92nd General Meeting Presentation
An Introduction to the U.S. Chemical Safety Board
Adam Henson
The following remarks were delivered at the General Session of the 92nd General Meeting on May 13, 2024. It has been edited for content and phrasing.
INTRODUCTION: Adam Henson is a chemical safety recommendations specialist at the U.S. Chemical Safety and Hazard Investigation Board (CSB). He gives presentations under the agency’s advocacy program to drive chemical safety excellence and protect communities, workers, and the environment. Henson has 15 years of experience in chemical safety and security. He also was a compliance officer for the Occupational Safety and Health Administration and a nuclear biological chemical defense specialist while serving in the U.S. Marine Corps.
His slide presentation can be found here.
MR. HENSON: Hello, everybody. I work for the United States Chemical Safety and Hazard Investigation Board. How many people in the room have ever heard of the CSB? That's better than what we get most of the time. Actually, I'll probably just leave then. I'll see you guys later.
In all seriousness, the reason the CSB exists is because every day people go to work in communities that have facilities where they face the threat of a chemical release. Starting in March 2020, we enacted a rule requiring owners/operators to report to us incidents that happen. And since that rule has come into effect, there have been 357 incidents where someone has either been killed or seriously injured or at least a million dollars of property damage has occurred. At least 40 more of these incidents have happened since last October. These numbers are from when I drafted this presentation two months ago, so there have been even more since then.
The CSB vision is a nation free from chemical disasters. And our mission is to drive chemical safety excellence. In simpler terms, based on the education and experience in this room and the interest, I would bet that just about everybody in this room has an idea that could make things better for people. They could make processes safer and safety inspections more efficient, but those ideas rarely come to light. The reason is either because of a lack of follow-through, a lack of responsibility, or the fear of feeling foolish.
My agency exists only to develop these ideas based on incidents that have occurred and to share them with the industry in the hope that they'll be implemented.
The CSB is an independent federal agency, meaning we do not report to a cabinet-level position. We report to a Senate Committee that reports to the president. As I said, we do root-cause investigations of these incidents to make recommendations in the hope that they'll never happen again. My job, specifically, is to go out and meet the people to help them implement the recommendations that come from the investigation side. A recommendation is a specific, measurable action designed to prevent a chemical incident.
We'll talk about this incident from the D.D. Williamson & Company incident. It was a catastrophic vessel failure in Kentucky that occurred in the early 2000s and killed one worker. On the morning of April 11, what ended up happening – do you guys know why this incident happened? In short, this incident occurred because a person got distracted while repackaging boxes and forgot to shut off the heat to a process vessel. In the time it took for them to notice that the process vessel was overheating, it was too late to do anything about it. They went to the top of this vessel, which did not have a pressure relief device, and attempted to manually release the pressure by opening the manual valve at the top of the relief vent, which, unbeknownst to this person, had been clogged.
Interestingly, the study notes there was no pressure relief valve on this piece of equipment. There’s the notion that if it had had one, then the consequences of the incident would have been mitigated and nothing bad would have happened, but there's that other line of thought: Who's to say that the line of the pressure relief valve wouldn't have been clogged as well.
The recommendations from this case went to the company, and we recommended that they get instrumentation and an alarm significant enough to be aware of these incidents before they turn catastrophic. What's interesting, too, is that this vessel was manufactured in the 1970s in the New York area and used for about 15 years in that area until eventually being shipped to Kentucky.
The company did not notify the jurisdiction of the installation of this vessel as required by state law. They also did not have this vessel inspected by a qualified inspector upon its installation, which is required by state law. The reason is because they had told themselves it was not a pressure vessel. That's it. That is a concept in safety that I like to call a “belief of convenience.”
It's not a pressure vessel. Why not? Because it has an open 1-inch line on the top of it. Except that every time we run this process, we close that 1-inch line for some amount of time, and we use an air compressor to charge it to 25 psi, but it's not a pressure vessel?
They circumvented all these existing laws, all these great practices, and all these great rules for reasons that no one will quite come to understand, except for a belief of convenience that this was not a pressure vessel.
These beliefs of convenience are going to come up again. I'd ask you to pay attention to these incidents as we describe them. There will be five of them and see if you can come up with any common themes. We'll go ahead and label this as the belief of convenience.
Another incident occurred in Houston, Texas, in the early 2000s. It had significantly less severe consequences, but it was nevertheless a very serious incident. These people make polyethylene wax. They have seven tanks. To move the process along, they pressure them up with nitrogen. The nitrogen generator that they used to create the nitrogen was described as small. It wasn't big enough or capable of operating the process the way that they wanted it to do, so what they would do periodically was arrange temporary piping and tubing to force nitrogen out of this thing using compressed air.
What does that sound like? We'll call that “get ’er done.” We can't do what we need to do with the equipment we have, so we're going to arrange for it to do it despite the consequences. What ended up happening was that using the compressed air to force the nitrogen obviously introduced oxygen into the process. Generally, as it was intended to operate, the oxygen level in this process should have been below 3%, which is low enough that it will not sustain combustion. Doing it the way they did, the level of oxygen jumped to 18%, which is more than enough to support combustion.
As if that were not bad enough in and of itself, what also happened to this vessel over time was that they were tired from having to re-raise the piping so frequently, so they decided to cut a hole in the vessel and hard pipe a way to do it, along with the process of adding a heating coil for a boiler. The people they hired were not qualified to perform this work. They were not National Board (NBBI) stamp holders of any kind. They were just a bunch of guys using some kind of torch. They opened this thing, added the heating coils, and closed it up. The people who did the welding were not qualified welders. The post-incident examination revealed that the weld penetration was only 25%, and it was riddled with porosity. Additionally, they didn't even bother with the basic surface preparation before starting.
What ended up happening is that this vessel, which was calculated to hypothetically withstand any psi, failed at 67 psi. At 67 psi, the patch they put on to close that thing flew off, causing the vessel to eject its contents, a superheated solution of polyethylene wax that caught on fire. That all sounds pretty bad enough by itself, but because of this process having the oxygen level that it did, the fireball was going to travel back into the process upstream and explode the tank, causing it to land where it did. So in this one, they “got ’er done.”
NDK Crystal, Inc. was a company outside of Chicago that made synthetic quartz. They had eight very large tanks about five stories tall that operated at 30,000 psi. The company they contracted to make these tanks made three in the first go. They came to find out the material they selected was not significant or strong enough to meet the minimum grade specifications from the codes. Did they scrap that design, return to the drawing board, and create three new vessels? No, they did not. They went to the state and asked for a special dispensation, which they received. Another issue that came up in the design of these vessels was that different parts were thicker than what's generally recognized as appropriate. So, there were some questions about how effective the heat treatment of these vessels was in general.
Regardless of the special dispensation of the first three vessels, they changed the material for the other five. They operated for a couple of years and got some runs out of them. A run of these vessels takes about 150 days to get the quality of the material that they need.
In 2007, two years before the incident, one of the tanks had the lid fly off. Nobody was injured, but it was a pretty big mess as you can imagine. The company shut down everything and hired an engineer to do an evaluation of the facility, look at the incident, figure out what the root cause was, and what could be done to prevent it from happening again. It all sounds very responsible after this point, doesn't it?
That person concluded that the incident was the result of stress-corrosion cracking because the vessel itself was made of carbon steel and was in-process service. The company's means to insulate the vessel from the corrosive atmosphere were ineffective and not even done that way on purpose. The engineer who performed the investigation recommended the company take down all the vessels to evaluate stress-corrosion cracking and the efficacy of the material they had put in to prevent the corrosion and not make any more production vessels until this was completed.
Do we think that happened? It did not. They ran for about two more years. What's interesting is that the vessel that had the lid blown off was one of the vessels that was made later, using the appropriate material to meet the code. And the vessel that failed catastrophically was one of the first three vessels which did not have the strength requirements to meet the code.
If we have to name this one the same as we've named the others, we might call this one “special dispensation.” Can you imagine going up to the code enforcement authority and saying, “We actually don't meet the minimum requirements of the law that you guys have enacted in this state and that you enforce every day of the week on everybody else, but we would like it if you would just let us keep these.”
I don't want to get too deep into it; obviously, I wasn't there. This incident was well before my time, but I think it's pretty clear in hindsight that that dispensation should not have been offered. Tragically enough, the person who was killed in this incident was across the street getting gas – not even an employee of the facility, just some poor bystander who got hit by a 700-pound piece of chunk steel.
This one is a more recent event, and that's why we've been talking about it a lot lately. In 2017, a massive steam explosion caused a steel pressure vessel weighing 2,000 pounds to launch 130 feet in the air, fly across the street, and kill three people.
A lot of the backstory involved in this one involves faulty operating practices, inadequate repair in 2012, and a company that just refused to learn lessons. This company operates Monday through Friday and shuts down over the weekend. One Friday morning, the stationary engineer responsible for operating this process notices that the pressure vessel is leaking. There's an observable puddle of water beneath it, but he's seen it happen so many times previously that he doesn't consider it an emergency.
They ran all of Friday before shutting down as usual for the weekend. They hired a repair crew to come in over the weekend to put a flush patch on the bottom of this vessel. On Monday, they went to start it up again as usual. During that start-up process, the flush patch that was applied in 2012 fell off.
The post-incident engineering analysis that we commissioned indicated that at the time the repair was made, the average thickness of the remaining material that was welded onto was, on average, half the thickness that it should have been to maintain that. Digging into it a little bit deeper, which was difficult because the repair company had gone out of business prior to the incident, we found out that it was their intention originally that it should be more of a temporary repair. A bid was submitted to do a more thorough and complete repair that was never followed up on by the company or the repair organization. It held for five years. I think that's pretty impressive. If they had done it the right way, it probably would have held indefinitely.
Finally, most recently, the Yenkin-Majestic Resin and Paint Corporation in Columbus, Ohio, where NBBI is headquartered. On April 8, 2021, they had an explosion and fire. You're not going to believe why this happened. It happened because, after a manufacturing batch of resin, the operator came in and added a flammable solvent as they always do. The problem was, though, that the agitator wasn't on. What's supposed to happen is it's supposed to mix continuously and not explode. Unfortunately, unbeknownst to this operator, an electrical shortage caused the agitator not to be running.
What ended up happening was the hot resin, like a layer of flammable liquid vapor built up on top of this hot resin. Then, the person responsible for the charger returns from the laboratory after dropping off the sample and observes through the window that the agitator is not on. It seemed simple enough: This thing is supposed to be on, and it’s not, so let me turn it on. He turned it on, and all that flammable solvent on top of that hot resin vaporized nearly instantaneously. The vessel went from 0 psi to 4 psi in under two seconds.
Again, we've properly protected; we have a rupture disk and a pressure relief valve. We have a vessel built and constructed to withstand certain pressures. The vessel that failed at 9 psi was stamped at 40 MWAP when it was manufactured in the '80s, which should be no problem.
Unfortunately, in the prior year, the company, as part of an automation process, needed a new manway to fulfill its goals for automation. It hired a company to come in and install that manway. Again, that company was not qualified to do the work it was hired to do.
The Yenkin-Majestic Resin Plant thought it could hire these people to do whatever it wanted to because of a convenience belief that they were also not covered by the state's boiler and pressure vessel law. They considered this an atmospheric vessel because, technically, it has a process vent that is continuously opened to the atmosphere.
Unfortunately, the nature of their process, though, is to create the vapor at such a rate that, in general, almost all the time, it's at 2 psi. And they say, well, it's an atmospheric vessel. Yeah, but it runs at 2 psi. To me, atmospheric means 0.5 basically or below. How do you run this thing at 2 psi and call it atmospheric? Because it is? But it's not.
Yeah, it is. We have an alarm that goes off at 4. Again, they convinced themselves they weren't covered by this rule for reasons that I'll never understand. Eventually, it exploded. The vessel itself didn't come apart; all that happened was that the gasket they put on the new manway failed, and it ejected all this form of liquid vapor and filled up the entire plant. If you have time, look up this incident online. They have surveillance footage of the vapor cloud chasing the operators out of the building. It's terrifying. It's a miracle that only one person was killed.
I like to think of this one as a “belief of convenience” combined with a “get ’er done.” Interestingly, a month after they put the seal on the first time, it failed. Instead of taking that as a sign that it was and going back and reading up and doing an analysis of what should be done to achieve what they wanted, they bought a thicker gasket and tightened it down a little tighter than they did the first time. It held for quite a bit longer and got them where they needed to be for a couple more months, but now, unfortunately, this plant is a parking lot.
They're going to rebuild it. We've made recommendations for when they rebuild so this doesn’t happen again. Still, the sort of company that can't follow the logic of operating over 15 psi on the air pressure vessel, and if you operate over zero psi, you're not atmospheric. I'm sorry to say that I'm hard-pressed to believe they will benefit substantially from these recommendations.
A lot of these incidents have really good recommendations. D.D. Williamson & Co. received three solid recommendations for improved processes, monitoring equipment, alarms, and pressure relief valves, and they did it. The jurisdiction of Kentucky got a recommendation to spread the word about unregistered, out-of-state pressure vessels coming in and not being inspected. NBBI got a recommendation to share the lessons learned from that incident with their membership because I'm sure before that recommendation came out, people were aware that vessels in their jurisdiction weren't being properly inspected, supervised, and registered.
But Loy Lange never hurt anybody, you know? Then we go back to the Marcus Oil investigation, and it's the same thing. The company received five really good recommendations, and they implemented all of them. They're still in business, and we haven't investigated them again since.
Once the CSB investigates a company, the likelihood that it will be investigated again is ridiculous. There's a company called Formosa that we investigated four different times over three years for four separate, distinct, individual catastrophes that happened in four separate, distinct, individual locations. Our recommendation should have been for this company to go out of business.
With NDK Crystal, we went so far as to ask the jurisdiction not to allow them to begin operating again until it could prove these vessels would pass fitness for service. The jurisdiction agreed. The business decided to close, and in 2015, it was demolished.
With Yenkin-Majestic, these are too new. We haven't even heard back yet from anybody about these recommendations, but we've asked ASME and NBBI to come up with fitness-for-service guidance for vessels that routinely operate under 15 psi. We've asked the company to implement a lot of safeguards and a lot of procedures and programs.
If you ever need anything from the CSB, please call me. If I can't be the one to help you, I can put you in touch with the person who can help you. I work in the Recommendations Department. We have an Investigations Department, and from time to time, they reach out to me, NBBI, ASME, and API to ask technical questions. We’re not looking to drag anybody down or to hang blame on anybody. We want to get the stories of these catastrophes out so people who weren't affected by them directly can use the lessons learned from them and prevent having their own disasters.
We have a couple more minutes. Does anybody have any questions about the Chemical Safety Board or any of these incidents or anything else?
MEMBER: How many inspectors do you employ that gives you the ability to send out investigations?
MR. HENSON: We have about 20. We're striving for 35. It's very hard to get people into these positions because of the type of background that they look for. Currently, we have three teams deployed. One in Martinez, California, for a reboiler tube that failed and killed a person. Another is in Texas, where a hydrofluoric acid incident happened a couple of months ago. By coincidence and happenstance, nobody was injured, but there was potential for it. And we really want to try to get ahead of that thing.
MEMBER: Did the city of St. Louis follow up on the recommendations for Loy Lange?
MR. HENSON: They did not. The city of St. Louis has never formally responded to our recommendations. We've had several informal meetings, which are held with the actual board, and that doesn't happen often. The actual president-appointed Chemical Safety Board is trying to get a meeting with the mayor of St. Louis, and the mayor of St. Louis won't meet with them. Their recommendations were to create a city law, an ordinance that follows NBIC Part 2, or similar, to train the people who do these inspections for the city as Authorized Inspectors. The one with the mayor's office was simply to send a notification to the city's licensed operators that this incident happened and what they ought to be doing to prevent it. And we can't get anything out of the city of St. Louis.
MEMBER: In Kentucky, where you have this problem, you say, well, this vessel was installed and operated not in accordance with their vessel law, but yet another jurisdiction doesn't even have a vessel law. So does the CSB make any recommendations to the jurisdictions to say, hey, you should have a vessel law?
MR. HENSON: We have done it on two occasions, and it's not worked. With Marcus Oil, the city of Houston said no thank you. There was another case where they wouldn't even call us back. The reason we found out they weren't going to do what we asked them to do is because they did an interview with a national public outlet talking about how they weren't going to do it because they didn't want to have to investigate every air-conditioner that was installed in the city, which wasn't our intention.
It's very interesting you bring that up. There was an incident in West Virginia with a rotary drum dryer. They were trying to dry out this water-purifying pool chemical. They overheated and over-contained it, and it exploded. And West Virginia is a state that doesn't have a pressure vessel law. Afterward, we went to the state and asked them who's responsible for this vessel? Because in the fire code it says rotary drum dryers shall comply with Section VIII of the ASME Boiler and Pressure Vessel Code.
The person in the Department of Labor who does boilers said, we do the boilers. I said that's all well and fine, but here in the fire code, it says that there are pressure vessel requirements that the state needs to enforce. Who enforces those? The fire marshal. We go to the fire marshal, who said it’s the boiler guy. No, we spoke with the boiler guy, and he said it's you – and it's in the fire code. So, logic, common sense, and the way things usually go, I asked if anyone on the staff is qualified to do pressure vessels. He said: The boiler guy does it. We got nowhere with them.
Every opportunity we get, we make recommendations to these jurisdictions that they adopt laws, especially those created by the committees and groups of people in this room because we believe in them. Many people who work at the CSB are engineers, and they have a ton of respect for both groups.
St. Louis is another one whose jurisdiction is loose as a goose. Missouri has a program. They didn't want to be under that program, so they made their own. But the vessel at Loy Lange that exploded wasn't even registered with the city. As part of their ordinance, they were supposed to go and look at the steam generators yearly. How many times did someone come to those steam generators in 20 years? Three. They never noticed the 18-foot-tall, 4-foot-round vessel sitting right next to them, which was not registered with their jurisdiction.
MEMBER: What are the criteria of the CSB to investigate an incident?
MR. HENSON: If there's an off-site fatality, we have to investigate that. Everything else is more or less up to our discretion and resource availability. An example is Loy Lange. If you read the statutory authority of the Chemical Safety Board, it says it has to be a hazardous chemical. What were they doing at Loy Lange? They were making steam. It's water. Obviously, we learned that water can be dangerous. A more recent incident in Wisconsin was a cornmeal plant that blew out – a combustible dust incident. That was corn. You don't typically consider corn to be a hazardous substance, but in this case, it killed five people.
MEMBER: Is the CSB strictly a reactive agency that reacts to accidents, or are you guys out proactively doing work to try to prevent them?
MR. HENSON: Unfortunately, we are a reactive agency. We try to come up with ideas to be proactive from time to time.
MEMBER: This is proactive.
MR. HENSON: No, it's based on something that already happened. Everything I've been talking about has already happened. I'm trying to prevent it from happening again. Everything we do in one way, shape, or form is a reaction to something that's already occurred.
MEMBER: Do you have regional offices?
MR. HENSON: No, we have one office in Washington, D.C., but all the professional employees are remote. I work in New Hampshire.