Opinion: Dust Explosions, what the incident records tell us
20 March 2012
This is intended as a follow up to the article Dust in the Factory - Still an Exposive Topic by Gordon Low on the HazardEx website and in the March edition of the journal.
A CSB reconstruction of an iron dust explosion at the Hoeganaes plant
A steady flow of dust explosion incidents in the USA, reported and then investigated to a high standard by the American government Chemical Safety Board, reminds us of the wide range of industries in which these risks may be found. The most recent report concerns Hoeganaes Industries, who make fine iron powder. A single site had three explosions inside six months, killing five persons. Sadly, the company is British owned.
To tackle these issues effectively, it is important to consider the accident statistics, and look for underlying features. The most important of these, I suggest, is that dust explosions nearly always start inside the process. This is because dust clouds of the required concentration form regularly inside the equipment, but with properly designed plant dense dust clouds rarely occur outside, or are very limited in extent.
That has implications for the pattern of ignition sources. I’d like to consider four of these in particular.
Self heating products are a significant issue, and examples are powdered milk, dried sewage sludge, biomass fuels, and some fine chemicals. All have caused dust explosions. We need to identify these at the design stage, characterise them, and design the process around the properties, but you cannot eliminate the ignition risk completely.
Drying and heating processes of all types figure very prominently in the accident statistics. Here you are deliberately applying heat to the materials in the process, and safety often depends on controlling both surface temperatures and contact time. Once again, you cannot design out the risk completely, nor can you exclude it by selecting ATEX equipment with a low temperature rating. You may try and control the formation of dust clouds, but even that as an approach is limited; drying is quickest when the product is suspended in a hot airstream. Often the external surface of the drier or heating plant is hot enough to ignite a dust layer or cloud. Some incidents where dust is suddenly released occur in this way.
Jeremy Smallwood identified in his article in the March 2012 edition of HazardEx the risks from static, which in the case of dust, is mainly concerned with preventing spark discharges, by ensuring suitable earthing paths. He did not have space to mention the one situation that really should be avoided, that of pouring powders into a vessel holding highly flammable liquids with air above the liquid level. This activity combines the risks from a vapour sensitive to ignition with the charging potential of moving powder. Inerting, or avoiding the act of pouring powder, is necessary to avoid an ignition.
Mechanical equipment features regularly in the causes of ignitions, and progress has been made in trying to specify how this risk can be reduced, through the standards in the BS EN 13463 series. Very frequently the mechanical moving parts are in contact with the dust cloud, or actually create it. Undetected faults are then a real risk.
There remains however, a large range of mechanical equipment where the options for meeting ATEX requirements are limited. Mills and grinding equipment are often high- powered, high-speed plant, creating their own internal dust clouds. Most of these simply cannot meet standards higher than ATEX category 3D, and this is officially recognised. Some mechanical equipment blurs the line of what is the inside, and what the outside. The Imperial Sugar explosion illustrates this (www.csb.gov) where a simple modification to a belt conveyor intended to limit sugar dust releases probably contributed to creating an explosive dust cloud inside the plant.
As a broad brush generalisation then, if you really want to minimise the risk of a primary dust explosion, look first at the inside of your process. Often the conclusion will be that the primary event cannot be excluded, and the plant must be designed to control the consequences.
Sadly, no matter how much time you spend worrying about area classification around the building and selection of ATEX equipment to suit, you will make little influence on the explosion risk. You may reduce the risk of fire, but that is another story.
Contact Details and Archive...