A pipeline pig (probably a “dumb” one). Photo credit: Harvey Barrison https://commons.wikimedia.org/w/index.php?curid=6032423

Sunoco promises to check its pipelines carefully for cracks and leaks. Before installation, the company says, it tests the welds using x-rays, and it uses a water-pressure test for leaks.

Once the pipe is in the ground, though, testing is more difficult. Large ruptures can be detected because of pressure changes. But smaller flaws are harder. Detecting them requires running a “smart pig” through the pipe to detect cracks. A smart pig is an electronic device that can detect cracks and changes in wall thickness using ultrasound or changes in a magnetic field. Once the pig has been run through the pipeline and recovered, the data retrieved from it is processed by computer software that interprets it and produces reports on the state of the pipeline.

Sunoco is required to do these smart-pig inspections periodically (at least once every 5 years), and Sunoco always mentions this process in response to safety-related questions. But how good is that technique at actually catching problems? Will it prevent a catastrophic leak in the Dragonpipe (Mariner East 2 pipeline)?

“As scientific as a roulette wheel.” A Wall Street Journal article published in 2013 (“Oil-Pipeline Cracks Evading Robot ‘Pigs’”) cast doubt on the ability of smart pigs to find cracks. It discussed spills (including one that sent 21,000 gallons of crude oil into Michigan’s Kalamazoo River) due to flaws that smart pigs had failed to detect.

A pipeline consultant, Don Deaver, is quoted in the article as saying that using smart pigs to find cracks is “as close to scientific as a roulette wheel”, but even so it is still the best tool the industry has. Testing by the Pipeline and Hazardous Materials Safety Administration (PHMSA) revealed that “smart-pig scans sometimes fail to detect anomalies.”

Do the pipeline companies really want to know about defects? Prior to that 2013 article, investigative reporter Greg Palast had already reported on pig failures, and he referred back to his early work when he wrote about the recent Keystone pipeline spill. His reporting gives cause for concern. The problem is not just that the technology doesn’t always work.

What Palast reported was that one smart-pig company sells a device with software that is known to be defective, causing it not to report problems in some cases. He claims the software developers at the company knew about the bug and created a fix for it, but the company decided not to implement the fix.

Palast reports that the company did not fix the bug because its customers (the pipeline operators like Sunoco) prefer NOT to know that their pipelines have cracks and leaks. He sees this as an economic decision: it is cheaper for them to deal with the aftereffects of a spill or explosion—even massive legal settlements for injuries and deaths—than to shut down the pipeline, dig it up, and fix the pipe before the problem reaches that stage. Using a pig with faulty software, they can meet the regulatory requirement for pig inspection and still not have to deal with problems that really should have been detected and addressed.

Palast lays out the story in his 2011 book “Vultures’ Picnic”, which I can recommend. (It is available from the Chester and Delaware county library systems.) A whistleblower (one of the programmers who worked on the pig software) approached Palast with the story. The programmer came forward because he felt remorse after the San Bruno natural gas explosion (see “What happened in San Bruno could happen here”) which he thought might have been prevented by a proper pig inspection. He felt guilty about the many injuries and 8 deaths that occurred in San Bruno. In principle, Palast reports, most cracks and leaks ought to be detectable by a smart pig long before there is any detectable external leakage.

PHMSA knows that pigging is not enough. PHMSA (the Pipeline and Hazardous Material Safety Administration mentioned earlier) is the federal agency in charge of pipeline safety standards. In the wake of the San Bruno natural gas explosion, Congress told PHMSA to require periodic pressure testing of natural gas pipelines, although not liquids pipelines (like the Dragonpipe). But subsequently there were two major ruptures of liquids pipelines, in 2007 and 2013, and both happened after smart-pig inspections that failed to find the problem (details are in this article).

Recognizing the need for better inspection, PHMSA launched a rewrite of the pipeline safety rules for hazardous liquids in 2015. It was called the “Hazardous Liquids Integrity Verification Process”. That effort sounded promising, but it disappeared into a black hole—nothing has been heard of it since 2015.

Sunoco’s record proves their inspections are not adequate. In practice, multiple leaks on the Sunoco Mariner East 1 have been discovered (by third parties, not Sunoco) only after substantial spills—one of them as recently as April of this year—that occurred after pig-based inspection failed to reveal a problem. And Sunoco’s record for leaks per mile of pipeline is the worst in the industry. Clearly, the company cannot be counted on to prevent serious leaks.

So the next time someone from Sunoco claims they keep their pipelines safe with smart-pig inspections, ask them how many times they have actually dug one up to fix something detected by the pig. Then compare that with the number of leaks that Sunoco pipelines have sprung. That will tell you something about how good their pig-based leak detection system is.

The bottom line: despite what Sunoco may say, there is no safe way to put a pipeline this dangerous through a populated area. It must be stopped.