Report on pipe mill ultrasonics by E. Ginzel, M. Hoff, J. Mihell
This report shows some reasons why in an automated ultrasonic testing installation for welded tubes, severe defects (cracks) pass the inspection undetected.
The first question to this matter is: What are the physical and technical reasons for this failure? The second question follows: Why are existing specifications not detailed enough to exclude such failures?
Unfortunately undetected flaws always blame the method in general, i.e. here ultrasonic testing is seen to be unreliable. But this is not true, if the suited techniques and the suited probes are used.
As Ed Ginzel explained, the main problem is the full coverage of the total wall thickness by ultrasound and related to this, beam size and angle have to be chosen, so that the flaw to be detected has a strong interaction with the searching sound beam.
This immediately leads to the next step, to establish specifications where these properties are checked periodically at a reference weld containing suited artificial reflectors.
It is a big problem that the tool 'ultrasonic beam' is seen as a constant, but really only the probe generating and receiving sound waves is always the same. Whereas the shape of the ultrasonic beam depends on the properties of the material and on the direction (beam angle). Therefore it has to be measured under real testing conditions before running an automated inspection. And the operator of the automated installation must be able using test reflectors to position the probes so that the beam covers the expected area in the tube weld)
This is very important with thermo-mechanically rolled plates, which are anisotropic and where e.g. beam angles and probe positions cannot be calculated using Snells law.
Interaction of sound waves with defects is best for 45° waves on defects close to surface (ID and OD). Because on tubes beam angles are smaller on the inner surface than on the outer surface flaw echoes from inner defects usually show higher amplitude than from outside defects. Internal defects are seen best, when hit perpendicularly. This affords suited sound beams for the defects in the different zones.
So 70°-beams positioned to surface defects only, will not detect internal cracks. But 45°-beams on defects close to surface will give much better echos than 70°-beams, whereas 70° beams , if positioned to the center of the plate (when not using tandem-probes) are best for internal defects.
If using noise suppression, i.e. suppresing all small signals, a lot of defect indications can be overlooked using pulse echo method. That is one of the reasons, why time-of-flight diffraction (TOFD) detects more defects in welds than ordinary pulse echo angle beam testing: With TOFD sensitivity has to be so high that all signals above noise level are shown in the B-scan-image (and not lost). On the other hand TOFD uses wide sound beams at angles between 50° and 60°, and diffraction on cracks is nearly independent of beam angle, so all internal defects are detected with a very high reliability. But because of the dead zones TOFD is insensitive for defects close to surfaces.
So,as shown by Ed. Ginzel, results are excellent if the advantages of both methods are combined.
One of the reasons why this is not treated very well by the existing codes and specifications is, that caused by globalization of production and high pressure on production costs, the main goal of standardization today is harmonizing of the existing standards and not so much pushing them to a higher level (higher safety but at higher costs).
To check the coverage of the wall thickness by the sound beams used, reference welds have to contain artificial defects at different positions. To our opinion through holes are not enough to check this. Additional internal and external notches have to be used and if necessary internal reflector(s).
Ed. Ginzel mentioned the standard EN 10208-2, but this standard only points to the detailed standards for automated ultrasonic testing of tubes, i.e.:
EN 10246-7 seamless and welded tubes (except submerged arc welded) EN 10246-8 electric -resistance and induction welded tubes EN 10246-9 submerged arc welded tubes
All these standards use notch reflectors. EN 10246-8 and EN 10246-9 use internal and external notches on both sides of the weld seam edges, but only one through hole at center line.
To summarize, I want to underline Ed Ginzels demand for
- new targets to evaluate the coverage of the weld by the sound beam, - new targets for the midwall area, - detailed procedures to position probes for the best interaction with the defects to be detected.