It is not common practice to UT through paint, however, that does not mean it cannot be done. First of all, what kind of welds are you examining? Butt welds? I do not have AWS in front of me, but you say AWS requires bare metal, so I cannot speak for AWS. I believe the intent of that is to ensure sound transmission is not affected by extranneous material that will inhibit the transmission of the ultrasound energy. This is important for evaluation, since AWS has an amplitude based acceptance/rejection criteria.
Now from a purely technical standpoint, however, my experince, has shown me that UT through coatings can be successfully accomplished by a couple of different methods. First of all, the coating must be tightly adhered to the material being inspected. Loose coatings "will" inhibit the transmission of the sound.
One way is by calibrating on the basic calibration block and performing a dB difference check (commonly referred to as a transfer technique) to compensate for the attenuation from the coating material. This attenuation can be anywhere from 2dB to 10 or even 12dB, depending on the material. This will allow you to compare the amplitude of indications with reference reflectors with reasonable reliablitiy.
You can use 2 angle beam search units either with a thru transmisson mode or pulse echo mode and a blank wedge of the same nominal angle on the calibration block. Peak your signal and set it at 80% FSH on the basic calibration block.
Then go to the part and do the same thing. Note the difference in dB. This is you dB difference (or transfer). This difference will be added or subtracted, as applicable, from your primary sensitivity level for evaluating indications. Accurate dB measurements must be taken in various areas and documented on the report.
The other way is to simply coat the calibration block with the same coating material and application process as the material being inspected. This is probably not practical, however. You are probably using an IIW Block for caibration, so I would lean towards option 1.
If cost of removing the coating is a concern and your "client permits" the use of transfer technique, I see no problem using that method.
I have done it and, if properly performed, works well. As long as your client accepts it. The coating MUST be tightly adhered, though.
By -
Date 06-14-2002 04:25
Actually, D1.1 does specifically address UT through paint in paragraph 6.20.2. There it is identified as a procedural "variation" that may be agreed to by the Engineer and must "be recorded in the contract records".
Assuming that this variation is approved, NDTIII correctly identifies two methods you can use to offset the effects of the coated surface. API RP-2X details a more accurate method of transfer correction, though it is often cumbersome due to the screen ranges required. I have found that by using the DAC curves that I've constructed for the joints to be tested (presuming you are using a basic sensitivity block/DAC type of calibration), and by comparing a one-skip-distance signal from the cal block and the work piece to the curves, that the results are somewhat more accurate than the method described by NDTIII.
I have also found that depending on the steel types involved, that attenuation and angle of transmission will vary in a nonlinear manner between the various ultrasonic angles and in different directions (direction of rolling vs. transverse to the direction of rolling). You may need to do transfer correction with all angles and in both directions to get a complete picture.
