If you work in a location that handles or processes hazardous materials, eventually you have to come to grips with the matter of risk and accidents. It is possible to design procedures that prevent certain kinds of accidents and casualties.  It is possible to install devices and automated contrivances that can eliminate specific failures and the resulting cascade of multistage calamities that might follow.

Over time and with plenty of thought, a chemical plant can be fool-proofed to a large extent. But in the end, residual margins of safety depend on the man-machine interface. People have to undergo recurrent training and certain staff must be assigned to specialize in safety.

All devices have a failure rate. The rate may be small or large. A device may fail safely or not. Most chemical plants are, to a large extent, hand built. They are fabricated by skilled tradesmen who connect pre-fabricated parts to one-of-a-kind assemblies built on-site. In this way, expertise is captured from the plant designers, contractors, and the manufacturers of the installed equipment.  In the end though, it is up to the designers to assure that there is compatibility and some margin of overdesign in the finished facility.

While it is possible to assemble a facility from the very best equipment, the fascinating question of design for the attenuation of accident propagation is not often discussed, at least openly.  Accidents can usually be reduced to a few characteristic phases. They are initiation, propagation, and termination. 

An incident begins with an initiating event. Some release of hazardous energy is presented to the surroundings that may begin a propagation of undesired events. Events can propagate in series or parallel chains.  Hazardous energy can be electrical, chemical, or mechanical. The initiation is the tipping of a domino as a triggering event that causes the release of other hazardous conditions to ensue. Eventually, the propagation of the hazardous energy release is suppressed, extinguished, or simply exhausted in the termination phase.

One of the best ways of learning about the phenomenon of accidents is reading about them. A website worth visiting is the US Chemical and Hazard Investigation Board. It is useful for chemists and engineers to study the anlyses of the CSB and gain useful insight into the dynamics of chemical plant accidents.

It is possible to configure a chemical plant in such a manner as to attenuate the propagation of hazardous energy during an incident. In general, a large distance between reservoirs of potential energy is the easiest solution. Explosives manufacturers have known this for a long time. One well known German manufacturer of energetic materials has a manufacturing site spread over a large rural area and has built in bunkers with berms and trees to attenuate the propagation of shockwaves and allow flying fragments to land safely in an uninhabited area.  Fortunately, not many manufacturers have processes and products requiring this kind of design consideration.

Situations of “ordinary” risk magnitude do require some thought, however. Consider the storage of drums of flammable materials. Most companies that handle palletized drums of flammable liquids meet the minimal fire and insurance codes for the handling of these materials. 

But consider this. What if a forklift driver spears a drum of solvent with his lift, and then in a panic, backs up and pulls the fork out of the drum resulting in a spill? At this point, policy and regulations are irrelevant. The only question is this-  Where does the liquid and the potential fire go?

Indoor storage of flammable materials requires fire suppression. Fire suppression is not the same as fire extinguishment. It is about knocking down the fire to a manageable level for emergency egress, to suppress the spread of the fire, and for firefighters to make some kind of attempt to extinguish the blaze. This is routine firefighter stuff.

What is less than routine, however, is the issue of BLEVE’s. I have written on this phenomenon previously.  Fire suppression is one thing, but BLEVE’s – Boiling Liquid Expanding Vapor Explosions- are quite another matter to deal with.

This is where a well designed facility with passive architectural features to attenuate the spread of hazardous energy can be helpful. An indoor BLEVE is virtually assured to accelerate the pace of a disaster.  So, in the planning phase of a plant, it is important to consider how energy release during an accident may propagate.  Drummed flammable liquids should be isolated from work areas and egress paths. This is pretty obvious to initial designers, but not necessarily years down the road during an expansion.

Consideration should be given to the anticipated direction in which energy is released. Where possible, energy should be released away from populated areas and away from major capital equipment. A fire in a materials storage area shouldn’t lead to an extended plant shutdown due to damaged process equipment. Segregation is key to plant safety and business viability.

Smoke is a potential killer and there are architectural tricks that can add provide slightly greater safety margins. Ceilings designed to collect and channel smoke out of the space could reduce the likelihood of suffocation of stranded workers and suppress the chances of a flashover. Smoke curtains properly placed can channel smoke away from hallways and the resulting spread.

Another concern is the fate of spilled flammable liquids in a storage area. Where should the spill go? Should the spill be concentrated in a small space or channeled to another space where a fire can burn with lower negative consequence? Nobody likes to pay for an overengineered warehouse, but fire resistant partitions in a solvent storage area can go a long way toward the isolation of a fire and attenuation of a larger scale calamity.

One major plant accident I am familiar with has a number of attributes that other operators would do well to consider.  A 750 gallon reactor explosion resulted in the complete fragmentation of the vessel.  A few pieces ejected from the hole in the roof were found lodged in the walls of neighboring structures off-site.  Fortunately, this reactor was in an enclosed space with no other reactors or stored hazardous materials. In one way, this accident was isolated due to passive attributes. However, the building space was interconnected to other spaces by a series of adjacent rooms and hallways. While fragmentation and fire damage were contained due to the happy fortune of isolation, the shockwave was able to follow all of the connected and enclosed pathways.  The connected pathways were a convenience to the workers, but this feature channeled a pressure wave throughout the entire facility, lifting the roof enough to damage large -remote- sections of it as well as badly damaging overhead doors and windows throughout the facility.

Take home lessons? 1) Leave open space walkways between production and storage buildings and the rest of the facility. Collateral damage is likely to be suppressed with this cheap, passive feature. 2) isolate and dedicate certain vessels to hazardous operations. 3) Store hazardous materials well away from processing areas. Storage and processing have their own hazards and a disaster in one area should not be allowed to propagate to the other.

The matter of flammable solvent storage and accident attenuation is only partially solved with enclosed flammable materials lockers. It seems to me that some research should be done to advance the level of best practices in this area.