I have mentioned “farm leak testing” several times in my ramblings, and every time I caution against industrial facilities, especially those falling under OSHA’s PSM or EPA’s RMP. Essentially, a “farmer’s leak test” uses NH3 as the leak-test medium before using their implement to knife the NH3 into the ground as a fertilizer, as this is common practice in the 1,000-acre cornfield. Although several RAGAGEPs, such as CGA 2.1 and NFPA 58, allow the leak test to be performed using HHC/EHS, it comes with serious risks.
For example, a 30,000-gallon NH3 storage tank’s pressure boundary will be broken for some modifications/maintenance. BEFORE this tank is placed back into service, those locations where the “pressure boundary” was broken should be leak tested (NOT pressure tested!) to ensure the integrity of the tank/piping. The contractor intends to use the “farmers test” method by pumping some liquid NH3 into the tank and allowing it to boil off, creating pressure within the tank and piping. Keep in mind this is a far cry from a farmer filling up a 3-foot-long, 1″ hose with NH3 gas! So I ran the numbers with the help of ChatGPT, and it turns out we would need 1,950 lbs of NH3 to achieve the 25 psig leak testing pressure. Keeping in mind that 100 pounds is the RQ for NH3. Trust me when I say, your back is against the wall when you have to report the release and explain it occurred during a “leak test”. PLEASE UNDERSTAND, if we wish to do our leak testing at a LOWER pressure, then less NH3 would be needed. The leak test pressure SHOULD be quantified in the facility’s Leak Test SOP.
The leak test is not the problem. Using NH3 as your test medium is not the problem. The problem is the VOLUME needed when the tank has to be leak tested. We need to rethink this “farmer’s test” method when we have nearly 2,000 pounds of NH3 involved. A failed test means we can create an IDLH environment around the leak point, requiring an EMERGENCY RESPONSE to stop the leak, all due to a failed “leak test”. And with many facilities disbanding their ERTs, that means a response by the local FD and a call to the NRC if the leak exceeds 100 pounds. Even with an on-site ERT capable of responding to this uncontrolled release, the time it takes to respond versus the time it takes to achieve the RQ is a VERY SMALL window.
It happens more than you may think; it has happened to me several times, which is why I never suggest it and always caution my clients against using the “farmer’s test” method.