By San Francisco Chronicle news reporter Jaxon Van Derbeken
Pacific Gas and Electric Co. knew about more than two dozen unexplained leaks dating back decades on its San Bruno natural-gas transmission line when it certified that the pipe was free of seam-weld leaks less than a year before it exploded, a Chronicle review of inspection reports and other documents shows.
PG&E workers found at least 26 leaks on the Milpitas-to-San Francisco pipeline from 1951 to 2009 whose causes were listed as “unknown,” according to a November 2009 report that the company kept on file to show that it had inspected the line in compliance with federal pipeline safety regulations.
Although federal law required PG&E to consider any previous longitudinal seam failures or leaks when it picked a pipe inspection method, there is no evidence the company ever found original documentation to rule out bad seam welds as a cause of any of the San Bruno line’s leaks. Instead, it relied on computerized records with scant information that were loaded onto a database starting in the 1990s.
Federal investigators learned only after the San Bruno blast that at least one of PG&E’s leaks, discovered in 1988, had in fact been caused by a defective weld on a longitudinal seam. PG&E never conducted an inspection that could detect such problems before another flawed weld ruptured on the San Bruno line in September 2010, setting off an explosion that killed eight people and destroyed 38 homes.
If PG&E had confirmed at least one seam weld leak, it would have been required to check the San Bruno line for other such problems using high-pressure water – a test the company seldom used on any line. PG&E inspected most of its lines, including the one in San Bruno, with a cheaper, less burdensome method best suited at finding corrosion.
‘Worse than roulette’
“Having histories of leaks and failures, that is fundamental to the regulations” governing inspection methods, said Royce Don Deaver, a pipeline industry veteran turned safety consultant in Texas.
Not knowing the cause of leaks is just as bad as denying they occurred at all, he said.
“There is absolutely no justification for doing this,” Deaver said of PG&E’s reliance on incomplete data to certify the San Bruno line as safe. “This is worse than roulette, because with roulette, there is a chance you can get the right answer. This was designed and functioned in way to come up with the wrong answer.”
The National Transportation Safety Board concluded in August that PG&E was to blame for the San Bruno disaster, citing what its chairwoman called “poor record keeping, inadequate inspection programs and an integrity management program without integrity.”
The company has conceded that its gas integrity management program was not what it should have been and has brought in new management to run the gas division since the explosion. It has also embarked on a program of pressure testing that it says will cover hundreds of miles of pipeline in the next several years.
“We’ve been open about the fact that our past operations should have been better,” spokesman David Eisenhauer said. “PG&E is looking forward now, because looking at how we can make our system safer is the most important thing we can do for our customers.”
Eisenhauer did not respond to questions about the leaks found over the years on the San Bruno line.
“We are committed to learning from mistakes,” he said. “But publicly rehashing issues over and over about things we may or may not have done wrong decades ago does not serve that purpose of making our system safer.”
Evidence that PG&E knew of 26 unexplained leaks on the San Bruno line since 1951 was found on computer-generated survey forms included in the November 2009 inspection report. Such forms have spaces to note critical details on each leak – including its exact milepost location and any findings about its cause, as well as tracking numbers that refer to original repair documents.
But none of the forms for the 26 leaks cited in the 2009 inspection report for the San Bruno line, known as Line 132, has a tracking number or other details – like a specific date – to assist in finding original documentation.
In each case, the cause of the leak is shown simply as “unknown.”
Another document in the November 2009 report – created to certify that Line 132 was free of seam weld problems and thus fit for a corrosion-only inspection – refers to 33 leaks that PG&E crews had found on the San Bruno line with “no causes given.” It did not specify a time frame or list the leaks.
Experts say that in the absence of hard evidence about the cause of a particular leak, pipeline operators should err on the side of caution and assume that a weld problem is to blame.
“You could have defects that excessively endanger public safety,” said Robert Bea, an engineering professor at UC Berkeley who has been following the San Bruno case. “In that case, the standard of care is guilty until proven innocent: You stop operations and find out whether the condition is posing an unacceptable risk.”
Even a small leak on a transmission line can lead to an explosion, Bea said.
“You should have zero tolerance for leaks,” he said, likening the danger to finding a crack on an airliner.
“At that point, they stop and determine if the crack is endangering the system,” Bea said. “They don’t keep flying the plane.”
A Chronicle review of PG&E inspection reports for other transmission lines in the company’s urban gas system found that a dozen had a total of more than 100 leaks that the company never accounted for. None, however, had more unexplained leaks in its history than Line 132.
Eleven leaks were detected from 1963 to 1984 on a half-mile segment of Line 132 about 14 miles north of the Milpitas terminal. Five leaks were found along a stretch of the line between 3 1/2 and 4 1/2 miles north of Milpitas from 1951 to 1974, the records show.
Another mysterious leak in October 1988 led PG&E to replace 12 feet of the pipeline in San Mateo near Crystal Springs Reservoir.
On the November 2009 inspection report, the cause of that leak is listed as “unknown.” But the National Transportation Safety Board discovered six months after the San Bruno blast that PG&E had documentation listing the actual cause as a defective longitudinal seam weld.
Based on available documents, the board concluded in its final report on the San Bruno blast: “It is probable that additional longitudinal seam weld defects have remained in service since 1948.”
Although PG&E accounting records indicate that the company investigated the 1988 leak, the company has not been able to produce the findings. Publicly, the company has sought to blame that leak on a pinhole in a girth weld around a pipe’s circumference, a less dangerous problem than a defective longitudinal seam weld.
In 1992, PG&E found another flaw in a longitudinal seam weld on Line 132, the federal safety board said. But no documentation on that discovery has been made public, and PG&E’s leak and inspection logs do not mention it.
In October 2009, one month before its inspection report came out, PG&E drafted a long-term safety plan that covered the part of the line where the 1988 leak had occurred at a defective seam weld. It concluded, however, that the line was free of seam weld troubles.
“There were no leaks referencing material or manufacturing failure that had occurred since this pipeline was installed,” the report concluded.
That same month, a corrosion inspection team stumbled onto a defective seam weld on Line 132 in South San Francisco near the California Golf Club course, 2 miles north of where the line exploded less than a year later.
PG&E officials have said that because there was no leak at the South San Francisco site, they were not obligated to pressure-test Line 132.
Tracking leaks is critical to knowing the health of a gas transmission line, said Richard Kuprewicz, a pipeline safety expert who is monitoring PG&E’s gas-safety planning as a consultant to a consumer advocacy group, The Utility Reform Network.
“Here, you have seriously incomplete forms related to a very serious subject – leaks and repairs,” Kuprewicz said. “Not having complete reports, it’s an illusion – the most dangerous of all illusions, the illusion of safety.”
The fact that PG&E would rely on such incomplete reports to push forward with a corrosion-only inspection, he said, “raises serious questions here. The safety culture appears to be fatally flawed.”