Deadly Hydrogen Sulfide Accident Puts Deer Park in Headlines Again

 

A friend once summarized much of engineering ethics to me in two words:  "No headlines."  If that's a good guideline for engineering ethics, the city of Deer Park has seen two major violations of it in less than a month. 

 

Following a giant pipeline fire that burned for four days in September, on last Thursday, Oct. 10, some contract employees at the PEMEX refinery in Deer Park were working on a pipe flange, and something went wrong, releasing the pipe's contents into the air.  They may or may not have known that the pipe was carrying hydrogen sulfide (H₂S), which is a byproduct of oil refining.  It is probably familiar to most readers as the "rotten egg" odor that comes from aged chicken products and sewer gas.  The human nose can detect it at concentrations as low as one part per billion.  Unfortunately, one of its toxic effects is to deaden the olfactory nerves, causing the perceived smell to go away and leading to a false sense of security as concentrations increase.  It is highly toxic, and concentrations as low as 100 parts per million are classified as "immediately dangerous to life and health."

 

Two contract workers died in the accident, which occurred around 4:40 in the afternoon, and 35 others were exposed to the gas to the extent of needing treatment.  The bodies were not recovered until 3:30 AM the next day after the area had been cleared of toxic gas.

 

The city of Deer Park sent out shelter-in-place orders to its residents around 6:30 PM, but due to technical difficulties with the alert system, some people were not alerted until they read about the incident on social media.  A supplemental siren system in the city is due for an upgrade soon.

 

The accident is under investigation by both local authorities and the U. S. Chemical Safety and Hazard Investigation Board, but no cause has yet been determined. 

 

As we noted two weeks ago when writing about the pipeline fire, residents of Deer Park and surrounding communities in Houston are no strangers to refinery-related emergencies.  Release of toxic chemicals in oil refineries has been happening ever since there were refineries, and the industry has adapted to steadily increasing standards for air and water pollution control and safety measures over the decades.  The PEMEX refinery where this accident occurred dates back to 1929, when it was built by Shell Oil.  In 1993, Shell sold half the facility to the Mexican national petroleum company PEMEX and operated it as a joint venture until 2022, at which point Shell sold its share and PEMEX became the sole owner and operator.  This change of ownership may or may not have anything to do with the accident, but management cultures can change with ownership changes, and the upcoming investigation may answer that question as well as many others.

 

Considering the extremes of temperature and pressure under which highly flammable and toxic chemicals are processed in refineries, it's a wonder that we don't have a refinery explosion every day.  But it's the job of engineers to make sure that every possible thing that can go wrong in a refinery is anticipated and forestalled, and far more than 99.9% of the time, this foresight prevents mishaps. 

 

As with airline accidents, by far the most frequent cause of chemical-plant accidents is human error rather than a simple failure of machinery, although the two can be mixed.  The accident in question could have happened because the workers involved misidentified a flange to be worked on.  An oil refinery is one of the most complicated pieces of plumbing on earth, with thousands of valves, flanges, pipes, processing units, and interconnections.  Refinery workers have to know exactly what they are dealing with before taking any action that could conceivably release a product, and it might have been a case of simply opening the wrong flange.  Or an operator may have believed that the pipe in question had been purged of H₂S when in fact it hadn't been.  You can't tell the contents of a steel pipe just by looking, so there must be elaborate protocols in place to verify what is where, especially when maintenance operations are in progress. 

 

It is incidents like this one which make refineries and petrochemical plants high on the NIMBY list—"not in my back yard."  Given that a country wants to have fossil-fuel products, and given that it has considerable expertise and resources to make them, we in the U. S. must have refineries somewhere.  According to a list of new refineries compiled by the U. S. Energy Information Agency, the U. S. refining industry has managed to add considerable refining capacity since 2014 by building new refineries, but they tend to be in or near existing ones—Houston, Corpus Christi, or various locations in Alaska.  It's a lot easier to upgrade an existing refinery or build a new one next to an existing one, than it is to install the infrastructure of pipelines and shipping facilities in a place without refineries at all.

 

For the foreseeable future, the global economy will rely on fossil fuels, and so we will have to put up with refineries and everything that goes with them.  But people who live near them and work in them have a right to expect that they will be operated as safely as human ingenuity can manage. 

 

That was obviously not the case at the PEMEX plant last week.  We will follow this accident in the future, and when the investigation concludes, perhaps we will learn what chain of events led to an accident that killed two people and endangered an entire community.  But until then, we can take some comfort in the fact that refineries rarely show up in headlines, despite all the dangerous stuff going on in them.

 

Sources:  I referred to articles on the H₂S accident by ABC News 13 in Houston at https://abc13.com/post/pemex-chemical-leak-crews-waiting-lower-levels-before-entering-unit-center-deadly-hydrogen-sulfide-deer-park/15416337/, an article in the Saturday Oct. 12 edition of the Austin American-Statesman, "2 dead, dozens of others injured in hydrogen sulfide leak near Houston," and the Wikipedia article on PEMEX Deer Park.  The Energy Information Agency data is from https://www.eia.gov/tools/faqs/faq.php.

Nearly $1 Million In Fines Proposed for Deadly Liquid-Nitrogen Accident

 

Last January 28, workers at the Gainesville, Georgia chicken-packing plant operated by Foundation Foods began to show up for work.  Several workers were assigned positions in the below-ground-level freezing room on Production Line 4, one of several in the large plant.  Not too long before, the old ammonia-refrigerant freezing system had been replaced with a new liquid-nitrogen (LN2) system.

 

On the face of it, liquid nitrogen is a safer chemical than ammonia, which is both toxic and explosive.  Nitrogen makes up 78% of the atmosphere, and while liquid nitrogen can cause painful burns, as long as it is confined to locations inaccessible to workers, no direct harm from it can result.  However, breathing nitrogen gas cannot support life.

 

The freezing stage itself was in a separate freezing room, and consisted of an immersion freezing line in which the chicken pieces were immersed in liquid nitrogen at around -321 F, followed by a spiral freezing chamber where the gaseous nitrogen boiling away from the liquid surrounded the pieces as they spiraled upward on a conveyor belt. 

 

Just after the shift started, Line 4 had a problem, and three maintenance workers entered the freezing room.  What they apparently didn't know was that the malfunction had filled the entire room with evaporated liquid nitrogen, displacing most of the oxygen in the room.  The workers immediately passed out and died shortly thereafter.  Two more workers entered the room and died, and a sixth was sufficiently overcome that he died on the way to the hospital.  The release of liquid nitrogen created a condensation fog that made escape difficult, and at least twelve other workers were injured enough to require hospital care. 

 

Both the U. S. Chemical Safety Board (CSB) and the Occupational Health and Safety Administration (OSHA) began investigations of the accident.  In February, the CCB issued some preliminary updates, and on July 23, OSHA announced its intention to assess Foundation Food Group Inc. (the plant operator), Messer LLC (the LN2  supplier), Packers Sanitation Services Inc. (the cleaning and sanitation service provider) and FS Group Inc. (the equipment and mechanical servicing provider) a total of nearly $1 million in fines.

 

OSHA found 26 violations committed by Foundation Food Group.  The agency discovered that the workers at the plant had received no training or warnings about the hazards of LN2 and safety precautions they should observe.  There were no written lockout procedures dealing with how maintenance should be performed.  While the following is reading between the lines, something like this accident could have occurred if someone had shut off the LN2 supply to the system while leaving an LN2 valve or joint open in the freezing room, and then someone else came along and opened the shut valve, not knowing it would kill six co-workers. 

 

Lockout procedures are vital for work on hazardous equipment of all kinds, ranging from conveyor belts to high-voltage power systems.  If equipment is being worked on, a typical lockout procedure specifies brightly-colored tags that must be placed on critical controls in a way that will prevent the control from being operated.  Evidently, nothing like this was done.

 

Another standard safety measure is to train everyone who has a reason to work inside a confined space where hazardous gases (such as pure nitrogen) can accumulate.  Ideally, an oxygen monitor or some way to indicate a problem with the air inside the room should have been operating, but at least the workers should have been made aware of the possible danger.  That way they might not have simply walked into the room to their certain deaths. 

 

If a person strays into an area where there is no oxygen but the air pressure and other conditions seem normal, there is usually no time to react, turn around, or do anything except pass out and drop.  I once saw a dashcam video taken from a state trooper's car.  The trooper had stopped to investigate an accident involving a liquid-ammonia truck.  The gas had been released invisibly, and you can watch as the trooper walks toward the truck and then falls as if shot through the heart, and eventually dies.  Even the miserable coal miners of nineteenth-century England were able to bring a canary into the mine with them to give advance warning of "black-damp"—air without enough oxygen to breathe.  But canaries wouldn't be allowed by the microbiologists who rule meat-packing plants—too unsanitary.

 

While there was probably some mechanical malfunction that caused the initial release of nitrogen, the fatalities could have been avoided with better communication among all the parties involved.  The factory operators needed to give instructions in their employees' native language (Spanish) about the hazards of working around LN2.  The LN2 supplier needed to communicate with the factory operators and maintenance workers concerning the delivery and transmission of their product.  And the maintenance and supervisory personnel in charge of fixing Line 4 needed much better lockout and safety procedures, which were signally lacking in this tragic case.

 

Vigilance is perhaps not taught that much in business schools.  It sounds like an old-fashioned, rather static virtue, somewhat passé when compared with flashier virtues like competitiveness and enterprise.  But clearly, whatever safety procedures and safety officers were in place in the Gainesville plant on last Jan. 28 were sadly inadequate to prevent a tragedy that didn't have to happen.  Not everybody enjoys going around their place of business imagining what could go wrong and how to keep it from happening.  But that's the main job description for those charged with the safety of workers, and it's a job that looks like a loss on a company's balance sheet until something like the LN2 accident happens.  Then it's a lost opportunity, and the beginning of something much worse. 

 

Sources:  OSHA's press release on the fines assessed so far regarding this accident can be found at https://www.osha.gov/news/newsreleases/national/07232021.  Some preliminary findings of the CSB are at https://www.csb.gov/third-csb-update-on-poultry-plant-incident/.  I also referred to a news article at https://www.npr.org/2021/07/23/1019784347/six-poultry-workers-died-from-a-nitrogen-leak-osha-has-issued-1-million-in-fines.  I previously blogged on this topic on Feb. 1, 2021 at https://engineeringethicsblog.blogspot.com/2021/02/six-die-from-liquid-nitrogen-accident.html.

 

Questions Remain About Visakhapatnam Gas Leak

Here's what we know so far.  In the early morning hours of Thursday, May 7, workers at a petrochemical plant in the southeastern India coastal city of Visakhapatnam were trying to restart the plant, which had been shut down earlier due to lockdown restrictions imposed in connection with the COVID-19 pandemic.  Around 3:30 AM, a gas leak occurred and spread rapidly throughout the densely-populated area surrounding the plant.  At least 200 people wound up hospitalized, and as of May 8, 13 had died from the effects of the leak. 

The plant is presently owned by the South Korean firm LG Chem, but was founded in 1961 and has passed through several changes of ownership since then.  Its main output is polystyrene plastic, and to make polystyrene requires styrene, a benzene-like molecule that is liquid at room temperature.  About 2,000 metric tons of the monomer were stored onsite in tanks.

A Chevron safety sheet on the styrene "monomer" (what the molecule is called before it is polymerized into a chain) emphasizes the main danger from storing it:  runaway polymerization.  Most polymers have to be forced into polymerizing, but evidently styrene is an exception:  it will polymerize if given half a chance, and especially at temperatures above about 68 F (20 C).  When it polymerizes, it gives off heat, which makes it polymerize faster, and the resultant heat and pressure buildup can cause an explosion.

This is why storage tanks of styrene are normally refrigerated to keep them cooler than 20 C, so that spontaneous polymerization doesn't happen.  While the exact sequence of events is not yet clear, it appears that a computer glitch or other problem interfered with the refrigeration of the styrene tanks.  Once the temperature rose much above 20 C, polymerization in the tanks would have raised the temperature and pressure, and eventually a safety valve somewhere must have opened, or else a rupture in the tank or piping occurred.

At any rate, a large amount of styrene monomer escaped the limits of the plant and must have traveled hundreds of meters, affecting several villages that have sprung up around the plant in the sixty or so years since its founding.  Styrene, being heavier than air, sticks to the ground, and in sufficient density it will suffocate you.  But lower concentrations than that will still cause intense respiratory problems and death, as it did for 13 people that night.  Eventually, authorities evacuated a 3-km-radius area around the plant, but by that time most of the damage had been done.

While this accident pales in comparison to the well-known Bhopal tragedy of 1984 in which about half a million people were exposed to a toxic chemical and thousands died, even one death of a resident near a chemical plant is unnecessary.  What lessons can be learned so far from the LG Chem plant accident?

A common thread that shows up in many chemical-plant accidents is that they tend to occur when a plant is being started up after a shutdown.  There are several reasons why starting up is a dangerous time.  Conditions in the system have to be brought from a low-pressure, low-temperature state to operating pressures and temperatures without straying into combinations that can be dangerous to equipment or personnel.  This requires more than typical vigilance from operating personnel, who may not have experienced that many shutdowns and restarts in their careers.  The procedures for starting up a plant can be much more complex than those required to keep it running, and more mistakes can be made in a complicated, time-sensitive process than simply one in which your job is to make sure everything stays the way it is and runs smoothly.  Last but not necessarily least, it seems that a favorite time for doing a startup is after the beginning of the midnight shift.  Whether the implied secrecy of early morning is appealing in case anything goes wrong, or whether it is simply a more convenient time with regard to plant schedules, I don't know.  But from the viewpoint of sounding an alarm to the general public if anything goes wrong, the period from late evening to early morning is the worst possible time to do something that might cause problems to people outside the plant, who will all be asleep and hard to evacuate in an emergency.

Another factor in this accident is the presence of densely-populated villages just outside the plant boundaries.  According to one news report, in 1961 the region where the plant was erected was rural, but with the subsequent population growth of cities such as Visakhapatnam, that changed.  The permit status of the plant is reportedly in a legal gray area, which might result from the fact that if the plant were to be built from scratch today in the same location, it might not be allowed at all, or at a minimum a large buffer zone would be required between the active plant and the surrounding populated areas.  As is true in most parts of the world, the land surrounding chemical plants is where you find some of the lowest-priced housing, and the kind of people who live in low-priced housing are generally poor people.  While they are not happy to be taking an unknown risk of sudden death or long-term illness by living within the sights, sounds, and smells of a chemical plant, they may not have much of a choice.

At last report, an investigative team from LG Chem's South Korean headquarters was onsite trying to determine the accident's cause.  But that is little comfort for those who lost loved ones or the hundreds who were injured in this accident. 

Absolute safety in industrial processes is virtually impossible without exiling plants to an uninhabited island operated entirely by robots.  And in any case, such an operation would be undercut in cost by operations such as LG Chem that runs with human beings and in proximity to people who may not know they are taking a chance every day of their lives just by living close by.  In a sovereign nation, the only force that can generally make sure powerful manufacturing interests don't hurt or kill too many people is the various branches of government, with perhaps private insurance companies coming in a distant second.  I hope that this accident teaches all concerned—corporations operating in India, the government officials responsible for regulation, and the Indian people—how to do things better next time, and to make it a long time before the next such accident occurs.

Sources:  I relied on the Wikipedia article " Visakhapatnam gas leak," as well as reports on the website thehindu.com at https://www.thehindu.com/news/national/andhra-pradesh/visakhapatnam-gas-leak-lg-chem-dispatches-technical-team-from-seoul-for-investigation-rehabilitation-works/article31573936.ece and https://www.thehindu.com/news/cities/Visakhapatnam/visakhapatnam-lg-polimers-chemical-plant-gas-leak-updates-may-7-2020/article31523178.ece?homepage=true.  I also referred to a Chevron safety sheet on styrene at http://www.cpchem.com/bl/aromatics/en-us/Documents/Safe_Handling_and_Storage_of_Styrene_Monomer.pdf.

One Accident, Three Headlines

There are probably as many reasons that people choose a career in engineering as there are engineers.  But one reason that may be pretty common was voiced by a graduate student of mine many years ago.  When he entered college, he thought at first about becoming a philosophy major.  But when he discovered that, like lawyers, philosophers prided themselves in being able to defend either side of any argument, he felt that he needed to find a field where objectivity was paramount and subjective opinions had to defer to facts.  So he chose engineering.

My student was on to something.  Modern science-based engineering does start from universally-recognized facts about the physical world.  But if you go into engineering with the hope that technical matters are all that matter and personal subjective beliefs and feelings are irrelevant, you will be in for a surprise.  A good example of this came about last spring in news reports of an incident at the Arkansas Nuclear One plant in Russellville, Ark., about seventy miles northwest of Little Rock, on Easter Sunday, March 31 of this year.  What was meant to be a taxing but routine industrial maintenance operation turned into a fatal accident.  But the same facts were construed at least three different ways by different news outlets.

The first report I’ll consider is from the U. S. Nuclear Regulatory Commission’s official website, http://public-blog.nrc-gateway.gov.  This blog is clearly intended to be the NRC’s public face.  Their headline on the incident reads, “Easter Sunday and Arkansas Nuclear One.”  From the headline alone, you might expect it to be a caption of a nice photo of, say, the sun rising over a cooling tower as a group of worshippers nearby celebrate an Easter Sunday sunrise service.  Only when you read the text do you find that early that morning, a group of workers at the Entergy Operations Inc. plant were using a temporary crane to move a 500-ton piece of equipment called a generator stator.  This assembly of copper and steel is the stationary part of the electrical generator, and is just like any other stator in a coal-fired or gas-fired plant.  As they were moving the huge stator, the temporary crane failed.  Five hundred tons falling even a few feet will acquire a tremendous amount of kinetic energy.  When it hit the floor, pieces flew around the generator hall.  One worker—a 25-year-old Arkansas Tech grad named Wade Walters—was killed, and eight others injured.  As a result of the damage, the outside power to part of the plant was interrupted, and the plant’s emergency generators started automatically to keep things under control.  Once outside power was restored, things returned to normal and cleanup from the accident began.  By law, the plant operators were required to notify the NRC of this incident, and this information was posted both on the NRC’s technical incident-notification site and on the public blog.  The parts of the plant containing nuclear material were never in any danger of leaking anything as a result of this industrial accident.

The next report I consider is a Russian news website called RT.com.  Their headline reads, “Arkansas nuclear plant incident kills one, injures eight.”  The phrase “nuclear plant incident” will arouse in the susceptible reader images of Three Mile Island and Chernobyl, both accidents in which extensive damage resulted to the nuclear parts of the plant.  Three Mile Island, a nuclear plant in Pennsylvania that partially melted down in 1979, released a small amount of nuclear-laced gas, but caused no deaths or injuries.  The Chernobyl disaster, which happened in 1986 in Ukraine (then a part of the old USSR) was much more serious, killing dozens promptly and releasing large amounts of nuclear material that led to the condemnation of many square miles around the plant and a rise in radiation-related illnesses for years afterwards.  So Russians have good reason to be touchy about anything that goes wrong at a nuclear facility. 

The RT article plays up the human-interest side, with extensive quotations from those mourning the loss of Mr. Walters.  The general tone of the report can be assessed from the first quotation in it:  “ ‘We are deeply saddened by what has happened today,’ executive vice president and chief nuclear officer Jeff Forbes said in a public statement, without providing details about the cause of the young man’s death or the severity of the other victims’ injuries.”  Now, there may not have been any intent to cover up or deceive in the fact that Mr. Forbes did not provide the aforementioned details.  He simply may not have known enough accurate information to say, and in any event, the details of injuries to private individuals are normally considered privileged medical information.  But the way the statement is framed leads the suspicious reader to think Mr. Forbes may be hiding something. 

Finally, a website that may be fairly characterized as anti-nuclear leads their version of the story with this headline:  “Fatal accident at Arkansas Nuclear One leaves unit without offsite power.”  This is from something called Enformable Nuclear News.  I have been unable to determine the etymology of the word “enformable” which sounds vaguely French and might mean “able to inform.”  At any rate, Enformable.com has a large web presence and a substantial team of well-qualified reporters who apparently spend most of their time looking for nuclear-related bad news.  In this, they are no different than any other reporters—bad news travels faster, is more interesting to read, and makes up the bulk of all news. 

Notice that the Enformable report begins with the worst outcome of the incident:  “fatal.”  And while it is true that part of the plant was without external power for a very short time, the emergency generators started normally and allowed operators and automatic controls to take the appropriate actions that prevented anything out of the ordinary from happening to the plant’s nuclear operation.  But if you read the headline alone, you may get the impression that someone at the plant might have died from radiation or other specifically nuclear-related cause, and that the plant is still sitting there disabled through lack of outside power.

Public-relations people call such slanting “spin,” and for some time we have lived in an age of spin.  But once information reaches the public domain, it is no longer “objective” in some abstract, depersonalized sense.  Those who convey it to their various publics—and in the age of the WorldWideWeb, there are more different kinds and divisions of publics than ever before—will inevitably emphasize certain aspects of a story over others, and lead their readers to do the same. 

The engineer who thinks all this sort of thing is beneath contempt, and that a simple objective statement of the facts ought to clear up all dispute, is fooling himself (or herself).  While public opinion may be misinformed, distorted, or even flat wrong, everyone in the nuclear industry, including engineers, has to operate in the real world, not in some idealized technical utopia where everyone sees everything the same way.  And the sooner engineering students recognize and understand that aspect of engineering, the better. 

Sources:  I thank Andy Taylor of Entergy Operations Inc. for drawing my attention to this incident.  The URLs for the three reports are: http://public-blog.nrc-gateway.gov/2013/04/01/easter-sunday-and-arkansas-nuclear-one/ (NRC); http://rt.com/usa/arkansas-nuclear-plant-accident-170/ (rt.com); and http://enformable.com/2013/04/accident-at-arkansas-nuclear-one-leaves-unit-without-offsite-power/ (enformable.com).  I also referred to the Wikipedia articles on the Three Mile Island accident and the Chernobyl disaster.