Emptying tanks??

[Luis H]

Simplified is good, and correct. Of course, the gasket under the copper disc will fail almost immediately. Same for the nylon valve seat.
Pesky

All the non-metallic components will degrade at relatively lower temperatures, but they do not provide a flow path large enough for quick depressurization. A leaky gasket will not be sufficient in an intense fire.


Simplified calculations implies an approximation... that it is correct, well I have made mistakes in the past and I will guarantee you I will make some in the future. I just hope to catch before anyone else does.

 

[pescador775]

Not a quick enough flow path? Well, let's see;
assuming 0.010 thickness and 0.310 dia, the leaking gasket will have a flow area of 0.01 sq ins. Valve ports are 0.105-0.125 dia. The blown valve plus gasket will have a area of 0.0186-0.022 sq ins. At 3000 psi the calculated flow is about 130-150 cfm.

I had some trouble with the disc gasket calc. The area of the gasket suggests that it could release air at 100 cfm but there are too many parts in there, the disc would probably just blow back against the plug, so, I arbitrarily used a number of 20 cfm. That would be leakage around the threads, mainly. I calculated all the flows based on the Hacker formula, sort of. The formula does not seem to work at sonic flow rates so I substituted "radius" for "diameter" and referenced the result to a known flow, the Scubapro MK25 regulator and that looked very close. This gave a valve flow of about 125 cfm. This is a simpler situation than the burst disc gasket so I feel that this is a fairly good number for a melted valve seat.

All the non-metallic components will degrade at relatively lower temperatures, but they do not provide a flow path large enough for quick depressurization. A leaky gasket will not be sufficient in an intense fire.

 

[NorthWoodsDiver]

I store my tanks full usually, though storage peroids range from hours to weeks usually and maybe in the winter a month or 2 depending on which tanks I use for ice diving. Heck I even have some steel 72's (low pressure) sitting full at about 3100 psi, they have been sitting full since october I think.

 

[Luis H]

Not a quick enough flow path? Well, let's see;
assuming 0.010 thickness and 0.310 dia, the leaking gasket will have a flow area of 0.01 sq ins. Valve ports are 0.105-0.125 dia. The blown valve plus gasket will have a area of 0.0186-0.022 sq ins. At 3000 psi the calculated flow is about 130-150 cfm.

I had some trouble with the disc gasket calc. The area of the gasket suggests that it could release air at 100 cfm but there are too many parts in there, the disc would probably just blow back against the plug, so, I arbitrarily used a number of 20 cfm. That would be leakage around the threads, mainly. I calculated all the flows based on the Hacker formula, sort of. The formula does not seem to work at sonic flow rates so I substituted "radius" for "diameter" and referenced the result to a known flow, the Scubapro MK25 regulator and that looked very close. This gave a valve flow of about 125 cfm. This is a simpler situation than the burst disc gasket so I feel that this is a fairly good number for a melted valve seat.

Hi Pescador

I was going to ask you how you did the flow calculations, considering that it would be a fully developed supersonic (compressible) flow. There would be a fully developed static shock wave with a very large density change (from around 15.3 lb/cu ft to around 0.075 lb /cu ft).

To fully calculate the time required to empty the tank you will have to set up some differential equations to account for the pressure change as a function of time.

I haven't done this particular set of calculations since college, but if I am really bored maybe I will later, my compressible flow books are at work, so it would be much later.


BTW, considering that it takes well over a minute to empty a full tank with a fully blown burst disc (with a flow area several times larger than you calculated) it seems that your assumptions as still off.

Your damaged seal flow area assumption does seem large, but that is probably not the worst thing.

The flow rate published by Scubapro for the Mk-25 is probably only attainable with a modified valve and of course only with a high constant input pressure.

 

[pescador775]

I looked at the equations to do the calc and I just didn't want to put in the time and trouble for this hypothetical that probably doesn't settle anything. I think I told the story about driving to the dive shop in VA Beach; on the way, a burst disc in a tank situated behind me blew. I didn't time the dump but it was not very long. The majority of the air seemed to vent in about a minute. Icing of the valve could be a factor which one should not expect in a fire.

For the calc, I used 3000 psi as the input pressure but, naturally, as the air in the tank vents this pressure will drop and the cfm number will drop as well. To estimate total vent time, one could insert half the starting pressure (1500 psi} to get a number for the time to drain the majority of contents. At least, this estimate might give better results until pressure drops really low and the character of the flow changes.

The fire thing is really about aluminum tanks. Will the tank release its pressure before it explodes or not? It depends on starting pressure, temperature and time. We are told that "half full" is bad. Apparently, 6061-T6 can withstand a temp of 350F for a certain period of time before the character of the metal changes dramatically. A temp of 400F is definitely bad and after a few minutes will damage the temper permanently. The melting point of nylon 6 is 390F. However, one could expect a valve seat to fail at somewhat lower temps , let's estimate 350F. So, we might observe that this is "too close to call". We don't know what will fail first, the aluminum tank which was originally half full, or the nylon components in the valve. However, we are pretty sure that one of these will give way before the overpressure disc blows. Without the benefit of actual tests we will never know. It goes back to your statement about the character of the fire. Tests would have to range from "slow burn" to "instant inferno". My bet is that an instant inferno would blow the tank.

 

[rjack321]

Anyone still using 6351 AL tanks would be wise to store them near empty to avoid neck cracking. Otherwise I would't go out of my way to drain anything, fire or no fire