Welding, hydrotesting a compressor tank

[aametalmaster]

My dad was a racecar driver when i was a kid. He always carried one of those pancake air tanks on the trailer. One day he filled it up and stopped to BS with a guy and sat the tank down. I went on ahead to the house, i was about 15. A few minutes later he picked up the tank and it blew removing flesh from his right leg. The tank had blown open from top to bottom and looked like a big pillow. He spent the next week in the hospital. That had to hurt...Bob

[usmcruz]

I was taught on the procedure for welding pressure vessels, so if you havent been taught this, dont attempt it. If you were going to attempt it, and theirs nothing we can do to stop you, go ahead, and die test the cracked area, and see where the crack start and ends, then if the crack is a perfect straight line, drill out the start and stop of the hairline crack with a 1/8" drill bit, to stop the crack from spreading when you weld it. Then carefully bevel out the crack on both sides of the crack, leaving a 3/32" to 1/8" gap in the middle with the same sized land. Meaning a 3/32' gap, should get a 3/32" land, etc. Then either 6010 the root in or tig the root in. Make sure you ask the manufacturer what grade metal they used for the tank, not letting them know what your doing, otherwise they wont give you the info. Run a root pass, hot pass, and cap, and if done correctly, should leave the weld slightly sticking above the base metal. Then die test again, then hydro. This is a very delicate procedure, and not to be done by the hobby welder, by any means. I would get another tank, but I wanted you to know the correct way of doing it, if you did try to, just so that you wouldnt get hurt or killed.

[aametalmaster]

And don't forget the "R" repair Stamp on the ASME tag...Bob

[Portable Welder]

Aametalmaster, Thats unfortunate what happened to your Dad.

I know when I fill up my portable tank that is rated for 125 psi. I'm always care full not to fill it to the 175 psi that my big compressor runs at.

On my plazma cutter I have one of those toilet paper water seperators that says max 125 Psi.
I made sure to put a regulator on that, It looks like its made of cast aluminum.

On the train it appears the blow off valve froze shut and over pressurized.

[Bodybagger]

If there are fatigue cracks you can see with the naked eye, imagine how many lurk under the surface that can only be detected by magnafluxing.

Sort of like the "see one cockroach, how many are inside the wall" problem.

Now imagine a crack leading in the direction around the "waist" of the tank. Let's say the tank is 30" in diameter.

Force=pressure X area. The cross sectional area is pi*15^2= 707sq in. At 120psi, there is a force of almost 85 thousand pounds trying to blow the tank in half across this section.

If that sounds like a lot, look at the stress on a lengthwise crack...
If the tank is 5' long, the cross sectional area is about 60x30=1800sq in (neglecting domed top/bottom). At 120psi, the force trying to rip the tank apart is 216 thousand pounds.

Yes, when these things fail, it's most commonly a crack that opens slowly enough to let the air out. But it's that failure mode where the crack seeds on an existing fatigue failure plane that allows them to let go suddenly, and with devastating results. This is not a lottery anybody wants to win!

Notice that when they go destructively, they blow out from the side along a longitudinal split (as in the locomotive picture), due to the fact that this plane carries the higher strain.

[Craig in Denver]

I was in a guy's garage (friend of a friend), leaning back against his bench, looking across the garage (open trusses). A light in my head went on and I said, "Hey, it looks like a can of paint exploded in here?" He pointed down behind my knees, under the bench, at a brand new air compressor. He said there was a gallon of paint on the floor, where I was standing, when the compressor tank blew. Luckily, his vehicles were out and the garage was uninhabited when it happened.

[DSW]

Working at the compressor shop I saw split/cracked compressor tanks on a regular basis. None were catistrophic failures luckliy. Most were small pits or cracks that leaked. Considdering the high volume of air screw machines produce, I could easily see a castistrophic/ large split if you had a OPV failure and/or reducing reg / hp limit switch failure.

At the dive shop, I have personally seen no fewer than 4 alum 3000psi- 3500psi tanks split necks. All have been small cracks that leaked. One split just after being refilled, having just come back from passing hydro. We would frequently find neck cracks on alum tanks on eddy current inspections and fail tanks prior to the tank actually leaking.

I have only personally seen the aftermath of 1 catastophic HP tank failure. That failure was in an O2 tank and there was 1 fatality and several injuries. I had a pict of the tank but can not find it right now. Again, most of the ones I have heard of happen either at filling or due to elevated temps such as leaving them in car trunks/sun etc. Again, most likely cause is burst disk failure.

[ichudov]

A follow up from the original poster of the question.

I have been able to find a sensible deal on a good condition compressed air tank. I will throw my tank away or reuse it for pig roasting (have not decided). In any case, I drilled a hole in the tank already to warn any possible future user that the tank is not suitable for compressed air.

if I actually throw it away, I will also weld a word "BAD" on it so as not to leave any doubt or room for unscrupulous sellers who may get it from garbage.

So, the bottom line is that I will not use a tank. If chance of its explosion is only 1%, I do not want it.

[engnerdan]

Here is another picture just for fun.


-Dan

[jweller]

IMO, it sure would be great to have a 80gal tank to build a real nice BBQ smoker out of. You could have worse problems