We all are aware of…
68.25(h) Consideration of passive mitigation. Passive mitigation systems may be considered for the analysis of worst case provided that the mitigation system is capable of withstanding the release event triggering the scenario and would still function as intended.
So I wish to challenge the claim that a WCS inside an engine room would be allowed to consider the engine room structure as an “enclosure”. It is very common to find WCS OCA’s in the refrigeration industry to lay claim that the WCS will be reduced by the engine room structure, but as demonstrated in this article the engine room would need to be built to some other standard than just local building codes. The pressures the WCS generates within an engine room can be WELL BEYOND what a normally constructed cinder block wall can withstand. Here is a scenario that is somewhat typically of a medium sized refrigeration process housed inside an industrial built engine room where the walls are cinder block and/or the engine room has windows or roll-up bay doors:
Using the “enclosure” as a mitigating device a facility can reduce their RMP*Comp WCS endpoint distance with a release of 15,000 pounds from 1.5 miles to 1.1 miles (or from 2.2 miles to 1.6 miles if rural topography). What we have upon release of the 15,000 lb over a 10 min period of time is a simple volume expansion from the liquid state to the gaseous state uniformly increasing the pressure inside the room.
Liquefied anhydrous ammonia at room temperature has a liquid density of 38.00 lb/ft3 at 70 degrees F, so 15,000 lb would take up about 394ft3 of space.
If 15,000 lb was released into the room, the very best you can do is assume only about 20% of it flashes to vapor and the rest forms a boiling liquid pool inside the room (even though most engine rooms do NOT have secondary containment). The vapor would be very cold, of course. Again, the best you could do would be to assume the vapor temperature warms up from the -28 degrees F boiling point of ammonia to say 0 degrees F. The vapor density at that temperature is 0.1042 lb/ft3 at one-atmosphere pressure.
The 20% that flashed from liquid to vapor would take up (15,000 lb * 0.2)/(0.1042 lb/ft3) = 28,800 ft3 in OPEN AIR. (In reality, the liquid will aerosolize and boil and the vapors will be warmer, so the volume would actually be GREATER!!!)
However, those vapors aren’t in the open air, so the only thing they can do is to pressurize the room until something gives way.
Assuming the engine room is 5000 ft3 (length x width x height), the 28,800 ft3 of ammonia vapors are going to be compressed into that 5,000 ft3 room volume, and assuming no significant leakage from the room and neglecting the initial liquid volume (i.e., tank volume), will raise the room pressure by:
14.7 * 28800 / 5000 = 85 psi
Since typical cinder block walls start giving out at roughly 3 to 5 psi depending on how they were built, windows go at about 0.5 psi, and large metal roll-up bay doors typically fail at something less than 1.25 psi we can begin to see that a typically built engine room will NOT stand a chance during the WCS of an ammonia pressure vessel cracking open like an egg and taking the liquid to a gas in 10 minutes!!
(Note: The above sample calculations only address the volume expansion hazard. Obviously, anyone inside the room would be fatally injured by the toxic vapors, cryo temperatures and/or pressure effects, unless they were able to immediately escape. The calculations also do not address what would happen if the vapors ignited, like in the Borden explosion in Houston, or if the bullet tank ruptures.)
Here is what EPA’s OCA Guidance says:
3.1.2 Releases of Toxic Gas in Enclosed Space
If a gas is released in an enclosure such as a building or shed, the release rate to the outside air may be lessened considerably. The dynamics of this type of release are complex; however, you may use the simplified method presented here to estimate an approximate release rate to the outside air from a release in an enclosed space. The mitigation factor (i.e., 55 percent) presented in this method assumes that the release occurs in a fully enclosed, non-airtight space that is directly adjacent to the outside air. If you are modeling a release in an interior room that is enclosed within a building, a smaller factor (i.e., more mitigation) may be appropriate. On the other hand, a larger factor (i.e., less mitigation) should be used for a space that has doors or windows that could be open during a release. If any of these special circumstances apply to your site, you may want to consider performing site-specific modeling to determine the appropriate amount of passive mitigation. In addition, you should not incorporate the passive mitigation effect of building enclosures into your modeling if you have reason to believe the enclosure would not withstand the force of the release or if the chemical is handled outside the building (e.g., moved from one building to another building).
So consider this when doing your 2014 submittal… will my engine room be able to withstand my WCS? If the answer is “no” then we should not be “clicking on the box” below
