ProPublica

Journalism in the Public Interest

Cancel

Fatal Accident Turns Attention to Nation’s Aging Pipelines

.

The remains of a gas line lie on the ground after an explosion September 10, 2010 in San Bruno, California. (Photo by Eric Risberg-Pool/Getty Images)

Following last week’s pipeline explosion in California that killed at least four people and destroyed 58 houses in suburban San Bruno, regulators have been searching for clues about what caused the blast. A piece in the Associated Press, meanwhile, suggests that thousands of pipelines like the one in San Bruno are aging and nearing their life expectancy:

Experts say the California disaster epitomizes the risks that communities face with old gas lines. The pipe was more than 50 years old — right around the life expectancy for steel pipes. It was part of a transmission line that in one section had an "unacceptably high" risk of failure. And it was in a densely populated area.

… Thousands of pipelines nationwide fit the same bill, and they frequently experience mishaps. Federal officials have recorded 2,840 significant gas pipeline accidents since 1990, more than a third causing deaths and significant injuries.

More than 60 percent of the nation’s gas pipelines are 40 years or older, AP reports. And age isn’t the only issue. Older pipelines are often configured in such a way that they’re unable to accommodate the latest pipeline safety technology.

Both The Wall Street Journal and Los Angeles Times have reported, for instance, that the pipeline itself was unusually built. The Times noted that the ruptured San Bruno pipeline, which was installed in 1956 by Pacific Gas & Electric Co., was so old and oddly configured that it could not accommodate a new pipeline-testing device known as a “pig,” which is considered to be “superior to other forms of testing.” 

"It's the best way to test," a former head of the government's pipeline oversight agency, told the Times. Other testing methods “can miss things that the pigs would not miss," he said, including "tiny, tiny losses" in the thickness of pipeline walls. The former official, Cesar De Leon, also told the Times that because replacing these older pipelines in order to accomodate the pigs would be so costly, owners aren't required to do so.  

According to PG&E President Chris Johns, the section of the pipe that exploded did not have such a device, but the testing methods used by the company do not “have any less capability.” He told reporters that the ruptured pipe was inspected twice in the past year, and those inspections turned up no problems.

As we’ve noted, the Pipeline and Hazardous Materials Safety Administration, the federal agency that regulates 2.3 million miles of oil and natural gas pipelines, largely relies on standards written by the oil and gas industry. It has about 100 inspectors, leaving industry a great deal of latitude with inspections. (Even after the blast, state utility regulators ordered PG&E to inspect its own network of gas pipelines.)

And according to The Washington Independent, federal regulators are required to inspect only about 7 percent of the country’s natural gas pipelines. That percentage is based on how populated the surrounding area is, and not the actual conditions of the pipelines.   

Couple things.

UT pigs are hardly new, they have been around for nearly 30 years and have been in widespread use since the early 90’s. While they do work very well, and very quickly, standard mag particle is more than sufficient.

Secondly I noticed the rather backhanded insinuation that the API’s standards for pipeline construction are somehow “tainted” because individuals from the O&G industry write them. The individuals who sit on the standard boards collectively represent hundreds of years in fairly narrow technical specialties directly related to the process. I would like to hear you elaborate on what specific portions of the various API codes you think are deficient (there are several pertaining to pipeline construction governing everything from material of construction, valveing, installation, sizing, inspection, etcetera). What would you, the bright and well educated Miriam Wang, change about ASTM E164-03, API 1104, or ASME B-31.3. What specific code deficiencies have you observed from your decades of designing, specifying, building, and operating pipelines? Specifics are always good in a debate like this. Perhaps you would like to debate with some of the hundreds of engineers who have contributed to these codes.

I am sure they would love that. These individuals put both their professional and personal reputations (as well as their professional credentials) on the line when writing these standards.

The United States has millions of miles of pipeline transporting everything from steam, to oil, to natural gas to anhydrous ammonia. An isolated incident does not make a trend.

And your earlier article’s claims about transparency in the publication and availability of standards and codes is complete hogwash. While bound copies of API/AGA/ASTM/ASME codes cost $100’s even $1000’s per copy, ANYONE can go down to a library and pick up a copy for free.

With regard to Mike H’s concluding comment that an isolated incident does not make a trend, the article states that there have been 2840 pipeline accidents since 1990, twenty years ago.  That’s an average of 142 accidents every year for twenty years, one about every two and a half days.  According to the article over a third cause death; that’s a fatal accident about once a week (7 1/2 days) for twenty years.  It not just a trend, but a deadly trend.

You know Ronald, I come to find when reading anything from Propublica a closer look of their “facts” is always warranted. According to the sources cited by Propublica (the Pipeline Safety Trust) there have been 2463 pipeline incidents since 2003. Out of these there have been 18 fatalities and 42 injuries. That averages out to 2.25 fatalities pr year. While 2.25 fatalities is certainly a tragedy on an individual level, considering that the US has millions of miles of pipelines I think the safety of the pipeline operators is quite good.

Ronald Harbin

Sep. 15, 2010, 7:13 a.m.

Thanks for your response, Mike.  I’d like to get into this a little more for a couple reasons.  One, I’m concerned about public safety, a concern I expect you share, and secondly I am an admirer of the efficient partnerships (MLPs) that operate many of the nation’s pipelines and have a significant investment in several of them.

You cite Ms Wang’s source as the Pipeline Safety Trust and she cites “federal officials”.  She uses 1990 as her baseline year and you use 2003.  Taking your figure from the Trust of 2463 accidents that’s 352 a year instead of the 142 you get using her figures.  Either way if her statement that over a third of accidents result in death is accurate the numer of fatalities is much greater than 2 1/4 per year.  If we assume only one fatality per fatal accident (there were 4 in San Bruno) that’s 947 fatalities in 20 years, 47+ annually,  using her numbers and 821 in 7 years, 119+ annually, using yours (again assuming that 1 in 3 accidents is a fatal one and that there was 1 fatality per deadly accident).  How good that record is could be debated, but its a lot more than 2 1/4.  My biggest concern is that If the claim that, like many of our bridges, this system has reached a point at which much of it has operated about as long as it could have been expected to safely we have a situaltion that is going to worsen.

On the crass level of economic self interest as an MLP investor I wonder if this is going to become a liability for operators of pipelines that explode.  Have pipeline owners been sued for negligence as a result of some of these accidents and are they likely to be in the future assuming that the lifetime of the pipes’ safe operation is up?  If these partnership need to replace pipe to insure the public safety I’m all for it, yet from an investor’s perspective that would limit the amount of cash for dividends, a primary benefit of MLP investment.

Ronald, I think the confusion in these numbers comes from the lack of clarity on the descriptor “significant” and the total fatalities. I have seen Propublica pull stunts like this in the past. A reader infers from the information presented that the danger is much greater than it actually is.

The database I found over at the Pipeline Safety Trust includes all reported pipeline incidents. These include “significant” and “non significant” incidents. The information I posted earlier was all incidents both “significant” and “non significant”. Compiling data from both sets of data available (1984-2004 and 2002-present) shows a grand total of 43 fatalities and 2038 injuries (using only 1990 – present data). That translates into 2.15 fatalities per year and 101.9 recordable injuries. This is for 5721 recorded incidents. I have no way of determining how many of these were “significant” as the 1984-2004 database does not break this down but the 2002-present database does. The grey are seems to be how many total fatalities there were and how many injuries there were.

Considering how readily available this information is, I am surprised by its omission from the article. Considering Propublica’s willingness and past track record of playing bait and switch with information though, I guess I am not too surprised.

To your point that the nations hazardous pipeline system is becoming more dangerous, there is much data contradicting that.

http://www.dot.gov/perfacc2006/images/pipelinebarchart.jpg

I would encourage you to dig into these numbers if you disagree. The can be found at the Pipeline Safety Trust’s website.

Ronald, I think the confusion in these numbers comes from the lack of clarity on the descriptor “significant” and the total fatalities. I have seen Propublica pull stunts like this in the past. A reader infers from the information presented that the danger is much greater than it actually is.

The database I found over at the Pipeline Safety Trust includes all reported pipeline incidents. These include “significant” and “non significant” incidents. The information I posted earlier was all incidents both “significant” and “non significant”. Compiling data from both sets of data available (1984-2004 and 2002-present) shows a grand total of 43 fatalities and 2038 injuries (using only 1990 – present data). That translates into 2.15 fatalities per year and 101.9 recordable injuries. This is for 5721 recorded incidents. I have no way of determining how many of these were “significant” as the 1984-2004 database does not break this down but the 2002-present database does. The grey are seems to be how many total fatalities there were and how many injuries there were.

Considering how readily available this information is, I am surprised by its omission from the article. Considering Propublica’s willingness and past track record of playing bait and switch with information though, I guess I am not too surprised.

To your point that the nations hazardous pipeline system is becoming more dangerous, there is much data contradicting that. I would post some, but every time I try and insert a link my comment is moderated. Google “pipeline incidents” and look at the images tab to see some interesting trend data from the DOT.

I would encourage you to dig into these numbers if you disagree. The can be found at the Pipeline Safety Trust’s website.

We saw what a good job the petroleum and gas industry did essentially self-regulating their drilling in the Gulf of Mexico.  We also see how open and honest the gas industry is with hydraulic fracturing.

Why should anyone trust these industries?  (We can’t even safeguard our food.)  Government doesn’t have the will or courage to do what is necessary to properly protect its citizens and the environment.

Add a comment

Email me when someone responds to this article.