Overfilling LP Steel Tanks -- How bad is it?

[Mark Michaud SELAUSAR]

Did you look at the list I posted?
http://www.scubaboard.com/forums/4558800-post218.html

Nov 2001 filling, killed one
Aug 9 1999 filling, severe injury
A very interesting one in Aug 1994 Calgary Canada, steel, filling, overfill/disk, one killed one moderate. Sounds like a cave fill thing to me.

DOT and PSI most likely do not get info on third world ruptures, there should be more then just these!!
I'm trying to get 2002 to present data. not sure if they will release it to me as of yet.

That is my point. Real stats as opposed to opinions. Again, it does not show a high failure rate just some that have failed. Mass production will see this but to say the fill process is the only cause is a bit out of order, as is saying they will fail a lot. There will be failures but it does not seem to be an unsafe process or practice even with the stats you showed.

 

[oxyhacker]

This was an interesting case. I think there is a full report on it available on the net. Best I can recall, someone was filling a tank with the burst disk plugged on a compressor that someone had fiddled with the relief valve on to get higher fill pressures, and the operator got distracted, with predictable results. Hardly an indictment of steel 72s (any tank will blow if overpressurized high enough) or judicious overfilling. I don't think any of the overfillers on is list would recommend overiding all the safety devices then going out for a coffee while the compressor strains and groans! But a reminder that it takes two to tango when it comes to effective redundancy - you cannot disable your side of the safety system on the assumption that you are still protected because the other person will not disable theirs.

Oh, re steel 72 explosions, there was a fatal one in Florida not too long ago, where someone filled an out-of-VIP tank (I think it was out of hydro too), and it burst killing one. An examination of the tank showed that it was so terribly rusted inside that the must cursory visual would have condemned it. Once again hardly a condemnation of steel 72s or overfilling, but a reminder that saying "tanks don't really explode" can be fatal.

Did you look at the list I posted?
http://www.scubaboard.com/forums/4558800-post218.html

Nov 2001 filling, killed one
Aug 9 1999 filling, severe injury
A very interesting one in Aug 1994 Calgary Canada, steel, filling, overfill/disk, one killed one moderate. Sounds like a cave fill thing to me.

 

[WD8CDH]

Really? Funny, I've never heard of a single incidence of a steel 72 blowing up from being overfilled. Surely this would be news worthy as fill stations would be destroyed and people would be dying. Do you have any examples, any documented case of this happening? I'd be interested in that. I'm not denying your statement, I just never heard this before.

I have always assumed that steel 72s were over-engineered for their service pressure. This is why it's almost unheard of for them to fail hydro, despite the fact that every one still in service is decades old. I spoke to a hydro guy once who told me he has never seen one fail.

I think that in part, the service pressure of those tanks was influenced by the fact that lower pressure for compressed air was the norm in the 1950s and 60s, and also by the fact that many steel 72s were reputedly used as fire extinguishers, which operated at much lower pressures.

I would be very interested to know the wall thickness and other specific structural information on LP72s as opposed to current 3AA 2400 tanks.[/QUOTE]

Just check the tank stampings. If they are stamped 3AA, as most were, they were made to the same specs as the current 3AA 2400 tanks as far as safety factor, strength of material and stress at service pressure.

 

[adm3745]

Code of federal regulations:
Section


Overfilling properly maintained LP Steels is safe. We just don't do it on faith as many people seem to think.

"(2) For cylinders with service pressure of 900 psig or more the
minimum wall must be such that the wall stress at the minimum specified
test pressure may not exceed 67 percent of the minimum tensile strength
of the steel"

That among many makes me pretty comfortable about 3600 fills.

 

[adm3745]

This link is the main page so you can look at 3A and 3Al specs as well as some other info.
2004 CFR Title 49, Volume 2

It does appear that the Aluminum and HP tanks have a lower safety margin. But in terms of actual tank fills (only considering the safety margins) a 3600 LP fill is equivelant to a 4000 Al fill.
Of course we don't do that to Al's, so i'll be looking into the dozen other factors determining tank strength and life span.

 

[DA Aquamaster]

Samples from a lot of 3 AL tank have to be tested to 2.5 times the service pressure without bursting - 7500 psi in the case of a 3000 psi AL tank. But it is a one time test.

Aluminum tanks are however prone to fatigue issues that are not an issue with steel tanks, consequently samples have to be tested with 10,000 cycles to the test pressure, or 100,000 cycles to the service pressure.

Aluminum is also much softer and suceptible to damage in service and that increased potential for stress risers has to also be considered.

Consequently, overfilling AL tanks is not a real smart idea.

 

[DA Aquamaster]

There are patterns in the data. Prior to 1985, the big killer of steel tanks was corrosion (PSI came along in 1982, which is interesting). After 1985 corrosion falls off the chart in regard to corrosion as a known cause, but probably accounts for many of the ?'s in the "cause" column after that as fatigue is not an issue and if the comparatively few shops catering to cave divers with 3600 psi overfills were blowing up, we'd have heard about it.

In contrast, AL tanks, which started coming on the scene in significant numbers in the early 1970's begin experiencing cracking issues in the 1980's with a sprinkling of heat related failures throughout the 1973-2001 time period. The SLC crack issue led to visual plus inspections and to my knowledge no properly inspected Al tank has catastrophically failed since 2001.

Consequenbtly based on the known data:

1. Steel tanks fail due to rust
2. AL tanks fail due to cracks or heat damage, and
3. There have only been 60 scuba related failures in the 40 years between 1960 and 2001 from all causes - out of how many hundred thousand tanks in use and how many million fills?

 

[ctcdg]

you cannot give these guys the answer the want. It's like a religion with them. They believe!!!

Once again, Leadking got it right. After the inane replies to my post, I gave up. Some guys can’t seem to grasp the fact that this is a cumulative effect. You can take arsenic and you may not die. You can even do it multiple times, thus convincing yourself it is safe. However, at some point you hit the cumulative limit and you die. You may not even take as much this time as you did last time, but when you hit the limit, you still die.

One guy actually looked up the info in 49 CFR 178, but then wrongly decided that since there is a safety factor, overfilling must be OK (thus eliminating the safety factor!) Try running any tool at 100% duty cycle, and see what you get. That's what an inexperienced fool does. A craftsman respects his tools.

For the super-genius who asks, "are you sure you know what you're talking about?", I won't introduce myself again, anyone interested can go to my first post: Plus ratings and hydros Post#39

I'm astounded at all the people willing to put into writing their misinformed statements of "I think...", "I'm comfortable with...", "I consider..." Without any engineering data to back it up.

The real doosey was, "The reason testing is done with hydraulic rather than gas pressure is that you can't have a "catastrophic" failure, just a very sudden drop in pressure as the material gives. Doesn't even make much, if any, noise."

Ha! Tell that to the test operator who has to go home and change his underwear!! We had a cylinder blow inside a test jacket that startled me 150 ft away, inside the front office. It blew the bottom completely off the cylinder. Don't tell ME they don't make noise! Have you ever witnessed a burst test??? I've been to 10 different cylinder manufacturing facilities and over a hundred retest facilities in 20 different countries and worked with over 4000 test operators. (But then again, maybe I'm "not sure I know what I'm talking about".)

I am posting two photos of calibrated cylinders that have ruptured do to fatigue failure. These are examples of the "unnatural" failure that occurs with fatigue. It is not a ductile failure - it cannot be detected by visual inspection or hydrostatic test. When it goes, it goes - sudden and catastrophic.

I say "unnatural" because cylinders (other than composite) generally fail with a ductile bulging in the sidewall before the tensile failure causing burst. If you look carefully at the second photo, you don't see the characteristic bulging. Both of these cylinders burst BELOW test pressure! At manufacture, the burst pressure of 3A and 3AA cylinders cannot be less than 1.5 times test pressure. By cycling these cylinders at higher pressures, you LOSE that safety factor. There was no ductile bulging in the wall, just a sudden, catastrophic failure.

I say "catastrophic" because the cylinder burst before test pressure. No warning. A cylinder failing the hydrostatic test simply exceeds 10% permanent expansion - not catastrophic. Had either of these been pneumatic rather than hydrostatic bursts, they probably would have also been "catastrophic" in the sense of loss of life and/or limb.

Just because you "can" overpressurize a cylinder doesn't make it safe practice. Some of these guys just don't seem to get it. Each time you do it, you are losing some of the safety factor that was built into the cylinder. Cylinders are extremely safe - they are "over-designed". That's what KEEPS them safe. 3AA cylinders are designed for UNLIMITED service life - that means they live forever. We have cylinders that are over 100 years old. But when you overpressurize them, you are eating into that service life. It may not be you who pays. It may be someone 50 years from now who bought the cylinder at the auction of your estate, and 10 years later it ruptures. That may be meaningless for some of the younger people reading this post, but as for me, I don't want ANYONE'S blood on my hands.

I will not continue to reply to this thread. This simply isn't an issue of debate for me. One guy said that rules are for stupid people. Each time I get on an airplane I will thank God that that guy is not the pilot.


Darrell Garton
CTC Seminars

 

[adm3745]

3600 fills do not push 100%. It's closer to 50-55% Yield strength. 4400psi is only 66%. I have been studying mechanics for the last three years and see no reason for why that pressure (3600) is unsafe. I'd even be willing to take a a few tanks that guys have been overfilling for 10-20 years and test the metal with my professors. I think we all accept that service life is degraded through cyclic fatigue. The amount of cycles even at 4000psi that a tank could go through is in the thousands. My tanks will still last 50 years but I probably won't keep mine past 10 years because I don't know the exact effect the loading is having on my individual tank. Hopefully by then, we can get the capacities we want in a reasonable sized and priced tank. Then there will be no need for overfilling.

 

[Luis H]

3600 fills do not push 100%. It's closer to 50-55% Yield strength. 4400psi is only 66%. I have been studying mechanics for the last three years and see no reason for why that pressure (3600) is unsafe. I'd even be willing to take a a few tanks that guys have been overfilling for 10-20 years and test the metal with my professors. I think we all accept that service life is degraded through cyclic fatigue. The amount of cycles even at 4000psi that a tank could go through is in the thousands. My tanks will still last 50 years but I probably won't keep mine past 10 years because I don't know the exact effect the loading is having on my individual tank. Hopefully by then, we can get the capacities we want in a reasonable sized and priced tank. Then there will be no need for overfilling.

During hydro test a cylinder will experience some permanent deformation… guess why that happens.

The hydro test is design to take a cylinder to just below the very low limit of the material yield strength. Every time you exceed the yield strength you are work hardening steel alloys, which in turns it makes it less ductile (more brittle).

Is like bending a coat hanger permanently, you can do it a number of times before fracture, but it won’t take long for it to fracture.

Keep in mind that all the steel in a cylinder does not have exact yield strength. The numbers used in most calculation are based on the lowest yield strength of the material. The actual properties are actually a range. That is why we get “certs” (material certifications) with every lot and that still specifies a range, a minimum and a maximum, also the minimum ultimate, ductility, etc.

The strength that you quoted above (in the section of CFR 49 that you quoted) is referring to the ultimate strength not the yield strength. I am sure you know the difference, but if you read the entire CFR you should have deduced this information.

As long as you stay below the yield strength, then all you have to deal with is with fatigue curves (S-N curves). The interesting thing is that (as I mentioned before) S-N curves are normally drawn in logarithmic scale. The number of cycles that scuba tanks have seen to this day is relatively small.

Just because it has not happen in 20 years is by no means proof that it will not happen sometime in the future.


BTW, I hope your professors are not reading this thread... I am not necessarily questioning your engineering knowledge...

Added:
If you are interested on pressure vessel calculations, I will be glad to send you some Excel spread sheets that dealing with the calculations for 49cfr178.37 and other stress calculations.