Scuba Cylinder Long-Term Storage: Fact and Fiction

[Doc Harry]

Yes, Dr Harry, 6351 contains some lead and yes, the theory was that cracks occurred in "migration" areas. However, attempting to discriminate after the fact between migration and defective mix at the outset might be difficult. It is possible that too much lead was used in the first place. There have been explosions, one in Miami, that occurred in absence of any corrosion or slag. In fact, it happened a couple days after hydrostatic test and visual inspection. Every time an anomaly like this happens there is more fancy footwork from the experts. Explanation of the Miami deal was along the lines of "instantaneous, catastrophic failure, not SLC related". The details of the accident sound a lot like the "slag" tank. Are we talking about the same thing here? If so, it is new to me but the wheels keep grinding, even when I'm not watching. However, if there can be multiple explanations of similar accidents there can also be more than one interpretation of "load". Moreover, we might want to focus our attention on SLC as it relates to the other aluminum alloy, 6061; here, Luxfer says "no problem". I agree. It never was.

Here's the link to the slag-related explosion analysis:
http://hazmat.dot.gov/pubs/reports/cylinder/dot3al_rupture_pgs1_14.pdf

Here's the the link to the DOT cylinder analysis failure report page:
Cylinder Failure Analysis Reports

Here's the link to the DOT information page on 3AL cylinders:
3AL CYLINDERS

It's great, there's so much stuff on line these days compared to 10 years ago.... Even a lot of old reports are now on line as PDF.

 

[SparticleBrane]

Every time a cylinder is pressurized to operating load it expands ever so slightly, even steel cylinders. Some of the expansion is permanent. The longer that a cylinder is pressurized, the greater the permanent expansion as a percentage of total expansion. This is what hydro testing checks. Are you saying this is not true? Or are you saying something else?

I believe what he is saying is that steel tends to be very 'stretchy' and as long as you aren't stretching it past the elastic limits, it can essentially bounce back and forth forever.

-----------------------------------------------------

While I agree with the findings of your study (thanks for all the time and effort you put into it!), I have to say that there are, as with everything in life, advantages and disadvantages in the way scuba cylinders are stored.

In my case, I recently completed a cabinet to store my cylinders (ok that's a lie, I need to finish staining/varnishing it...). Space for up to 4 sets of doubles and 4 single cylinders, with the singles on their sides. I would really like to use this as it helps keep much of my gear in the same location. After looking over all the evidence presented, it looks like I'll have to open up my single cylinders stored horizontally a bit more often. I don't want all those woodworking hours to go to waste.

Also, I prefer to store cylinders full (or at least "not close to empty") for two reasons:
1: Since I get my air "free" at school ("free" meaning I volunteer a lot of hours/week with OW classes) I don't always have access to the fill station whenever I want it -- I have to work around other people's schedule. I like to get my cylinders filled when I have the chance so they're ready to go if my buddies and I decide to go for a dive at some random time. If we stored cylinders empty, we couldn't decide on Saturday to go for a dive on Sunday, unless we want to pay for fills. Why pay when we can get it for free, and fill our own cylinders? If we have 1500psi left over after a dive, I'm not big on the idea of emptying our cylinders down to ~300psi just to possibly reduce corrosion -- I would rather leave it, so there's less air to pump and less heat generated which would result in a short fill.
2: I believe the fire issue was mentioned above -- if you store cylinders empty and there's a fire, the full cylinders have their burst discs blow before the cylinder's alloy fails, unlike the empty cylinder, which would likely go boom.



It would appear that the old adage "An ounce of prevention is worth a pound of cure" is quite applicable here. Follow a few simple rules and your cylinders should last a good long time...

• Only have your cylinders filled at places that pump clean gas with a minimum of moisture. If you suspect otherwise, open up your cylinder to check and see. Better to find compressor oil or gunk before the dive rather than after...
• Never drain your tank all the way down while diving. If you do, open it up, clean it, dry it, and make sure it's good to go before re-filling.
• If possible, try to store cylinders, especially steel cylinders, vertically. In my particular case I'll just check them more often.

In a perfect world, moisture and gunk would never get inside any cylinder and they would all be pristine inside. Unfortunately we know this isn't the case -- hence why we have annual visual inspections. I suppose the big message to take away is that preventing corrosion is better than fixing corrosion.

 

[Luis H]

Every time a cylinder is pressurized to operating load it expands ever so slightly, even steel cylinders. Some of the expansion is permanent. The longer that a cylinder is pressurized, the greater the permanent expansion as a percentage of total expansion. This is what hydro testing checks. Are you saying this is not true? Or are you saying something else?

Yes, I am saying that is incorrect.

To incur any (however miniscule) permanent expansion/ deformation you have to stress the metal beyond its elastic limit (also know as yield strength). In the case of a pressure cylinder, they are designed so that the hydro test pressure is at the lowest end of the yield strength.

The hydro test is measuring any miniscule permanent deformation which is a reflection of stress exceeding the material yield strength limit.

The yield strength (or elastic limit) is exceeded if the material properties have been compromised by fire (heat affecting material heat treatment) or the average wall thickness has been reduced by corrosion.

That being said, metals are not totally homogeneous. That means there are variations even on one chuck of steel as found in a pressure cylinder. That is the reason why most steel structures (including pressure vessels) are normally design with a maximum working limit much lower than its yield strength (in this case 60% of the lowest expected yield level). This should provide sufficient safety factor to avoid any yielding at all.

We always talk about a yield point, but in reality it is a yield stress region and for different metals it is defined slightly different. It is not a precise stress point were the metal is purely elastic below that stress level and then the deformation becomes what we call plastic deformation.

As the work implies, elastic deformation (below the yield stress) level means that when the load is removed the material will fully return to its original shape. Where plastic deformation means there is some permanent deformation.

Most metals (in particularly steel) are affected by time under static stress. A steel structure in theory can stand for ever if it is not attached by corrosion or cyclic loadings.

Steel cylinders are affected by corrosion, fire, and over stresses, time in itself is not an issue.

Now aluminum...that can be kind of a weird metal in some respects.


During hydro test of steel cylinders it is very important to do a round out procedure because the test is intended to check for yield stress in the material not the geometry change of a slightly out of round cylinder.

 

[Luis H]

It would appear that the old adage "An ounce of prevention is worth a pound of cure" is quite applicable here. Follow a few simple rules and your cylinders should last a good long time...

• Only have your cylinders filled at places that pump clean gas with a minimum of moisture. If you suspect otherwise, open up your cylinder to check and see. Better to find compressor oil or gunk before the dive rather than after...
• Never drain your tank all the way down while diving. If you do, open it up, clean it, dry it, and make sure it's good to go before re-filling.
• If possible, try to store cylinders, especially steel cylinders, vertically. In my particular case I'll just check them more often.

In a perfect world, moisture and gunk would never get inside any cylinder and they would all be pristine inside. Unfortunately we know this isn't the case -- hence why we have annual visual inspections. I suppose the big message to take away is that preventing corrosion is better than fixing corrosion.

That is good advice, but I would add that most important is to make sure that the valve fill connection is dry before filling a tank. I personally fill y tanks in most dive shops or very carefully supervise the individual filling it.

I blow my valve and make sure the filling whip is also blow dried.

There is no faster way of pushing water into a tank that with high pressure air behind it.


I have been storing my tanks full for close to 40 years now and never had an issue. But, I have always done my own VIP and have always been extra careful not to allow moisture into the tanks.


Only once about 35 years ago I allowed someone to fill my cylinders in a marina that used a salt water tank for cooling water. I suspected something so wrong with that fill so I VIP my new tanks. I was tumbling them on Monday when I went back to work (I bought them from the LDS I was working 5 days before).

 

[Luis H]

I just did a small analytical study for PSI/ PCI on the subject of ideal tank pressure in a fire and the observations are interesting. This was just based on analysis not testing.

The calculations obviously show that the higher the initial pressure, obviously the lower the temperature required to burst the relief pressure disc.

It was interesting, looking at the calculated data for material strength as a function of temperature and noticing that the burst disc will also let go at lower pressures as the temperature rises. Now it seems obvious.

The bottom line is that higher pressure will release earlier in a fire. But, in a steel tank as long as you have a reasonably functional burst disc, it should let go before the cylinder structure is compromised for any given gas pressure. Copper is affected by lower temperature heat than steel.

For aluminum tanks the latest recommendation (from Luxfer) is to have them at least half full, but preferably higher.

 

[SparticleBrane]

That is good advice, but I would add that most important is to make sure that the valve fill connection is dry before filling a tank.

It must be late...I can't believe I forgot to mention that. I ALWAYS make it a habit to thoroughly blow out any moisture left in the valve, especially in DIN valves. It's amazing the amount of water that can be trapped in those threads...

I've never used one but DeepSeaSupply sells what appears to be a nice tank inspection light, with a cord long enough to use on large bank bottles if need be.

 

[SparticleBrane]

Any ideas when the tank o-ring would melt?

 

[Doc Harry]

Yes, I am saying that is incorrect.

To incur any (however miniscule) permanent expansion/ deformation you have to stress the metal beyond its elastic limit (also know as yield strength). In the case of a pressure cylinder, they are designed so that the hydro test pressure is at the lowest end of the yield strength.

Okay, I understand. At normal operating pressures (e.g., 3,400 psig) the steel is completely elastic.

At hydro pressures (e.g., 8,130 psig) the steel is in the lower end of the inelastic range.

Thanks for the explanation.

So it doesn't matter if you store your steel tanks empty or full. How about Aluminum?

 

[Luis H]

Any ideas when the tank o-ring would melt?

Depends on the actual rubber the useful temperature varies, but many rubbers will not just melt.

In any case, if someone is dreaming that the O-ring will behave as an alternate pressure relief...it is not going to happen. Yes, it could leak, but if the valve was set properly the gap between the tank and valve is not large enough to relief the pressure fast enough...even if the O-ring disappeared, which it will not. A deteriorated O-ring will still tend to seal in the tank neck geometry. By design the gap should be zero or very close to zero.

Even if the O-ring seal starts leaking, it will still tend clog the metal to metal contact on the neck area. The neck O-ring will only leak slowly. It was designed as a very effective seal, not as a gas/ pressure release.

 

[SparticleBrane]

Okay, I understand. At normal operating pressures (e.g., 3,400 psig) the steel is completely elastic.

At hydro pressures (e.g., 8,130 psig) the steel is in the lower end of the inelastic range.

Thanks for the explanation.

Just out of curiosity, where did you get that 8,130psi number? :shocked2: :shocked2:
Tanks are hydro'd at either 5/3rd's of working pressure or 3/2 or working pressure.

Most "normal" non-exemption/non-special-permit cylinders are tested at 5/3rd's -- so 3000psi Al80s are tested at 5000psi, and 2400psi LP steel cylinders get tested at 4000psi.

The 3442 series gets tested at 3/2 working pressure, or essentially 5,250psi. Note "TP5250" stamped on your cylinders -- "Test Pressure 5250".


That makes sense Luis, thanks for the info.