. "Take one or two of those things away, and maybe you don't have a problem."
Although hydrogen is rarely cited in liquid pipeline failures, the Pegasus is exactly the kind of pipe where it's most likely to occur, according to a recent report on pipes that fail along their long seams.
The failed segment of the Pegasus was made in either 1947 or 1948 by Youngstown Steel & Tube, using low-frequency electric resistance welded (ERW) seams, according to PHMSA. The pipeline's seam area therefore was more vulnerable to cracks as well as particularly receptive to taking in hydrogen atoms that can create cracks or accelerate their growth.
Kiefner & Associates, which is conducting pipeline research for PHMSA, analyzed 280 seam failures in pipe made electric resistance welding or a similarly faulty process. One of its 19 findings, published a year ago, was specific to Youngstown pipe from the 1940s and 1950s — "Operators who have that vintage pipe have to take steps to minimize the chances of atomic hydrogen being generated at the (internal) surface of the pipe from internal sour components or from excess cathodic protection at the [outside] surface."
The report cited a case that closely mirrors the circumstances of the Pegasus. That incident involved a 1949 Youngstown pipe that fractured even though, like the Pegasus, it was operating well below its maximum stress levels. The pipe had a small hook crack that had survived multiple hydrostatic pressure tests, but it ultimately grew and fractured in part because it was located in the extra brittle seam area.
Kiefner's report called the accident a case of hydrogen-induced cracking, saying the pipe sustained a "sudden hydrogen embrittlement failure, with the hook crack providing the stress concentration needed" to attract extra hydrogen atoms.
Kuprewicz, the pipeline safety expert, said he isn't convinced there's a link between excess hydrogen and the Pegasus failure, but that he "can't rule it out 100 percent."
No matter what caused the cracks to grow, the Pegasus rupture will have ripple effects in the pipeline industry and at PHMSA, which is charged with ensuring the safety of the nation's pipelines.
"There's so many miles of these pipelines ... that even though it may be a fraction of those or a percentage of those [with similar problems], it's still a big number," said Pizzo, the pipeline materials expert.
This story is part of a joint investigative project by Arkansas Times and Pulitzer Prize-winning InsideClimate News. Funding for the project comes from readers who donated to an ioby.org crowdfunding campaign that raised nearly $27,000 and from the Fund for Investigative Journalism.