A cloud of ammonia coming from the open valve on the P34 receiver. Photo by Technical Safety BC.One worker was killed during the dismantling of an ammonia refrigeration system at the Artic Glacier facility in Kamloops, B.C., on May 26, 2022. The worker opened a one-inch quarter-turn ball valve on a cut-off pipe that was connected to a high-pressure receiver for the P34 ice machine, believing that the vessel was empty.
The vessel was not empty; as a result, there was a large release of ammonia, and the worker was sprayed with the liquid refrigerant. Immediately following the ammonia exposure, the worker moved further into the building away from the receiver. Other workers were able to evacuate the area to the outdoors, but the injuries sustained by the worker who opened the valve were fatal.
It would take more space than we have here to go through every detail of the incident. And if you are interested, Technical Safety BC’s report and appendices are available on its website.
Miscommunication
In this article, we will focus on one of the themes that emerged during the investigation: terminology and language.
The facility stopped producing ice in 2016 and the two on-site ammonia systems (P24 and P34) were shut down. As part of the shutdown, a refrigeration mechanic “pumped down” the systems and stored the ammonia in the high-pressure receiver and “dismantled” the ice machines. There is a note on the invoice for this work that the ammonia will need to be removed later. However, in future conversations regarding the invoicing for that work, the work was characterized in an internal email exchange as “repairs to the facility freezer and NH3 removal” and “NH3 removal and blow down of P24 and P34.”
Additionally, around the time the equipment was being shut down, there was also a call to the fire department about an ammonia odour. The fire department was told that the system was being “decommissioned” and they were “venting residual ammonia.”
This seemingly conflicting information led those who remained in charge of the facility in the following years to believe there was no ammonia in the systems. In 2018, they hired another refrigeration mechanic to remove the evaporative condenser from the P34 system so it could be reused at another facility and this mechanic did so.
They don’t remember checking the rest of the system for ammonia but are certain that they did not find any residual ammonia in the condenser or the associated piping, which reinforced the understanding that the system was empty and didn’t contain ammonia.
The quarter turn valve that was the source of the ammonia leak is shown “as found” in the open position. Photo by Technical Safety BC.More to the story
There is of course more to the story than this, but you likely get the
picture and can see how communication could result in assumptions that the ammonia was removed and that the system was empty. It seems that the contractor who shut down the system and pumped it down vented the residual ammonia from the piping and other components and left the bulk of the ammonia in the receiver.
Several words contributed to miscommunication throughout the end of life of this equipment that are worth discussing.
- Pump down
- Vented
- Disassembled
- Dismantle
- Decommissioned
- Pump out
- Evacuated
I think it is mostly clear in the industry what “pumped down” means, but as I start to think about more complex systems, some of this clarity disappears. For example, in small commercial refrigeration systems, the act of pumping something down almost always means closing the king valve and using the compressor to move most of the refrigerant to the receiver. In some cases where the receiver isn’t big enough (or when there isn’t one, like in residential A/C units), a system would still be considered pumped down as long as the liquid is mostly contained on the high side of the system. I don’t think this definition is controversial. But what about a complex system or larger industrial system where there isn’t a large enough receiver to contain all the charge (which is common), or where there are multiple low-pressure receivers? You would, in these cases, be able to pump down a portion of the system. Would the system ever be considered pumped down? Is that different than a pump out?
To each their own
As a refrigeration mechanic, I think I can understand the following sentence. “We are going to pump out the header and vent the residual ammonia so that we can decommission the piping to an evaporator before dismantling a freezer; the header will have to be evacuated before being put back in service for the other coils.”
Consider the following change: “We are going to pump down the system and evacuate the header so that we can decommission a freezer and disassemble the evaporator; once this is done, the header will be put back into service.”
Do these sentences mean different things? Does their meaning change depending on specific details about the system? Does their meaning change if you are an ammonia mechanic or a freon mechanic? Does the meaning change in different geographical regions?
The truth is that out of everyone who reads these two sentences, there will probably be dozens of interpretations. What does decommissioning mean? Is it permanent or can something be temporarily decommissioned? Can pump out and pump down sometimes mean the same thing? Does evacuation mean using a vacuum pump or does it simply mean emptying something?
An ammonia leak could cause lethal results if not contained properly, making safety training essential for refrigeration mechanics. (Photo courtesy of the Ammonia Safety Training Institute)Creating a safe workplace
I know some people will have strong opinions about the questions I asked, and I have mine. The reality is that we work in an industry that many people don’t understand so it’s important to remove the ambiguity that language and word choice bring.
One of the ways we can do this is by implementing procedures and industry practices that always leave the systems we work on safe. There are standards developed in many industries (including ammonia refrigeration) that outline procedures for properly decommissioning systems and they can provide useful guidance.
There are also steps we can take every day to help when we are working on systems and leaving systems offline. This might include labelling of systems, lock-out and tag-out of isolation valves, marking gauges as “out of service” if they are connected to the system but have been isolated or vented. And the list could go on.
Going back to the Kamloops incident, it seems clear that communication played a central role in leading the worker who opened the valve to believe the system was empty. This worker was not a refrigeration mechanic, and neither were any of the riggers that were there performing the demolition. They didn’t think they needed a refrigeration mechanic because the system was empty.
There is much more that could be discussed concerning this incident, and I have skimmed over some of the explanations of what exactly happened for the sake of brevity. The Technical Safety BC report does a good job of explaining the incident in much more detail. Perhaps in a future article, we can discuss some of the engineering analysis that was done to support the investigation.