Weird sounds during pneumatic test 2

[ElCidCampeador]

Hi, a friend of mine who works in a workshop has told me about a strange fact happened during a pneumatic test of a stainless steel "transition piece".

This transition piece is a short section of pipe (lenght about 1m) with 2 flanges ad the end, but at small end both section and flange are circular (O.D about 200mm) and at large end both flange and section are rectangular (about 500x700mm), something like this one below:

https://www.grainsupply.com/wp-content/uploads/2020/10/Square-to-Round-Transition..-225x300.jpg

Flanges at both sided were connected to other flanges with classic bolts+nuts.

What is really weird was that happened during the phase of pressurization of this piping: when pneumatic pressure overcome 0,5 bar, this piece started to "bang", making sounds of blows or something like that, with a slight deformation of the external wall. At 0,7 bar he decided not to continue and vent out the vessel. Please notice that vessel wall (3mm thickness) returns in a as welded original state, without permanent deformation.

Do you think it's normal? What causes these sounds? There were no air leaks but they were all scared of these banging sounds...

Thank you

 

[LittleInch]

That's a complex shape going from circular to square and especially being welded has a whole load of stresses and inbuilt strains that when pressurised are slowly being reversed until the part flexes into a different shape.

The stress concentration in those welds is large and I've no idea how you actually design rate such a strange shape. FEA is the only thing I can thing of.

Normal? No, but this is not a normal shaped element.

Just put it in a safe location and pressure it to what ever you think is a good number and see if it breaks?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[r6155]

1) Never use a pneumatic test on this design!
2) Several times I heard little sound during hydrostatic testing on the pressure vessel.

Regards

 

[EdStainless]

Those flatter sections will 'oil can' easily.
I agree, using air for a shape like this is iffy.
What pressure were they looking to go to?

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[ElCidCampeador]

I don't know real design pressure of these but final pneumatic test pressure should have been 0,8 bar or something like...He was very closed to pressure test but he decided to stop the test and not to risk.
Now he will cut all welds (N.B. later he found that weld seams had small defects), reweld and add some L-shape as reinforcement. Then he will try again pneumatic test...

Thinking about it, considering the small thickness and the change of external shape (from a circular curve to a plane) I suppose that weird sounds are "normal" because material is stretching due to internal pressure but my friend can't quantify how dangerous the situation is since he can't understand if material continues to work in elastic range...

 

[DriveMeNuts]

When you put a dent into a tin can or soft drink bottle and then pressurise, it will make banks as it buckles back into its original shape. This is not a safety issue.
But the sound may be annoying if the equipment is in cyclic service.
3mm does sound very thin for a square section though. Has the design had a detailed analysis? Fatigue, ratcheting etc.

 

[r6155]

Mechanical stress relief serves to cause small flaws, particularly in the weld zone, to yield plastically at the flaw tip resulting in a local relief of stress and a blunting of the crack tip. To achieve the maximum benefit from mechanical stress relief, it is necessary that the stresses so imposed be more severe than those expected in normal service life. At the same time, it is necessary that the stresses which are imposed are not so high as to result in appreciable deformation or general yielding.
The mechanical stress relief shall be carried out using water as the pressurizing medium.

This is the cause of a sound.

Regards

 

[DriveMeNuts]

R6155,
That is the 'oops I've stuffed up, how do I get rid of buckle popping sounds' solution.
And sometimes the pressure required to permanently buckle the popping dent outwards causes so much plastic deformation that the equipment is no longer fit for use.
This is particularly applicable to thin walls, as per this example. Sounds to me like the transition is too thin.

 

[LittleInch]

Why in gods name is anyone pneumatically testing something when you don't know the design pressure??

This thing looks like it doesn't see pressure per se and is a very complex thing to calculate such a pressure. Square tubes are not good at holding pressure more than a few inches of water pressure, especially at 3mm thick and some pretty dodgy looking welding.

I think your friend has been very lucky not to injure someone during this testing. I would strongly advise him not to do this again.

Even 0.5 bar can result in explosive failure and shards of metal flying about.

Chalk this one down to a near miss and LEARN THE LESSON.

IMHO

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[DriveMeNuts]

Perhaps it is a good old design by experiment project.
Such a small volume pneumatically pressurised to 1 bar isn't particularly hazardous while standing say five metres away.

 

[FacEngrPE]

Design by experiment. ASME allows this for some cases, the test derived operating pressure is some percentage of burst pressure. However even 1/2 bar can be a dangerous pneumatic test. ME PCC-2 (or the safe distance calculator). The safe distance for this test is 30m, dependino on what "safe" means to you.

 

[DriveMeNuts]

30 metres is far in excess of a safe distance. Half a metre cubed of air doubling in size doesn't warrant 30 metres of distance. The 30 metres in PCC-2 is a rule of thumb lowest possible distance. If you chose to read the website, then you would realise that this rule requires that an inch pipe, an inch long being pressurised to 0.001 bar, would require a pneumatic test distance of 30 metres. I assume that FacEngrPE is quoting a number he read on the interweb somewhere and hasn't actually employed any engineering judgement to assessing if the number is reasonable.
There is a second calculation in the same chapter of PCC-2 (and on the website) which has an actual energy based equation which reflects the real world and specifies 6 metres for half a cubic metre pneumatically pressurised to 1 bar.
With the low pressure, the equipment is not within the scope of ASME. Engineering Judgement with the aid of code rules is the appropriate approach for the development of this equipment. There is no need to bureaucratically follow the rules in this case......in my opinion.

 

[r6155]

It is sometimes necessary to consider a pneumatic test if it is not practical to perform a hydrostatic test.

Why not hydrostatic test?

Regards

 

[LittleInch]

It's really about the biggest bit of material that can be ejected at high velocity or break off, not the expansion of the air volume.

My point remains. There is NO indication that anyone has actually established a design pressure for this thin metal complex shape or a test pressure which doesn't exceed the UTS or the weld strength.

Air testing this component looks like a poor decision to me based on what we know.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[The Obturator]

Looks like a square to round transition piece of ducting used for dust extraction/LEV - looking at the website address of the photo '...grainsupply..." I may be right - for a grain silo? In which case the working pressure is generally below atmospheric.


*** Per ISO-4126, the generic term
'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***

 

[LittleInch]

Or a grain connection from a hopper to a round pipe?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[ElCidCampeador]

Photo I posted has nothing to do with this real case, I've chosen from web only for example to let me show the shape...

Transition piece of my friend is part of an air heater system, this is what he told me

 

[LittleInch]

But it's still a circular to square duct transition piece no?

All my comments still apply if this complex shape is being tested.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[ElCidCampeador]

But it's still a circular to square duct transition piece no?

Yes

 

[DriveMeNuts]

PCC-2 does have a section to calculate shrapnel ejection distance. This is only applicable if the user assess there to be a risk of shrapnel.
I would assess there is no risk. This transition would only tear open or distort the flanges into leaking, without ejecting material.
For a pneumatic test to destruction where you are certain that failure will occur under 2 bar, a 20 metre distance with a shield is more than adaquate. While not ideal, I don't have any issues with pneumatic testing for this small size as long as there is adaquate distance.