On the morning of June 13th, 2013, a massive explosion rocked the Williams Olefins plant in Geismar,
killing 2 men and injuring 167 workers. It also released 30,000 pounds
of flammable petrochemical products and by-products into the air that
day. The plant shut down for 18 months to make repairs and investigate
Arnold & Itkin has just successfully resolved one of the first lawsuits
brought against Williams Companies, Inc. and Williams Olefins LLC (the
child company who owned the plant) on behalf of four workers. For us,
this case became a prime example of what industrial workers face and why
they deserve better safety standards.
“Process safety management program weaknesses at the Williams Geismar
facility during the 12 years leading to the incident caused the reboiler
to be unprotected from overpressure.” – Chemical Safety Board
Today, the Chemical Safety Board released a study that may prevent workers
from suffering the same injuries in the future. Part of the Chemical Safety
Board’s mission is to thoroughly investigate the cause of disastrous
events, and provide a plan for companies so they don’t happen again.
Now that their findings are public, we believe it would benefit our readers
to know exactly what occurred and why.
But first, a little background.
The History of the Geismar Plant
The Geismar plant was built in 1967 and passed ownership from company to
company until 1999, when Williams Olefins bought it. It was originally
designed to produce 600 million pounds of ethylene a year, but the plant
was upgraded to double its production of ethylene (and produce propylene).
The desire for increased output is where this story really begins:
The movement from original design to upgraded capability requires an MOC,
or a Management of Change. The evaluation outlines the hazards, procedures,
and work required for the desired update. It should be written prior to
an upgrade to ensure the safest possible transition, whether it involves
equipment, personnel, or procedure.
In 2001, the plant made an upgrade to the reboiler system next to a propylene
fractionator, or distillation column. The distillation column is a tall
tower where a propane mixture will be heated, separating propane from
propylene. The propane sinks to the bottom of the distillation column.
This 2001 upgrade is the key to understanding the explosion 12 years later.
Why the Reboilers Were Changed
Propylene production requires a powerful heat source to work—that’s
what reboilers do. The reboilers are 24-foot long holding tanks where
hot water will heat propane as it's sent into the fractionator. The
hot water is contained in tubes in the center of the reboiler, while the
propane flows around the tubes in order to heat up.
The water that enters the reboilers contains oily tar. Over the years,
tar creates buildup in the tubes, making the system less efficient. This
is known as “fouling.” Every few years, the plant would stop
production and shut off the reboilers in order to clean them. To keep
production moving, Williams decided to change the system.
In 2001, they added valves to each reboiler—valves to stop the water
from entering and leaving, and valves to prevent propane from entering
or leaving. The idea was to isolate one reboiler and keep the other operating.
That way, when a reboiler would foul, they could engage the other reboiler
and disengage the one that needed cleaning. This would allow them to never
cease production of propylene.
It came with a risk though: if the water valve was ever opened while the
propane exit valve was closed, it would isolate the reboiler from its
most crucial component—the safety valve at the top of the fractionator.
The Events on the Day of the Incident
On June 13th, operations and maintenance personnel convened for their daily morning
meeting. The plant manager noticed that the water flow to Reboiler A had
declined throughout the day on June 12th. The Operations Supervisor, Scott Thrower, volunteered to troubleshoot
the issue. Scott found that fouling was likely the issue.
He attempted to contact the Operations Manager to get the right personnel
to take care of the issue. However, the manager was unavailable. Scott
returned to the reboilers, where Reboiler B was on standby, as per procedure.
He turned on the water valves in order to begin operation of the standby tank.
What Scott didn’t realize was that Reboiler B had no access to a
pressure safety valve—and it was 65% full of liquid propane. As
the hot water heated the propane, it expanded with nowhere to go. The
pressure climbed to 1,200 psi when the tank ruptured. The fireball exerted
enough force to launch a reboiler pipe 30 feet in the air.
The water valves were turned on at 8:33 am. It only took until 8:36 am
for the tank to explode.
The resulting fire took three and a half hours to put out, but the damage
was done. A nearby operator who was working on the fractionator was killed
on the scene—Scott sadly succumbed to his injuries the next day.
While there were many factors to this event, it can be boiled down to a
few critical elements:
The First Mistake Was 12 Years Earlier
Rather than fully evaluate the risks posed by such an upgrade, Williams
installed the valves,
and wrote the MOC after the fact. The hazards of the upgrade were never reported or addressed, upgraded
procedures were never written, and a deadly new risk flew under the radar.
The Reboiler Should Have Never Ignited
One new procedure was written to ensure that the reboiler on “standby”
was properly prepared—all valves would be shut, and the reboiler
would be filled with nitrogen (a highly stable gas). There it would sit
until the other reboiler fouled. However, one important fact was left
out: the block valves were known to leak.
While these valves leaked within industry regulations, Reboiler B slowly
filled with propane during the 16 months it was offline and displaced
the nitrogen. What should have been a tank of inert gas was instead a
pressure bomb. In addition, the plant had no tools with which to monitor
what was inside the reboiler.
The Repair Was Completed Incorrectly
In 2006 and 2008, internal and external sources recommended the same thing
regarding the reboilers: installing car sealing open reboiler block valves,
which would give the tanks an open path to pressure relief. In 2010, the
car seal installation was applied…but to only the active reboiler,
which defeated the purpose of the recommendation in the first place. Again,
an MOC was never performed for the installation.
In 2011, a Process Hazard Assessment tragically reported by mistake that
both reboilers were car sealed when they were not.
No Action Was Taken Upon Discovering the Mistake
In May 2012, only a year before the fated explosion, the engineering records
coordinator discovered that only the active reboiler was car sealed. The
records were updated, but plant management approved the change regardless—even
though the full repair was
There were many moments where this crisis could have been prevented. Events
had to align in a specific way for a disaster to occur—that’s
perhaps why the explosion occurred 12 years after the upgrade. There’s
no use in speculating what could have happened or what might have happened,
but even the absence of the Operations Manager that morning tempts readers
to believe this explosion was fate.
It wasn’t fate. It was negligence.
The confluence of these elements—of not acting on safety recommendations,
of treating safety regulations as an after-the-fact obligation rather
than a legitimate tool, of not responding to the findings of their own
employees—is what led to the explosion. Simple, human mistakes.
Poor safety culture will
always lead to preventable tragedy, whether it takes 2 weeks or 20 years. It doesn't
make the result any less heinous.
Ultimately, that’s the lesson the Chemical Safety Board wants to
teach companies. Safety protocol isn’t an obstacle to overcome,
it’s not a checklist to fulfill out of obligation, and it’s
not something that needs to be outsmarted or outmaneuvered. It's the
most important practice that a business could have.
Whether we acknowledge it or not, it’s the only thing keeping our