Preventing Dust Explosions
CSB's investigation raised awareness of these dangerous incidents. Detection/suppression systems work fast to snuff them out.
- By OH&S Staff
- Nov 13, 2007
Editor's note: The U.S. Chemical Safety and Hazard Investigations Board approved a final report last November after a two-year investigation of industrial dust explosions. CSB counted 281 combustible dust fires or explosions between 1980 and 2005 and, finding no general industry OSHA standard in place for them, recommended that OSHA issue one. CSB's database is available here.
Randy Davis, director of Industrial Explosion Protection for Boston-based Fenwal Protection Systems, discussed the causes of industrial dust explosions and the systems that detect and suppress them in an Aug. 17 conversation with Occupational Health & Safety's editor. Today's systems suppress explosions within 100 to 200 milliseconds after the start of ignition, and they're about to be held to a higher standard--the 2008 edition of NFPA 69, Standard on Explosion Prevention Systems, which is a complete rewrite intended to make explosion prevention systems more reliable, with third-party approvals added for the first time.
Excerpts from the conversation follow.
Looking at the CSB's database of dust explosions, they involve many substances and many causes. Is there anything common to dust explosions--any substance or any mistake?
Randy Davis: There is no single common substance or mistake involved in dust explosions. Dust explosions occur in a wide variety of industries and applications. One common thing that is a major part of the secondary explosions is housekeeping. Typically, there's a process vessel inside a facility that is processing a combustible material. If you have an explosion in an unprotected vessel (the primary explosion), the vessel is destroyed or ruptured and a fireball is released into the larger process area. When there is improper housecleaning, that's where you get the secondary explosion. This is when you have major damage and the risk to life.
The database indicates a lot of them started in dust collection systems. Dust collectors seem to be where the initial spark was, and it then propagated out into some larger area.
Davis: Correct. Dust collectors are definitely the single-biggest-hazard piece of process equipment, although we do see explosions in dryers and silos and other pieces. But if you said just one piece of equipment, it would be a dust collector.
Is that because those are installed and perhaps forgotten--not cleaned or maintained as they should be?
Davis: Not necessarily. Because dust collectors' function is to pick up dust, you have the first ingredient--which is a combustible atmosphere--because you have dust in air. And just because of where they pick up: Some dust collectors will pick up their dust from a mill, so if any kind of foreign material gets into it, a piece of metal--that's going to provide your ignition source. Or they can pick up a burning ember from a dryer or any other process, and then typically you get the right conditions in the dust collector to start the explosion.
I see what you mean: The way they normally operate can allow this to happen. The systems your company provides keep this initial problem from propagating into a much bigger one, correct?
Davis: That's correct. The way the systems work is, after an explosion starts, they'll detect it and suppress it, or vent it, which is releasing the energy before it ruptures the vessel and allows it to escape into the larger process area.
Detection, depending on the size of the vessel--because we protect vessels from 10 cubic feet up to 40,000 cubic feet--is typically in the range of 2 to 20 milliseconds. That's mostly a function of the size of the vessel.
Do all of these systems use pressure detection sensors?
Davis: That's a primary means [of detection]. We also do some room protection for aerosol filling, where they'll fill hairspray cans and paint cans with combustible propellants. For those applications we'll use a combination UV/IR optical flame detector.
For dust environments, which are the majority of systems, there are two different types of pressure detection: There's a threshold pressure detector, which is the older technology that's just looking for a certain pressure rise within the vessel. The new detection uses rate of pressure rise technology, and it's actually monitoring the pressure rise and comparing to the parameters for whatever that material being processed is. Different materials burn and generate pressure at different rates.
The idea is that pressure fluctuations are inherent in most processes, and you want to be able to distinguish between a process fluctuation and a pressure rise due to an explosion.
Once you analyze the process, you can set the system so pressure fluctuations are taken into account?
Davis: That's correct.
How expensive are such systems?
Davis: It varies…. Cost is going to be dependent upon the size of the vessel. For a typical dust collector, anywhere from $10,000 to $25,000 is probably a good ballpark figure.
Are you trying to save what's inside the vessel or not?
Davis: Not typically. The focus is to protect that vessel from being damaged and then also prevent [the energy] from propagating out of that vessel, either through interconnecting ductwork or through rupture of the vessel.
Once the system has worked and prevented an explosion, what's the next step? Do you go in and do thorough housecleaning? How do you eliminate the hazard and keep going?
Davis: It's a two-pronged attack. The first thing you want to do is find out what caused the event so it doesn't occur again. The last thing you want to do is put the system back online and have an incident happen immediately again.
This involves a thorough investigation to find out what caused the ignition source. Again, the fuel's going to be present just in normal operation, but you must find out what caused the explosion to start and then eliminate the source of the ignition.
The second step is to clean the process equipment and process area and recondition the protection system prior to restarting the process.
Do these systems send an immediate alert to the operators so they can take that action?
Davis: Right. The process is interlocked; whatever process controls are working would be interlocked with the control unit for the protection system. It does two things. One, it notifies the operators that the system discharged and they've had an incident. Two, and probably just as important, it shuts down the process so that it doesn't have a restrike immediately.
These systems are commonly used in this country. Are they commonly used all around the world?
Davis: Yes. We have sister companies in Europe . . . . As a matter of fact, the regulations are much more stringent in Europe. For example, in England, you have a government agency that can go into a processing facility and do a review of the dust explosion hazards. And if they don't have proper protection for dust explosions, they can shut down the process and prevent them from manufacturing until they get adequate dust explosion protection.
And they sometimes do that?
Davis: Yes, they do. I can't say they often shut down facilities, but they do have that threat. So it's much more ingrained into the manufacturing society that they have to address the hazards.
Other countries? Is it also stringent elsewhere?
Davis: Throughout Europe, it is. Once you get away from Europe the regulations are much more lax. In other parts of the world, it is typically only the large, multinational companies that utilize explosion protection.
Let's talk about housekeeping again. You mentioned it is a common element. Is it fair to say it is not generally ignored but is not as well practiced in some sectors as it should be?
Davis: Yes. As I said, dust explosion hazards are present in many different industries. In the pharmaceutical industry, typically, housecleaning is more enforced just by the nature of the business. Some others, such as the wood industry and others--I won't say they're bad industries, but enforcement on housecleaning probably isn't as strict as it is in pharmaceuticals.
I thought sawdust was often stored outside anyway.
Davis: We're talking about a particle size that is usually under 420 microns that is explosive. Sawdust generally is larger than that. In the wood industry, it's more the sanding dust, so it's a finer dust that would not be stored outdoors.
Is the composition of the dust a factor in which system you decide to use?
Davis: Not only in what type of system, but also the size of the system, the number of extinguishers, etc. Systems are designed based on a computer model, and one of the key inputs is the combustibility of the dust.
Can your company determine combustibility? I assume the customers generally know that.
Davis: Yes. We have a lab associated with us, the Combustion Research Center (CRC), that can determine the combustibility properties for a particular material. There are other labs around the country that'll provide similar service, as well.
Does the CRC conduct research on an ongoing basis? And do you have past studies to share with customers?
Davis: Yes, the CRC conducts fire and explosion testing on a full-time basis. Now, you'd think all wood dust is the same, but it does vary. Oak has different combustible characteristics than pine. We have a database of previously tested materials, but this data is customer proprietary. NFPA 68 requires that the combustibility parameters of material should be tested per ASTM standards and recommends that the testing should be conducted on dust that has been created by the process to be protected. This is because different processes will produce a material with varying characteristics such as shape or moisture content that have a direct effect on the combustibility characteristics.
Are there always new substances being worked on and therefore new types of dust being created?
Davis: Definitely. There are new chemical compounds that are being developed. Even some standard compounds, such as coal, have different combustibility characteristics if they are mined from different areas of the country.
So that's why you need a lab that does this work full time. Is your business growing? Are these systems being used more frequently and getting more sophisticated?
Davis: Yes, the market continues to grow. You mentioned the CSB site; that has generated more interest. Their recommendation was for OSHA to write regulations around dust explosions, and we expect that as dust hazards get more visibility in all industries, we'll see it continue to increase.
A key area of focus is education, to educate customers that dust explosions are a hazard. We've seen customers run the same process for 20 years and never have an explosion, and a week later, they blow up a dust collector or a dryer. Conditions were just right, and an incident happened.
A lot of people are not aware that they have a hazard until they start analyzing their process or someone else in the industry has an explosion. Then everybody steps back and says, "Wait a minute, we have that same process here. We have that risk."
This article originally appeared in the November 2007 issue of Occupational Health & Safety.