150# Ammonia Cylinder Rupture

[chiefbecker]

We recently had this cylinder rupture during the transfer of ammonia into a flooded refer system. The cylinder was connected with a 5/8" charging line to the suction manifold of a horizontal plate freezer. The freezer was in operation at the time of failure. It was about 2 hours into a 3-1/2 hour freezer cycle. The bottle was in the vertical position and had a frost line around the bottom third of the bottle.
The suction pressure in the manifold was running right around -1 to 0 psi. The manifold has a safety valve that is set to release at 203 psi and it has a hot gas return valve set at 85 psi. Both valves were operational at the time of failure. The compressors have an alarm at 200 psi on the high side and shut down at 250 psi.

The first thought when seeing the failure is an over pressure situation. If that is the case, I have no idea of where we could generate the required pressure to burst the cylinder.

Any ideas?

Rupture 2.jpg

[RANGER1]

Chiefbecker,
So no one hurt I hope?
When charging, is it just vapour, not liquid with cylinder standing upright, as that would take a very long time.
Normally you would charge liquid in liquid line making it reasonably quick, like 10 minutes.
Vapour would take hours with probably no one watching it.

You would have thought cylinder would have a relief valve to protect it as well.

The reason most likely is hydraulic expansion of cylinder 100% full causing it to rupture, how I do not know exactly.

Is the suction line gas only, or liquid/gas being flooded unit.
If bottle chilled down liquid could migrate back into bottle overfilling it, but just guessing.

[Segei]

I agree with Ranger1. I never heard about charging ammonia system by the vapor only liquid. This charging should not be left unattended. Another guessing that manifold has several valves and one of them is not tight and worked like check valve. It let vapor/liquid get into the cylinder but not back. Overfilled and ruptured.

[gesi]

not enough of informations. any fire or noise? all valves worked correctly? Did during all the time was any technicians there? Ambient temperature?
In my opinion cylinder was filled liquid to much, with no out from cylinder. Temperature was raise and cylinder rupture.

[NH3LVR]

I agree something unusual happened. It would have to have filled with liquid and then be shut off somehow.
I have seen one cylinder with a rupture like this. It was dropped off a crane on to the deck. Very similar rupture.

[chiefbecker]

not enough of informations. any fire or noise? all valves worked correctly? Did during all the time was any technicians there? Ambient temperature?
In my opinion cylinder was filled liquid to much, with no out from cylinder. Temperature was raise and cylinder rupture.

There was no fire in the space. All of the valves were working correctly. The plate freezer was operating in freeze mode at the time of the rupture. In this mode, the only valves that are open are the suction valve on the low side header and the liquid feed valve on the opposing header. The liquid pumps are regulated to 45 psi and dump over a relief if the liquid pressure exceeds that setting. The suction pressure at the plate header was around 0 psi to -2 psi.

At the point in the cycle, there should have been a limited amount of liquid coming through the plates into the suction header. The service port is at the top of the header. This header is 4" pipe and is vertical. The header is connected to a line that drops through the deck and into the cargo hold overhead. There is just one manual isolation valve in that line before it enters the low pressure receiver that is located in the lower engine room.

The bottles location was about 30' from the service port/freezer in the breezeway and was secured in a vertical position to a set of hand rails for an access stairway into the main engine room. The ambient temperature in that space is normally around 70^f. From the report from the vessels engineers, there was a frost line on the bottle about a third of the way up from the bottom at the time of the rupture. There are also two relief vales in the top of the header. One is the hot gas return regulating valve set at 85 psi and the other is a safety valve set to open at 203 psi. The hot gas return valve is working correctly.

I don't know where we would have gotten the liquid from to fill the bottle from the suction header. Any liquid that would be coming through the plates would drop into the header and return to the LPR. To reverse flow into the bottle, the pressure in the suction header would have had to have been higher than the pressure in the bottle. The pressure would have had to exceed the ambient temp/pressure of around 100 psi. The hot gas regulating valve would have had to have failed in order for it to reach that high of a pressure. It is working correctly at this time......

[chiefbecker]

From my understanding, the valve on the bottle at the time of rupture was open. When we received the bottle for inspection, it was in the open position. In order to completely fill the bottle, we would have had to have a high pressure pump to fill it with. At least one that could overcome the ambient pressure. The only liquid pumps that are in the system are set to relieve at 45 psi. The bottle was connected to the lower side suction header and had a direct connection from the header to the LPR.....

[chiefbecker]

Charging Port Photo 5.jpg

This is a picture of the suction header on the plate freezer. The service port is just below the pressure gauge.

[Segei]

Your liquid pump can fill up this bottle very easy because liquid is cold and subcooled. Even if the bottle initially had some pressure cold liquid will condense vapor inside the bottle and bottle pressure will drop. How cold liquid got into the bottle is another question?

[RANGER1]

So when bottle was connected it was vertical,
they were charging with gas, or does it have a dip tube?

[Brian_UK]

I wonder whether in addition to any of the possible gas/liquid movements there has been any impact damage suffered by the cylinder in its present location. Even a mechanic wacking it with a spanner on passing can inflict damage.

[Rob White]

I wonder whether in addition to any of the possible gas/liquid movements there has been any impact damage suffered by the cylinder in its present location. Even a mechanic wacking it with a spanner on passing can inflict damage.

So when my dad used to store cylinders in our yard in the 70's and I used to play
drums with them, with a big stick or a metal bar that I found, that might not have been the safest thing.
Is that what you are saying.

Rob

.

[Brian_UK]

So when my dad used to store cylinders in our yard in the 70's and I used to play
drums with them, with a big stick or a metal bar that I found, that might not have been the safest thing.
Is that what you are saying.

Rob

.

Maybe not Rob but stranger things have happened. What age is that cylinder, why the heavy rust marks, how secure was the mounting, who knows?

[chiefbecker]

So when bottle was connected it was vertical,
they were charging with gas, or does it have a dip tube?

The bottle was in in the vertical position. They were charging with gas. The only reason that they do this, is it's safer in rough weather to just keep the bottle secured instead on laying it down and re-securing it in a horizontal position. From what the vessel engineers have been telling me, it usually takes a couple hours to empty a bottle by this method. If there is any residual liquid in the bottle, when the freeze cycle complete on that freezer, they will isolate the bottle, allow it to warm up and then start charging again when the suction header is back down around 0 psi during the next freeze cycle.

[chiefbecker]

Your liquid pump can fill up this bottle very easy because liquid is cold and subcooled. Even if the bottle initially had some pressure cold liquid will condense vapor inside the bottle and bottle pressure will drop. How cold liquid got into the bottle is another question?

We have tried to back fill these bottles when we are removing the liquid from the plant when the vessel comes back for service. With the liquid pumps we have, you can only get them about half full before they stop filling. We back fill 1000# R-22 cylinders every year on the vessels that use that refrigerant. Those bottles have the ability to connect a suction line to them to reduce the pressure in the cylinder when we are filling them.

We always remove the refrigerant charge from all of our vessels when they return each November for maintenance. This vessel has about 7000#'s for a full charge.

[chiefbecker]

Maybe not Rob but stranger things have happened. What age is that cylinder, why the heavy rust marks, how secure was the mounting, who knows?

The cylinder is a standard DOT 4AA480 Cylinder. It was date stamped February 2019. There are no prior date stamps on it. The rust marks on the cylinder are from a previous vessel storing them on the deck of the boat somewhere. These cylinders take a beating when out on deck for any length of time. We normally store them below deck and out of the salt. The cylinder was in a rack built for that purpose and secured with chain binders.

I received some additional info on the cylinder. It was manufactured in 1964 by Pressed Steel. It was re-qualified on 2/15/19 and met all of the DOT requirements.

[seanf]

What age is that cylinder, why the heavy rust marks?

What Brian said.
Whats the condition of inside the cylinder?
Any marks on the fracture surface that would suggest a certian failure mode?
Is the thinkness of the steel uniform around the cylinder?
Was the cylinder from a know source?

[RANGER1]

Chiefbecker,
Just thoughts.

Is the cylinder rented or owned outright?
Could it have been overfilled this or another time in it's life!
If inspected a few years ago, nice to know how it gets recertified, as in what do they do, what standard, paperwork etc.

If supplied by someone they should investigate, or maybe need to get checked out by a vessel inspector or similar.
Do not know how or why, but to me can only be hydraulic expansion forces from a overfull cylinder, as things do not rupture at 1psi, if before connecting cylinder to system was probably 100psi.

On safety subject, this type of cylinder looks a bit vulnerable with valve sticking up with no guard to protect it, although has been around since 1964.
Charging hoses, do they get checked out & tested as well.
If you were to buy a new cylinder today, wonder what would be the difference, in standards, or protection of cylinder.

[chiefbecker]

What Brian said.
Whats the condition of inside the cylinder?
Any marks on the fracture surface that would suggest a certian failure mode?
Is the thinkness of the steel uniform around the cylinder?
Was the cylinder from a know source?

The inside of the cylinder had some rust. More than likely it was from where it was placed after the failure and before it was shipped to us. There was no indication on the exterior surface of any kind of impact damage or excessive corrosion. The material thickness looked to be uniform in the area of the failure. The ammonia was purchased from our vendor in Alaska. We pay a monthly rent on the bottles that we keep onboard the vessel. They rotate on and off the vessels as they are consumed.

[chiefbecker]

Chiefbecker,
Just thoughts.

Is the cylinder rented or owned outright?
Could it have been overfilled this or another time in it's life!
If inspected a few years ago, nice to know how it gets recertified, as in what do they do, what standard, paperwork etc.

If supplied by someone they should investigate, or maybe need to get checked out by a vessel inspector or similar.
Do not know how or why, but to me can only be hydraulic expansion forces from a overfull cylinder, as things do not rupture at 1psi, if before connecting cylinder to system was probably 100psi.

On safety subject, this type of cylinder looks a bit vulnerable with valve sticking up with no guard to protect it, although has been around since 1964.
Charging hoses, do they get checked out & tested as well.
If you were to buy a new cylinder today, wonder what would be the difference, in standards, or protection of cylinder.

The cylinder was a rental.
It could have been overfilled in it's past life, but I have no idea of were it was filled or how many fillings that it may have had during it's years of service.
The bottle is currently with a forensic engineering company that is going to do a thorough job of trying to determine the cause of failure.
The bottles that we receive and store on board all have caps on them. We do not allow uncapped bottles to be loaded on the vessel and the gas suppliers will not accept one without the cap. It just didn't come down with the failed cylinder.
We found out who the company was that did the inspection in 2019. I have not seen the documents for that specific bottle yet, but they assured us that it met all DOT requirements at the time of inspection. From my understanding, this bottle came to them in a fairly large shipment of empty cylinders from a gas supplier that required the testing.
As far as the DOT requirements for these cylinders, I would tend to believe that the older bottles would need to meet current rules as far as the testing. I don't do a lot with low pressure bottles, but I do have a lot of experience with high pressure scuba bottles. I had some older aluminum tanks that I had to scrap because they could not meet the current regulations for testing. It had to do with the thickness of the material that is threaded for the valve installation.

[RANGER1]

The cylinder was a rental.
It could have been overfilled in it's past life, but I have no idea of were it was filled or how many fillings that it may have had during it's years of service.
The bottle is currently with a forensic engineering company that is going to do a thorough job of trying to determine the cause of failure.
The bottles that we receive and store on board all have caps on them. We do not allow uncapped bottles to be loaded on the vessel and the gas suppliers will not accept one without the cap. It just didn't come down with the failed cylinder.
We found out who the company was that did the inspection in 2019. I have not seen the documents for that specific bottle yet, but they assured us that it met all DOT requirements at the time of inspection. From my understanding, this bottle came to them in a fairly large shipment of empty cylinders from a gas supplier that required the testing.
As far as the DOT requirements for these cylinders, I would tend to believe that the older bottles would need to meet current rules as far as the testing. I don't do a lot with low pressure bottles, but I do have a lot of experience with high pressure scuba bottles. I had some older aluminum tanks that I had to scrap because they could not meet the current regulations for testing. It had to do with the thickness of the material that is threaded for the valve installation.

Hopefully you can pass on what forensic test concluded.
Thanks

[NH3LVR]

I will look forward to the Forensic results.
I am assuming that you were not on board at the time. I mention this because as I said the only failure I have seen was from mishandling. Perhaps you are not getting the full story?

[chiefbecker]

I will look forward to the Forensic results.
I am assuming that you were not on board at the time. I mention this because as I said the only failure I have seen was from mishandling. Perhaps you are not getting the full story?

I will post the results of the forensic tests once I get that information. It may be a while as the group completing the tests want to tour the vessel once it arrives back in Seattle on November 15th. They are also trying to get all of the players in this event together to witness the exams.

I was not on the vessel when this event occurred. I do have trust in the information that I have received from the engineers involved. The cylinder could have been dropped by the deck crew, but I have no direct knowledge one way or the other on that. It could have been miss handled by the dock side crew or the company we purchased it from as well.

From my understanding of our history with this bottle is, it came onboard in March. It was removed with a number of empty cylinders and returned to the supplier. The supplier found that it still had about 3/4's of liquid left in the bottle and they returned it to us in April so that the crew could finish emptying it. So when they started to charge it into the system it was only a partial at the time.