The siting and fire safety studies commissioned by Puget Sound Energy’s proposed liquified natural gas plant on the Tacoma Tideflats were thorough evaluations of the common risks associated with such plant, according to industry experts asked to review the work by The News Tribune.
The studies contemplated broken pipes, leaks and other accidental spills of LNG involving potential failure of every pipe larger than two inches. The studies say such leaks would be contained on the site.
The documents spend far less time considering the effects of an unlikely but catastrophic failure of the plant’s massive LNG storage tank, according to the two engineering experts who reviewed them. That situation is analyzed only as a breakage of its inner tank or as a top-burning fire that firefighters would have to watch burn itself out.
One of the experts expressed concern the studies did not delve deeply enough into that potential situation, even though he agrees such an event is extremely unlikely.
Some Tacomans opposed to the plant have cited a failure of the 8 million gallon storage tank as a primary fear about the LNG to be produced at the $275 million, 30-acre plant proposed for the Port of Tacoma.
“They did their due diligence, I believe, on these reports, but they’re essentially not seeing the tank failing as a credible scenario,” said Peter McDonough, a Salt Lake City-based engineer who has worked extensively on how gas plants can withstand disasters, including seismic hazards.
McDonough said the likelihood of the tank sustaining that level of damage is “very, very, very, very slight” and would require a catastrophic event, such as a plane crashing into it.
“They do not see that as a credible event in their planning, and I probably would agree with that,” he said. Nonetheless, he said, the possibility deserved a deeper look.
The disaster scenarios that were analyzed for a tank nearly as tall as the Tacoma Dome would not create harmful effects beyond the plant’s property line, according to the studies, which PSE paid the international engineering firm Chicago Bridge & Iron to conduct. The experts said that nothing in the study gave them reason to question this finding.
The siting study calculates that a tank fire in which the roof was destroyed could have a flame more than 200 feet long at a nearly 50 degree diagonal tilt. The study does not say how long such a fire could burn.
“If for some terrible, fluky reason that were to ignite, basically all bets are off for the containment of the fire,” McDonough said, “but the computer simulations are predicting that damages would be confined to the property itself.”
A disaster of that scale, the experts emphasized, is highly unlikely. The tank has a double-walled construction, and the industrial precautions LNG requires are not new. If LNG is spilled, it becomes flammable after it rapidly warms to a gas and mixes with air — which eventually dilutes it too much to burn. The gas can ignite when it comprises between 5 and 15 percent of the mix in a volume of air. It can explode if ignited in an enclosed space.
“We’ve been running this stuff over 60 years, so there shouldn’t be any problem,” said John Y. Mak, technical director with Texas-based engineering and construction company Fluor and co-author of the 624-page book “Handbook of Liquified Natural Gas.”
Mak said the studies focused on “basic calculations” that explored mainly smaller-scale industrial mishaps. His company has worked on LNG plants around the world for more than 50 years, and he said the proposed Tacoma storage tank is “quite small” by the standards of the industry. The safety evaluations in the Tacoma siting study are not as extensive as those at larger plants, Mak said.
“Pretty standard,” Mak said. “Nothing special in there. Just sort of boilerplate stuff.”
PSE spokesman H. Grant Ringel said the siting study is more thorough than that description indicates.
“It was a systematic review of the specific design of the PSE project in Tacoma,” Ringel said, “and it looked for specific potential failure points at our facility and studied those points.”
SMALLER LNG LEAKS AND SPILLS
Ringel said the plant design will mean any spill is shut off well before the 10-minute spillage scenarios written to adhere to federal standards. The studies showed spills would be contained on-site.
Gas dispersion models in the study show that the flammable vapor cloud from any envisioned spill also would be contained within the plant site, with proper design. The models examine where the gas cloud would go with winds of 1 or 2 miles per hour.
“It would have been very relevant for calculations to also have been made at much higher wind speeds,” particularly at a waterfront site, McDonough wrote.
The gas dispersion aspect of the study, which was performed by Gexcon US, does include a note that their models showed lower wind speeds resulted in flammable vapor clouds stretching for longer distances, making them a larger concern. Faster winds disperse the gas into the atmosphere more quickly.
The dispersion study notes that “vapor barriers” — basically, chain-link fencing with slats — would help keep dangerous vapor leaks from escaping the plant site.
Such fencing can slow the along-the-ground movement of clouds of the chilled gas, created when LNG warms above -260 degrees Fahrenheit, while allowing wind to help dilute the mixture.
That could work for smaller spills, McDonough and Mak both said.
The chain-link fencing is not a strong security barrier, Mak said. It can, however, offer a structural advantage over more solid construction when dealing with LNG.
A paper expressing concerns over the Jordan Cove LNG facility in Oregon — for which FERC denied a permit — by two industry experts argued that building a tall spill-containing wall out of impervious concrete would create worse problems. The authors, chemical engineer Jerry Havens and the late mechanical engineer James Venart, wrote that a solid vapor barrier would “increase the potential for damaging vapor cloud explosions” by holding the vapor cloud from LNG and other volatile gases in one place to accumulate.
Experts said one small-leak response described in the PSE studies departs from common industry practice.
The fire protection evaluation says a small leak might be temporarily stanched by “several wraps of a towel wetted with water.”
Both McDonough and Mak said this is not a normal practice, and McDonough called it “downright dangerous” for a worker who might try it.
“It is also unlikely that a towel could be held against any size leak and provide a secure closure at anything but the lowest of pressures,” McDonough said.
IF THE BIG TANK FAILS
Although no structural weakness of the massive double-walled storage tank is identified, the most disastrous scenario in the studies is if the tank somehow ruptured.
“The failure that is studied for this particular thing is the inside tank failing,” Ringel said, “because if that were to happen, which is an incredibly remote possibility, the outside tank then steps in to contain the spill.”
The siting study does not measure what would happen if the outer tank was also broken.
“Failure of both of them is not considered — and this is by federal guidelines for the study of these things — is not considered a credible event,” Ringel said.
The reports take up the possibility of the tank catching fire with a roof failure. In such a case, the fire protection evaluation says, “there is no firefighting technique … that can be employed to put out the fire” except letting the fire burn itself out after a plant evacuation.
“They say, ‘Well, if it were to fail, basically hang on to your shorts,’ ” McDonough said.
In a follow-up email, he added that he was “a bit concerned” that the disaster scenario of “a very low probability catastrophic leak of the main tank” was not fully explored, especially in the spill-modeling computer animations.
“Although it is highly unlikely, I think it would have been prudent of the owners to make a computer simulation of a failure of the tank at several wind speeds,” McDonough wrote. “They ran a number of studies showing small failures, which show containment of these leaks.”
He added in a follow-up email that a tank-failure computer animation could have backed up the fire protection evaluation drawings that assert a disaster-level fire’s heat would not get off-site.
“They were looking at higher-probability failures,” McDonough wrote, “but I think the low-probability, high-consequence event should have been addressed better.”
Ringel, the PSE spokesman, said the studies’ examinations of an inner tank failure and of a roof failure on the tank, along with a range of prolonged spill simulations, meant the key areas of potential problems had been thoroughly analyzed.
“We chose very conservative assumptions there to study really truly worst-case scenarios,” Ringel said.
The fire protection evaluation mentions the possibility of evacuating “all members of the public near the facility” in the event of “an uncontrollable liquified natural gas spill” from the main tank, but no off-site evacuation zone is defined in either document.
Evacuation plans for people on and beyond the LNG plant site are expected to be completed in 2018, Ringel said.