Ep 69 Fernie Arena Ammonia Accident
Engineering News – Harvesting Freshwater from the Ocean’s Surface (1:10)
This week's engineering failure is the Fernie Arena Ammonia Accident (4:00). Built over 50 years ago, the arena (9:15) and its ice making equipment (11:10) were in need of repair. After ignoring warning signs for years that the plant was failing (15:35) in October 2017, a tragic accident occurred (22:35) sparking an investigation into what went wrong in Fernie (27:55).
Sources:
Engineering News
https://www.sciencedaily.com/releases/2022/12/221206083115.htm
Fernie Arena Ammonia Leak
Episode Summary
Hi and welcome to Failurology; a podcast about engineering failures. I’m your host, Nicole
And I’m Brian. And we’re both from Calgary, AB.
This week in engineering news, harvesting freshwater from the ocean's surface.
When ocean water evaporates, it leaves the salt behind, creating freshwater vapor in the air above.
A study from the University of Illinois Urbana-Champaign, 200km south of Chicago, suggests new infrastructure to harvest the ocean water vapor as a solution to fresh water supply shortages.
The study looked at 14 water stressed regions globally and found they could feasibly capture oceanic water vapor as a solution.
One of the stronger projections of climate change is that humid areas will get more humid and dryer areas will be more dry. But the projections also show that oceanic vapor will continue to increase so this solution is more robust to climate change than others. Also, we have tried wastewater recycling, conservation, cloud seeding, and desalination techniques and while they offer some relief, they are inadequate.
While this is a really interesting solution to the lack of freshwater, I personally think that maybe we should stop relocating to the desert, but I don’t even know how one would accomplish that. And pretty much every area is prone to some kind of weather related disaster. Although California has a few between deserts, wildfires, and earthquakes.
If you want to read more about the study, check out the link on the web page for this episode at failurology.ca
Now on to this week’s engineering failure; the Fernie Arena Ammonia Leak.
I think this is the closest to home we’ve come with any of our failures to date. And while this is a tragic story, arenas are a huge part of Canadian culture, especially in small towns. So we are talking about the ammonia leak today in hopes that it prevents another similar incident from occurring.
Fernie is located in SE British Columbia, about 200 km south and 100 km west of Calgary. It’s about a 3hr drive from here to Fernie, depending on the roads and it has some phenomenal skiing and snowboarding.
The City of Fernie has a population of just over 5,000 people, with about 7,000 in the metro area. It’s the only city-class municipality in Canada that is fully encircled by the Rocky Mountains.
On the way to Fernie, you pass what used to be the town of Frank (named after Henry Frank, who along with his partner Sam Gebo, developed the first coal mine in the Crowsnest Pass), which was destroyed in a landslide almost 120 years ago. While the story of Frank is more of a geoscience failure, it is still fascinating. So for those of you who are on our patreon, stay tuned for the next mini failure where we’ll talk about Frank Slide. And if you aren’t on our Patreon, check it out; it’s only $5/month, or bones, or clams, or whatever you call them.
Fernie Memorial Arena
The Fernie Memorial Arena, where the ammonia leak occurred, is located right in the middle of town, off of highway three, commonly referred to as the Crowsnet Highway. The arena site is an entire city block and also houses the curling rink and community centre. The City of Fernie, under their Leisure Services Department, owns and operates the rink and its ice-making equipment.
The arena is over 50 years old and a compressor room between the arena and curling rink houses the refrigeration equipment that makes ice for the entire complex. The compressor room has a main room where most of the equipment is located, an entrance room to the SE that is separated from the main room by a steel mesh wall with a lockable door, and a northwest entrance for controls, and emergency showers. On the southwest side of the main room, there is another set of double doors that were sealed and bolted shut when the leak occurred.
There are three emergency stop buttons for the refrigeration system, one in the NW vestibule, one at the SE exit, and one outside that connects to an emergency discharge valve and a vent stack.
Ice Making Basics
In order to make ice, a chiller plant pulls all of the heat out of the brine water, or salt water. The salt in the water allows it to be much colder than 0°C before freezing. I don’t know the standard brine concentration, but as an example, a 23.3% salt concentration has a freezing point of -21°C. That brine water is then pumped through a series of pipes within the concrete slab below the arena ice pad. When the slab is cold enough, they flood the slab with water and that water freezes. Depending on the use of the ice surface, whether it be general skating, hockey, figure skating, etc the ice temperature and the way they layer the flooding to create the ice are all different. There’s a very specific science to it, this is a really generalized explanation.
As the brine water circulates through the concrete, it picks up heat due to the concrete being warmer than the refrigerant; i.e. the concrete rejects its heat into the brine water. The chiller plant uses a refrigerant to reject that heat to outside, cool the brine water back down and send it out through the concrete again.
Sequence of Events
Just want to preface the sequence of events by saying there were alot of things going on. We attempted to summarize the events from the BC investigation.
On the day of the incident, the curling rink chiller had been in operation for about 30 years, suggesting it was not original to the building, as the complex was 50 years old. The average life expectancy of a chiller is 20-25 years. And the way it's going these days with refrigerants changing so fast, you likely want to upgrade in that time frame just to be able to easily purchase replacement refrigerant when you do repairs.
Starting in 2010, the refrigeration mechanics contracted to service the chiller plants had flagged that the chiller was past its life expectancy and should be replaced. Even though a master plan was developed in 2013 to undergo significant capital upgrades, the work was expected to be completed over a seven year period and the City was continuing with maintenance until the plant could be replaced.
And this is common. Chiller plants aren’t cheap. And if it's working, many people try to keep it going until they can get the funds in order to replace it. The challenge with this, is that additional assessments are needed to assess risk and worker safety and those don’t appear to have happened here.
Throughout 2015 and up until the leak in October 2017, further discussions were had about replacing the plant, several quotes were received, and individual components were replaced or rebuilt to keep the plant running until it could be replaced. Over this time, periodic ammonia leaks occurred, some triggered the monitoring system, some did not. The sensors were replaced, settings were adjusted, and various other repairs were completed.
In April 2017, the plant was shut down for the summer, the ammonia was pumped out into a receiver, and the operators noted a smell of ammonia that was not normal. The brine water was tested for ammonia, which the ammonia and brine water loops should be completely separate with no cross contamination between them, and 3320 ppm of ammonia was present in the brine water as well as higher than normal iron levels at 31 ppm. The recommended concentration of ammonia in the brine water is 0 ppm and the range for iron is 10 ppm. This suggests there was a leak somewhere between the two systems.
A second brine water sample was taken in August 2017 and found 1830 ppm of ammonia and 61.5ppm of iron in the brine water. Again there was another discussion about replacement of the old equipment, but no action was taken. Throughout 2016, the compressors were making some strange noises.
Day of the Leak
At 3:52am on October 17th, the monitoring company received an ammonia alarm for the compressor room and the fire department was called to the arena. Workers for the arena met them on site, presumably contacted by the monitoring company as well.
When firefighters entered the compressor room, with breathing apparatus on; the warning lights were flashing, and they detected 300ppm of ammonia from their handheld devices.
At 4:24am, the workers and fire department re-entered the compressor room to shut down the curling rink equipment using local controls and valves within the room. They opened the SE door to provide extra ventilation and tried to reset the ammonia alarm, but it kept coming back on.
At 4:48am a refrigeration mechanic called his supervisor and told him that the curling rink plant was off,but they were going to try to restart the arena plant.
The arena workers told the fire department that the curling rink would likely be down for the season, that they would lock up, and that the fire department was free to go.
At 5:09am, the arena worker was contacted by the alarm company. He advised them the fire department was on site, but didn’t mention they were leaving, and told the alarm company to disregard alarms for the next two hours.
At 5:18am, the refrigeration mechanic told the refrigeration supervisor over the phone that the gas compressors were not working and the oil was foamy. The supervisor noted that there was brine water in the oil and ordered replacement oil and set a plan in motion to have it replaced.
At 7:31am the two hours to disregard alarms had expired and the arena worker responded that they needed until 4pm to complete repairs and no dispatches should occur until then.
At 9:03am the refrigeration mechanic sent to replace the oil arrived on site. As per the report, he was not told about the earlier ammonia leaks, only about the oil replacement.
The last outgoing call from the three workers, two arena workers and the refrigeration mechanic was at 9:27am.
Despite a strong smell of ammonia two blocks away, cleaning staff claim they didn’t smell ammonia within the area.
The workers were not found until 12:45pm when an electrician arrived to complete work at the facility, went around to the compressor room and found the workers on the floor.
At 1pm the fire department arrived again and obtained readings of 400ppm of ammonia in the compressor room and evacuated the arena. Due to rising ammonia readings, at 2:07pm, the fire department evacuated the area surrounding the arena.
Investigation
WorkSafeBC completed an investigation, their report provided a lot of the information for our research.
They noted the following causes of the leak → Efforts to restart both the arena plant and curling rink plant should not have occurred after finding high concentrations of ammonia in the brine water after the spring shut down earlier that year. And the hole between the brine and ammonia circuits, which was reflected in the ammonia concentration in the brine water, subjected the brine piping to a much higher pressure than was allowable and created a large pipe failure and the subsequent leak.
It was discovered through the course of the investigation that the welding method used by the chiller manufacturer created welded seams that were highly susceptible to corrosion in the presence of chlorides, such as the calcium chloride brine that was used at the arena. The welded chiller tubes were only designed to handle an aqueous brine.
It's unclear from the investigation if this limitation was known during design, or if the brine solution was changed at some point.
The City of Fernie did not have a process to address or assess the potential safety hazards of the aging equipment during start up or shut down.
The refrigeration service company also did not provide direction or caution against restarting the chiller plants, but rather recommended additional monitoring.
The incident response measures were not present, resulting in workers not wearing protective equipment, the fire department was sent away, and the monitoring company was told to ignore further alarms.
The chiller plant was replaced with one with a synthetic refrigerant and the area reopened for the 2018-2019 season.
So there you have it, aging equipment, ignored operating issues, and a lack of safety policies led to the ammonia leak in Oct 2017 and the loss of three lives. This accident, like pretty much all of the others we discuss, was entirely preventable.
For photos, sources and an episode summary from this week’s episode head to Failurology.ca. If you’re enjoying what you’re hearing, please rate, review and subscribe to Failurology, so more people can find us. If you want to chat with us, our Twitter handle is @failurology, you can email us thefailurologypodcast@gmail.com, you can connect with us on Linked In or you can message us on our Patreon page. Check out the show notes for links to all of these. Thanks, everyone for listening. And tune in to the next episode, which is episode 70, an engineering marvel. We’re going to be talking about one of the American Civil Engineering Society’s seven wonders of the engineering world, the Golden Gate Bridge.
Bye everyone, talk soon!