Air Testing of Roof

[AK8824]

Air Testing of roof is required as per the following clause in API 650:

F.4.4 When the entire tank is completed, it shall be filled with water to the top angle or the design liquid level, and the design internal air pressure shall be applied to the enclosed space above the water level and held for 15 minutes. The air pressure shall then be reduced to one-half the design pressure, and all welded joints above the liquid level shall be checked for leaks by means of a soap film, linseed oil, or another suitable material. Tank vents shall be tested during or after this test.

I have tanks with design pressure about 1.5 kPa(g). The contractor wants wants to fill the tank upto the apex of the cone roof and use a stand pipe and fill it to exert the required pressure.

API does not say that Air Test or a hydrotest may be used.

Can somebody advise me if hydrotest can be used to substitute Air Test?

 

[JStephen]

I think that would be an equivalent or more rigorous test. However, the tank has to be designed for it. The section you're referencing is for tanks with fairly low pressures, and filling the roof plus a standpipe may stress lower parts of the roof, including the top angle, to a considerably higher pressure than the design pressure.

I believe API-650 allows this type of test, provided the tank is designed for it.

 

[Duwe6]

"I think that would be an equivalent or more rigorous test"
If you can touch all the seams to feel for 'weeps & seeps', I would call it slightly more rigorous, but if you just stand and look at it from a distance, that would not be as revealing a test as soap-bubble.

Also, keep in mind that a standpipe 2-ft taller than the roof peak is what is used to 'pop' back up a sucked-in roof on a 650 tank. 24" W.G. is 14 ounces/inch^2 is 0.866 psi. Make sure that your roof will not fail at this pressure. 650 is for Atmospheric storage tanks.

 

[fegenbush]

Duwe6,

The air testing of the roof clause is in Annex F, which is for tanks designed with small internal pressures up to 2.5 psig. Your concern still holds, though as any tank other than one that is very small is not economical to construct with that high a design pressure.

 

[AK8824]

Thank you for your replies.

I am sorry, I should have converted 1.45 kPa to water column. Well, the stand-pipe would be about 150 mm tall. My feeling is that testing with water may not be able to reveal the leak for a gas-tight roof. But, in the past I have heard from some fabricators they replaced Air Test with Water-fill test. I am not convinced.

 

[Roberto1956]

Much better than positive air test is to do negative air test. You will do it outside the roof tank, the negative pressure is higher than a positive air test, it is safe and so on. The parts impossible to do negative air test I suggest you to do capilarity test. I kick procedure would be to wet with dye penetrant inside the tank parts as one wet for each hour and repeat it 5 times. 24 hours after the first layer applied you develop outside the tank with the white developer of dye penetrant kit. For your safe work, try do not use the can that use propane as propelent gas. I prefer the large quantities applying with a brush. A tip for you is to do it on everyplaces that received any weld or any welding point, welding risk. I ever found defects on stack weld, electrode risk and never on the welding seam.

 

[JStephen]

Note that the standpipe may be 150mm, but due to the roof slope, the center of the roof may be several feet higher than the shell, so filling to reach the center will give much larger pressures at the shell. For a small tank, you can design the tank for it and test it that way; for a large tank, it may be impractical.

So far as detecting leaks with gas vs liquid, keep in mind that the shell itself is tested with liquid, never with gas, and roof seams would generally be considered less critical than shell seams.

 

[AK8824]

JStephen,

Thank you for your advice.

I agree with you. The tank in my case is to be blanketed using fuel gas. Therefore, we must make sure there is no leakage whatsoever.

Ashok