This article is for ANYONE with anhydrous ammonia, regardless if you have 100 pounds or 100,000 pounds of ammonia and your primary means of containment is Carbon Steel. I have written numerous times about the Stress Corrosion Cracking (SCC) failure mode and shared numerous accidents and alerts regarding this failure mode. In this article, I want to point out a little PSM/RMP requirement that often gets overlooked…
1910.119(f)(1)(iii)(D) Quality control for raw materials and control of hazardous chemical inventory levels
I have written about how key it is to control our inventory levels, but now we need to discuss the “quality control for raw materials” and in regards to using anhydrous ammonia, it is never more critical. Now, a lot of refrigeration processes will sample their ammonia for water content; unfortunately, this sampling is often done for “efficiency reasons” and not process safety reasons. Actually, too much water in your ammonia hampers the ammonia’s ability to act as a refrigerant. Anytime the process is opened, we allow humidity to enter the process, and without proper purging of the process, water will begin to accumulate in the ammonia, and thus, efficiency goes down and costs go up. But I want to talk about controlling the water content in our NH3 from a process safety perspective.
Anhydrous Ammonia comes in three (3) different grades for different uses.
- Metallurgical Grade is at least 99.999%
- Refrigeration Grade is at least 99.98%
- Commercial Grade (also called agricultural grade) is at least 99.5%
It is important to point out that the “refrigeration grade” NH3 is stated to have at least 0.02% water – OUTSIDE our SAFE LOWER LIMIT for water in our NH3 when the Primary Containment is Carbon Steel. Remember the articles and alerts about SCC failures, and the critical factor in these failures was due to the lack of water content in the ammonia solution, specifically the minimum of 0.2% water? Well, as we can see from above, using Metallurgical Grade NH3 (99.999%) or Refrigeration Grade (99.98%) in a process that uses carbon steel is setting us up for a LOPC event. Sure, 99.999% NH3 would perform better as a refrigerant. Still, without the proper process design (i.e., materials of construction and/or Post-Weld Heat Treating), we are setting ourselves up for an LOPC event.
In some situations, we may use our MI program to identify SCC issues; however, this “control” is too far down the chain of controls, and there is a much better means to control this risk, and it’s OSHA’s and EPA’s requirement to control the quality of our raw materials in our process (e.g., 29 CFR 1910.119(f)(1)(iii)(D) and 40 CFR Part §68.69). In doing so, we should establish safe upper and lower limits for ammonia concentrations in our process and then test/sample our ammonia on some established frequency to ensure we maintain our quality within our defined operating envelope.
NOTE: We should also include a reminder in our Truck or railcar unloading SOP(s) to ensure that the Certificate of Analysis (C of A) clearly states that the grade of NH3 being received is at least 99.8% ammonia or less.
One shipment of a couple thousand pounds of Metallurgical Grade or Refrigerant Grade ammonia being added to a process with over 10,000 pounds of ammonia will NOT be the end of the world as we know it. Not establishing a quality control process for our NH3 grade(s) and then running a process that uses carbon steel as the materials of construction with ammonia that contains less than 0.2 % water can lead to a premature failure of the containment system via SCC and result in injuries and environmental issues.
And finally, dare I suggest that if we are not doing either the quality control or the MI inspections for SCC, our vessels and piping should be Post-Weld Heat Treated, and this should be clear in our documentation for our vessels (U-1 forms) and our piping specs.
See my articles on PWHT for more on this topic.