Steel Tank Condemned: Cracks in Threads?

[rjack321]

The machine marks look straighter

they usually are but cracks propagate from the inside of the tank out, so the absence of any cracking on the hills of the threads is diagnostic of tool marks, which are most common in steel tanks

 

[Diving Dubai]

I'll defend the inspector as making a good call. If in doubt err on the side of caution

Now I know I'll be upsetting all those who've sat their Vis courses, on the other hand I'm an ex NDT Engineer, who for 15 years, not only carried out inspections, but developed new procedures, Taught all the courses (Basically and IT) and helped develop wild new methods which are still in use today.

In this case we have a photo whereas the inspector had the ability to look around and move the camera. And at teh end of the day, vis is just a basic and rudimentary inspection method. Sorry

While people say it can't be a crack because you can't see it on the top of the threads, there's nothing to say it's not there, OR has travelled in a straight path root to root through the thread. You might disagree, I will say to that, I've seen defect which apparently can't be there according to the models and the maths, meaning they revised the models and the maths

Even with my eye, I can flip flop between crack or not crack and can think of a number of feasible hypothesis - none provable from a static picture

As for methods to try to prove

Those suggesting Penetrant or Magnetic Particle (Magna flux) NO! Risk with MPIU and a certainty with Penetrant, that the oil based dye will contaminate the defect and you'll never really get it out if it's just a mechanical defect.

Eddy Current would be the best, but unlikely anyone had a suitable probe (the mickey mouse Eddy current tester sold to shops doesn't count - even if the operators have the knowledge and skill)

Radiography (X-Ray) certainly - if you found somewhere with an appropriate size machine for "thin" steel.

Ults for sure using shear wave - although I'd be tempted to go for and internal surface wave, but here like Eddy Current you need a really experienced operator to interpret teh results

Not seeing a defect doesn't mean it's not there and havign a defect indication equally doesn't make it a defect (hence people like me wrote carefully crafted procedures that had near zero interpretation required after lots of R&D)



For the OP the sensible option is to take teh offer of a trade cylinder. Certainly its a curiosity warranting further investigation.

The impression I see is people are too willing to dismiss the idea of a crack in steel - Expectation of results and a dismissal of results that don't meet that expectation is a very real problem.

See Aloha Airlines Flight 243 as textbook example .53 fatigue cracks found post crash in an area than doesn't normally exhibit defects.

It's never about the size of defect you find, but always about the size of defect you miss or dismiss

 

[Doc Harry]

It is HIGHLY unlikely that there was actually a crack in the threads...

I'm a cylinder inspector.
It is HIGHLY unlikely that there was actually a crack in the threads. Get another opinion. Probably just a tool mark.

 

[AfterDark]

What about taking it directly to a hydro facility?

I had looked at Eddy current looks like, per wiki, they use it for different steels.

Could you xray? It would show the fault in the material but at what cost?

If this were a submarine part it would be X-rayed and either wet magnetic particle checked or liquid penetrant checked. The X-ray would readily detect a deep crack, liquid MP would detect a surface crack. Being steel the florescent mag particle testing and X-ray is what Electric Boat would use.

I vote tooling marks for the reason that the crack is not contiguous across the threads.

 

[richboslice]

I have seen videos of "condemned" tanks that merely had a very fine thread of metal left from the valve thread. It is very common in mechanic work to find a very fine wire on threads that could be mistaken as a crack.

 

[richboslice]

Possibly have Eddy current done?

 

[Diving Dubai]

NDT 101 coming up:

If this were a submarine part it would be X-rayed and either wet magnetic particle checked or liquid penetrant checked. The X-ray would readily detect a deep crack, liquid MP would detect a surface crack. Being steel the florescent mag particle testing and X-ray is what Electric Boat would use.

You'd neither use Penetrant nor Mag Particle on this tanks since the defect would be contaminated with oil based products which would be very difficult to remove

Penetrant as you say detects surface breaking defects, Mag Particle both for surface and sub-surface

Depth of defect has no impact on X-Ray, you're looking for changes of relative density. The ability to resolve an image is controlled by a number of factors - but say its a very tight defect in a single wall but you have two thick tanks walls to shoot through you probably won't see it. Resolution is defined by material thickness vs defect size, type of film (or now sensor) beam angle of X-Ray tube, KVa and MA settings to name a few.

You'll use an IQI (image Quality Indicator) to prove your shot

Possibly have Eddy current done?

Requires a reference block of the same material having had the same heat treatments and (if a galvanised tank) to have same protective treatments - basically have the same conductivity characteristics (with regard to the Eddy current field)

Reference block needs to have a simulated mechanical defect as well as a simulated crack. Also need to know the optimum probe frequency for that material, even the probe coil size will have an impact on defect resolution

Eddy current on steel is possible IF you're a skilled experienced operator - but unless you've got a reference block you have no way to determine if you signal is a defect, or scratch in surface treatment or maybe a Fe anomaly .


NDT is not magic, it's physics and you do need to have a reference of the type of defect to prove this is what you're actually seeing and not either getting "false positives" or not detecting the defect at all.

Operators just don't NDT something, they follow a procedure which is specifically written for that type of defect in that type of material or component (Penetrant and Mag Particle somewhat less so) or something close to it.

I've spent many weeks - even months developing and testing procedures so that appropriately trained military and civilian operators can repeatedly carry out an inspection with the results being clear and requiring almost no interpretation

 

[rjack321]

I'll defend the inspector as making a good call. If in doubt err on the side of caution

Same

I happen to think these are tool marks, but their jagged nature is still worrisome and there is no shame in condemning this tank. Compare to having it explode, kill a fill station operator and do $50,000 worth of damage to a compressor, shop wall and inventory. That would be the shameful potential outcome.

 

[Diving Dubai]

Same

I happen to think these are tool marks, but their jagged nature is still worrisome and there is no shame in condemning this tank.

Yep. Especially as it's HIS signature on the inspection sheet.

On occasions you do get cases which divide opinion. The only way to be conclusive (not in this case since it's a low value item) it to test using another method

On really rare occasions, I've had two or 3, different tests give different conclusions. One case in particular sticks in my mind because after a week of throwing all the different test methods at it, we were still 50/50.

The component was condemned and when we broke it open we indeed found a defect. Unfortunately it was of a type that had no place being there. Lots of smart people spent lots of late nights trying to figure out why it was there.

That rabbit hole ended with 18 months of urgent fleet wide inspections, and some new pages being added to text books

That's one of the reasons I'm always dubious when people are quick to declare it can't possibly be a defect.

You just never know.

 

[OTF]

Regarding penetrant indicators contaminating the defect - wouldn't that only happen if indeed the defect is a crack and not just a surface mark, in which case the the tank is garbage anyway?