Rupture Discs, FYI:

[Norm]

I have never seen one go underwater either. I've had two to fail from fatigue and one to fail because it was grossly overfilled and stored too long. I have seen many 'used' ones that were dimpled. I have debated replacing them with stainless ones for years, and have even been advised to on other boards, but I wonder about liability repercussions in our society of litigation?
(Rick it was in a cool basement at 4:30 A.M)
Norm

 

[devjr]

Fred, that's an interesting point about the aluminum tank, fire, fill pressure and the burst disc. Fire may cause the tank wall to fail before the burst disc goes if the tank pressure falls in the mid range. Good material for some DI who is looking for a trick test question. Also interesting about the torque used for the plug. Sometimes those things need to be wrenched a bit to prevent leaks.

I'm not too worried about discs anymore. A while back, the disc in one of my 3000+ psi steel tanks blew after about 10 years service. Actually, the tank was in my Suburban, right behind the front seat and made quite a racket. Since I was driving to a dive shop anyway I let them replace the disc. Turned out that the original disc/plug which came with the tank was only rated for 2250 psi service, and it had lasted 10 YEARS through hot temperatures, salt water, etc. I checked my other two tanks, all identical and bought at the same time. One had a 2250 psi disc, the other tank's disc was 3000 psi.

 

[devjr]

Fred, one piece of trivia which you may find interesting; Scubapro mentions a torque setting of 90-110 inch pounds for tightening the disc plug.

>>Most specs only call for tightening the disk plug to 4 to 8 INCH-pounds to prohibit early failure of the disk.<<

 

[neil]

FredT,

Sherwood says 50-55 inch pounds. Maybe you meant 4 FOOT/pounds? And certainly not as high as 8 ft./lbs?

Neil

 

[sharpenu]

The Burst disk is not there just for firefighters. It is there to protect from cylinder failure due to temperature changes combined with overfilling. The disk on a 3000 psi tank will fail at 4500 to 5000 psi. A 2250 disk fails at 3250-3750psi.

The relationship of pressure change due to temperature is expressed by Amontons' Law. This is expressed by P1/T1=P2/T2 .
A cylinder filled to 3200psi at 80degF that is heated to 100degF (such as in a hot car) will have a final pressure of 3334psi. That same cylinder filled at 60degF, and heated to 100degF will hit 3461psi. That means a the disk on a tank filled to 3200psi at 80degF will fail at some time after the temperature rises to above 300degF. So during extremes of normal use, there is not a real chance of a disk letting go due to overpressure. Taht being said, however, they DO fail as they get older. The stress from filling/draining will eventually cause a failure.

Aluminum melts at 1150degF and will fail several hundred degrees less than this. At the point of flashover in a fire, the air is approximately 1150degF. At this point, all combustible materials in the room (even carbon monoxide in the air) become involved in fire. So we have exceeded the limitations of the metal. The disk bursting at around 600degF should prevent catastrophic failure of the tank. Any pressure below 2100psi will not ensure disk failure at temps below 600degF. THe only other answer is to keep tanks at less than 200psi. That way a failure won't be as hazardous.

 

[FredT]

Originally posted by sharpenu

Aluminum melts at 1150degF and will fail several hundred degrees less than this. At the point of flashover in a fire, the air is approximately 1150degF. At this point, all combustible materials in the room (even carbon monoxide in the air) become involved in fire. So we have exceeded the limitations of the metal. The disk bursting at around 600degF should prevent catastrophic failure of the tank. Any pressure below 2100psi will not ensure disk failure at temps below 600degF. THe only other answer is to keep tanks at less than 200psi. That way a failure won't be as hazardous.

You may want to refigure. The magic temperature for an aluminum tank in the 6000 alloy series is 350°F. At that temperature the aluminum will anneal to less than 1/4 it's previously rated strength.

Considering that pressure vessels are generally constructed with calculated burst points 3 to 4 times the _TEST_ pressure (not _working_ pressure) it makes sense that anannealed AL tank WILL fail at below the _working_ pressure! The nylon seal washer under the burst disk will melt and fail at between 400 and 500°F.

BTW The torque on the burst disk plug is necessary to simply prevent the disk from lifting off the seal washer. Overtorquing smashes the seal washer and makes it hard to remove, as well as "prefailing" the disk to some extent. Once the plastic disk is compressed increasing torque does little to "help" the seal. Doing a stress-strain deflectiion calulation on a TORQUE basis is generally useless, although it is a good indicator _if_ all threads are clean and new. Minor changes in thread condition make big changes in final bolt or plug tension/compression! As a rule of thumb, when tightenng a bolt stack with nylon or delrin in the gripped material roughly a 40° "turn of the nut" from the first frim seating of the fastener will bring you to the mazimim sustainable grip tension/compression. Beyond that you are simply moving plastic.

FT

 

[devjr]

You're both right. Until the 1960's SCUBA tanks were fitted with a disc, phenolic washer and lead plug which had a melting point around 600 deg. This was standard practice to protect firefighters. Overfilling was apparently not a concern since, at normal temps, the lead plug could withstand pressures close to tank burst pressure. The dive industry lobbied for a safety disc with a hollow plug, and so the lead plug disappeared.

The aluminum tank's tensile strength is due to heat treatment. This is compromised at around 400 deg ambient.

The relationship between temp and air pressure is not so simple as stated. Only applies to an "ideal" gas. Close enough for govt work.

 

[sharpenu]

That is why I said it will fail at several hundred deg below the melting point. There is no concrete temp for failure, because there are too many variables. The sheer number of SCUBA tanks here ensures that we, as a fire department, will see alot of them. I have seen many tanks involved in fires, both full and empty, yet have seen very few failures. I don't know if this is due to conditions or good engineering, but we do see quite a few failures of vessels with liquid/gas combos, such as propane.

And I DO know that Amonton's Law refers to an ideal gas. That is why it is a part of the Ideal Gas Law. Predicting the behavior of any real world material becomes diffcult due to the sheer number of variables. That goes WAY beyond my paltry understanding of math and physics.

 

[devjr]

Fred, no SCUBA tank is going to withstand 3-4 times the test pressure. A tank with 2450 psi working pressure is rated for about 6500 psi burst.

 

[sharpenu]

Hydro testing is done at three fifths of the working pressure and burst disks are set at that pressure. 3000psig tanks have a 5000psi (approx) hydro pressure