Answering a central question surrounding the San Bruno natural gas explosion that killed eight people and destroyed 38 homes in September, newly released documents show that PG&E employees, not outside contractors, installed the compromised pipeline as part of a construction job 55 years ago.
"It was PG&E that did the work in 1956," PG&E spokesman Brian Swanson confirmed late Tuesday.
The information, including internal company work orders, is contained in 225,000 pages of documents that PG&E released Monday evening in response to an order from state regulators who are investigating the explosion. The records may increase pressure on PG&E to reveal more information about the crew that did the work, including other jobs they might have performed in Bay Area neighborhoods.
The documents also reveal that PG&E for decades has suffered scores of welding problems on its natural gas transmission lines -- including the poorly welded pipe that exploded Sept. 9 in San Bruno -- raising serious questions about the company's decision to rely heavily on an inspection method ill-suited for spotting flawed welds in its underground pipes that crisscross neighborhoods throughout Northern California.
It was details about construction of the doomed pipe in 1956, however, that made the most impact Tuesday.
"People are going to want to know where else in the PG&E pipeline system did this particular crew work," said Mark Toney, executive director of TURN, The Utility Reform Network, a consumer group in San Francisco. "And those pipe sections would be a priority for inspection. It seems like they ought to be top of the list. And if they can't find out the exact crew, all pipeline installations done in that time period ought to be looked at."
Many flawed welds
During federal hearings in March, former welder Frank Maffei, now 80, came forward to say he worked on the job and that the crew was made up of PG&E employees. But PG&E's top executives until now had refused to confirm his account or provide details about who installed the 30-inch steel pipeline, under what circumstances and other key specifics, claiming the matter was under investigation.
"Our records review is still ongoing," Swanson said Tuesday. "We've really had a full-court press since the tragedy to review, research and verify the completeness of our records, and we're still in the process."
Responded Toney of TURN: "Nine months is a long time for PG&E to figure out that they are the ones who installed the pipe."
In January, metallurgists working with the National Transportation Safety Board found the section of pipeline, known as line 132, that exploded in San Bruno's Crestmoor neighborhood was made up of six unusual small sections, joined together with more than 150 flaws in the welds. The pipe failed on a longitudinal seam, while operating below its maximum allowable pressure.
The line that exploded runs from Milpitas to San Francisco, providing gas for heat and cooking to hundreds of thousands of Bay Area residents. It was built in two phases, in 1944 and 1948. In 1956, when a new San Bruno subdivision was being built between Skyline Drive and Crestmoor Canyon, PG&E employees relocated nearly 1,900 feet of 30-inch gas mains through the neighborhood. They did not buy new pipe for the job but used existing pipe they had left over from other jobs in 1948, 1949 and 1953.
The newly released records also show that the relocated line was tested for leaks using a "soap test," in which soapy water was sprayed on the pipes to look for bubbles. PG&E's attorneys wrote that the soap test "was a common method for identifying welding leaks during that era.''
Bills, work orders and other company records document how the San Bruno relocation job was assigned to PG&E's General Construction Department in July 1956, and overseen by division supervisor G.H. Taylor, according to documents PG&E turned over to the Public Utilities Commission this week. The job cost $55,405.
State regulators in February demanded all of PG&E's documents detailing "weld failures or defects" from 1955 through Sept. 10 after federal investigators found welding flaws in the line that exploded.
The mountain of records released this week, if stacked in one pile, would reach 75 feet high.
The company's response makes it clear the welding imperfections were hardly unique. Weld problems linked to gas leaks were identified in more than a dozen cities, including San Jose in July 2010, Oakland in 2005 and San Rafael in 1997 and 1998. The flaws also were found on line 132.
Besides two other welding problems previously revealed on that line -- one in a 1988 leak in San Mateo and another in a 2009 leak in South San Francisco -- the documents said an additional leak happened on line 132 in Sunnyvale in 1968.
A 2002 record also described a "substantial leak history" in a Palo Alto portion of line 132 that had been installed in 1947 with "nonstandard" welds. More red flags were raised in an October 2009 report about 3,590 feet of line 132 that was built from vintage 1944 pipe. Although the exact location of the pipe wasn't given, the report expressed concern that those portions of the line might suffer weld failure.
It also wasn't immediately clear from the documents how many bad welds overall the company has detected over the years. But the many flaws that were evident from an initial review of the records suggest the company seriously erred in having primarily relied on an inspection method known as direct assessment, which is good for detecting corrosion but not weak welds, according to Washington state pipeline safety expert Richard Kuprewicz.
Given its persistent welding woes over the years, PG&E had a duty "to really be looking for stuff like that, trying to pick the right tools to capture it," he said. "Direct assessment is an inappropriate technique."
Asked if PG&E should have used other inspection techniques more suited to finding imperfect welds, such as robotic devices known as pigs and pressure testing pipes with water to check for leaks, PG&E spokesman Swanson said the company already has acknowledged "that our past operations need to be improved." He added that "we're learning valuable lessons" from the investigation into the San Bruno disaster.
Contact Paul Rogers at 408-920-5045.
The information PG&E gave state regulators Monday in 103 gigabytes of files includes 16,000 documents on 225,000 pages.
The records are in response to a Feb. 24 order by the California Public Utilities Commission to provide details on all of the utility's "gas pipe weld failures or defects" from 1955 to September 2010.
The files represent only a portion of what the commission demanded. PG&E has said it will need until Sept. 30 to submit records for all its urban gas lines and another 15 months to provide the data for all of its transmission pipes.