UT inspection (flaw detection) on a 6" round x 18" long Bar.

[noviceEng]

Hi,
I had a 6" round Bar UT inspected for cracks and the results came back as "flawless" and acceptable for use. Yet after machining off 1/2" from the OD using a manual lathe, we found numerous hair line cracks. When we confronted the inspection company they implied that the test was done accurately and that the machining process could have caused the cracks..... (rubbish... i know) Anyway what I don't understand is how come the UT inspection didn't pick up the cracks that were obviously there all along. Is it that the UT wasn't done properly or am I missing something?

 

[HouseHark]

Should be easy enough to find, especially in such a short piece.

Bear in mind it is quite easy to miss something with UT.

Calibration may have been done wrong and the technician may be more familiar with testing something else like welds.

Problem is sensitivity as if calibrated on a 3mm side drilled hole on a 40mm WT block as common for weld inspection then the set wont be set up perfect for the inspection.

If no calibration block available I would have pumped the DB's up and scanned every direction, any indication is reject.

A better excuse by inspection company would have been you didn't clearly state spec/procedure and we used the wrong calibration block as you didnt prodcue one. I always ask for a prodcued calibration block that gives me a little out.


Also bear in mind that item should be scanned longitundinal and transverse.

 

[noviceEng]

The mat. is EN19 Chrome with a BHN of 252, and the test had to be performed according to ASTM E587 so they had to inspection criteria required.

I was convinced that the UT test was questionable, so I had it done again using a different inspection company and guess what... they got the same results stating that the bar was O.K.
This issue can really make someone question the validity of UT inspections. I'll definitely being performing LP inspections before & after from now on.

 

[JR97]

noviceeng,

Interesting that both NDE houses missed the cracks. What is the orientation of the visible cracks, i.e., primary length axial or circumferential? Is it correct to assume they appear to be propgating radially?

The ASTM spec imposed only angle beam testing. Do you know if both circumferential and axial scans were performed? Also, as HouseHark said the calibration sensitivity is important. The ASTM Spec does not specify a calibration reflector. In paragraph 6.1.4 it states "Reference reflectors of known dimensions, artificial reflectors, or distance-amplitude relationships of known reflector sized for a particular search unit and material MAY BE used." It does not specify and particular shape of size calibration reflector. Do the reports from the NDE houses mention any calibration or reference standards?

JR97

 

[redpicker]

To understand why imperfections were not detected in an inspection, you need to look at the inspection method and the calibration standard.

E587 is an angle beam standard, generally using an OD notch as a calibration standard. The reflector is the corner the notch surface makes to the OD of the bar. While this inspection is very sensitive to surface laps and seams, sub-sruface imperfectons might not be detected using this method. Sub-surface imperfections would have to have a favorable orientation to the transducer to produce an echo of sufficient strength to show up as an indication. That might happen, but it might not.

Also, a minor nit-pick. I doubt the inpsection report described the material as "flawless" and acceptable for use. More likely, they reported they inspected the material per the standard and found no indications. This is apparently accurate, as you have had verified by another inspection company.

Finally, LP testing won't solve your problems if, as I suspect, these imperfections are sub-surface. Instead of putting the inpsection company on the defensive by implying the inspection was not done properly, as for their help. If the inspection performed is not capable of detecting these defects, ask them (or find another Level III inspector) what inspection method would they suggest to detect these defects. You can also ask your material supplier what you need to do to get material that does not have cracks in it.

rp

 

[HouseHark]

Should have bought better quality bar. Any bar being used for machining should be well tested at the mill, automatically which reduces room for error.

It is interesting both inspectors missed it (unless they are mates haha) but also very true UT is not the best method in the world, it finds large defects in general testing, lab conditions are different.

If the second company new the crack was there it should have been very easy to find though, pump up the DB's and scan for anything not close to back wall reflection?

 

[WBH]

NoviceEng

I have been there and done that, the problem is the UT instrument is calibrated using a block that has a machined notch or side drilled holes, these reflectors are significantly larger than the cracks.

Take the bar with know cracks use this as a calibation standard if cracks are detectable by UT proceed with UT inspections. I had the same problem with forging burst in 8" rounds.

It is important to scan 100%, around the circumference and from each end with a straight beam.

 

[smurali1]

Hello,

I do not have much experience with UT. But I have heard UT technicians talk about deadzone. i.e. they say UT does not detect any cracks in the first 15 mm from the surface where UT probe is kept.

These undetected cracks can show up after machining.

Thanks,

SPM

 

[HouseHark]

dead zone shouldnt matter as you get 100% coverage from over side.

dead zone matters for testing very thin things and for considering scanning paths to ensure 100% coverage.

 

[rooskie]

Allow me to chip in here, if I may:
As background, I've done a bit of UT and been in the field a few years, but where I don't know something, I will bow my head and ask instead of trying to BS my way through. With that in mind: What is the chrome content of EN19 chrome material? What we call '9 chrome' in the refinery field is a heat-sensitive material when it comes to welding processes and how much heat is introduced and preheat,interpass and postheat requirements to avoid cracking are fairly stringent.
If you are finding cracks at 0.50", did you check at 0.25"? at 0.125"? In other words, were these surface connected cracks? Are the indications found after machining straight and well-pronounced or finer and not straight? In one instance, you may be looking at seams/stringers made during the bar-making process. In the other, the indications might indicate a heat-induced problem from the machining process.
What was the surface condition of the bar? If calibration was made on the bar itself, that's one thing. But if calibration was made on a cal std. and the surface of the bar to be inspected was as received from the bar mill, there's another opportunity for missed indications. It is an error in reporting and asking for trouble to declare a specimen 'flawless'. Better to state that it meets, in this case, the requirements of ASTM E578 if no reportable indications were noted. That's why you see the disclaimer at the bottom of any reputable NDE report. I also ascribe wholeheartedly to HouseHark's and redpicker's posts.