4/20/2018 UPDATE: To further this discussion… consider a diffuser tank is 95% effective (Source: IIAR Technical Paper), meaning that 5% of NH3 would theoretically make it to the atm. WHEN that 5% is greater than 100 pounds, you have a REPORTABLE RELEASE. Using an actual “oil separator” at a client, the RV has a 0.75″ discharge with a 19 pounds/minute rate @ 300 psi (Parker RV) would equal 1,140 pounds going to the diffuser tank over a 1-hr period. And with the assumed diffuser tank having a 95% effectiveness means that 57 pounds would make it to atm, falling short of the 100 pound RQ. It appears, using the 95% efficiency assumption on a diffuser tank, we would have needed a scenario of a considerable RV lifting, OR the RV is at full throat for longer than 1-hour, or we have a situation where multiple RV’s lifted, and their flow rates exceeded that as stated above, THEN even with a Diffuser Tank, we could have a reportable release.
One of my go-to guys for NH3 Refrigeration questions/discussions, Brian Chapin, posed a question he was debating with a group… “If my NH3 refrigeration process has a diffuser tank, if an RV lifts and sends NH3 to the diffuser tank, is this considered a “release”?” My discussion with another friend who must remain nameless was indeed interesting. But the short answer is NO, as long as…
the Treatment/Destruction system works as designed AND it is not overloaded.
First, we have to understand the limitation of a “water diffusion system” designed using IIAR-2, 2014. Section 15.5.3 states the system must be sized to handle the amount of NH3 that would be released in 1 hour from the largest relief device.
This is very easy to design to… simply take the largest RV flow (usually provided in cubic feet per minute) and establish the pounds of NH3 that could pass through this RV in 1-hour and then ensure we have 1-gallon of water in the tank. Of course, we have to ensure the water does NOT freeze. And the code requires the discharge pipe from the pressure relief device to distribute the ammonia in the bottom of the tank but no lower than 33 ft (10 m) below the maximum liquid level. Plus, the tank must be large enough to contain the volume of water AND ammonia WITHOUT overflowing.
So with all that said, a single RV lifting for a time LESS THAN 1-hour should NOT result in a “loss of primary containment” and thus this would NOT be a “reportable release” as our “water diffusion system” should have easily contained 100% of the NH3 vapor that bubbled up through the tank.
However, should any failure in our “water diffusion system” design and/or operation and we could indeed have a reportable release even with a “water diffusion system” in place. Improper design (vent pipe not long enough/too long, tank too small), not enough water, “dirty water”, frozen water (ice), etc. could cause us to merely pass the NH3 through the tank and into the atmosphere and if the amount of NH3 that exits the process is 100 pounds or more then we have a “reportable release”.
But even with a “water diffusion system” designed perfectly to comply with the current RAGAGEP designs, the system has some inherent design limitations. Most notable is that the system is designed for a single (largest) RV to lift and deliver NH3 to the diffuser tank for no longer than 1-hour. And by the way, this water quantity design seems to be 99.9% effective, but it is NOT 100% effective as it was long ago when the “water diffusion system” had to hold 1-gallon of water for every pound of NH3 in the process.
(NOTE: in the very early days of “water diffusion systems” the UFC required 2-gallons of water for every pound of NH3!!!)
We just have to look at the “fire scenario” in the mechanical room and see the proximity of the equipment is such that a single fire in the mechanical room could, in fact, lift several RVs (including the largest RV in the room) and this scenario would overload the “water diffusion system” and if the system were overload such that 100 pounds of NH3 was discharged from the diffuser tank vent we would indeed have a “reportable release” of NH3.
But as I said earlier, a properly designed, operated, and maintained “water diffusion system” will contain our largest RV discharge for up to 1-hour and thus the vast majority of NH3 that discharges thru an RV will NOT become a “reportable release.”
But in the past year, we have found properly designed “water diffusion systems” NOT operated/maintained properly during our work.
- One tank was empty – it was emptied for some tank work during a shutdown, and it was not refilled before the process was returned to operation.
- NOTE: I am working on another article about what it takes to safely take a diffuser tank out of service when the process still contains NH3! It is not an easy task and when done improperly creates a VERY DANGEROUS entry into the tank!!!
- One tank we came across could only be called an aqua ammonia storage tank! An RV, or maybe more than one, had been “leaking by” or quite possible chattering since the process was found to have been operated very close to the RV set points. Either way, the tank was no longer holding water, but a dilute concentration of aqua ammonia; meaning that it may have had 1-gallon of a solution for every pound of NH3 from the largest RV, but that solution was NOT going to absorb all the NH3 since it was ALREADY an NH3 solution.
In our PHA’s and Engineering Design Reviews on “water diffusion systems” we always recommend “low level” detection and alarms, Operating Limits w/ SOPs, and that the system is included in the PSSR process. We put a section in the “start-up” procedures to ensure someone is doing a check on the “water diffusion system” before the process returns to service. We even suggest a sampling program to ensure the “water” is being verified to be 100% H20.
But a “Water Diffusion System” that is properly designed, operated and maintained, should indeed reduce the number of NH3 reportable releases from an NH3 process. And merely discharging NH3 vapor to the water diffusion system would NOT be a reportable release. But please understand that any discharge to a “Water Diffusion System” needs to be investigated as a PSM/RMP incident so we can understand how the failure came about. The “Water Diffusion System” is our last line of defense from an NH3 release to the atmosphere, and when we have a failure that led us to this tank, it is a serious failure that needs to be investigated.