I can calculate the response of defects on the UT beam centre line using Ermolov, and I can calculate the response from a rough defect using Ogilvy. How do I calculate the response from a defect detected by the beam edges, ie a defect tilted slightly from the specular, but detected by the specular reflection from the beam edges ? Any ideas ? Don

09:56 Feb-04-2004 Paul A. Meyer R & D, GE Inspection Technologies, USA, Joined Nov 1998 ^{47}

Re: Calculation of UT signal response Hi Don, There are modeling programs available that calculate the response of a flaw anywhere in the sound field of a transducer. The flaw must be modeled as an array of points. The degree of difficulty will depend on the geometry of the flaw you wish to model. One such program is FIELD II available at www.es.oersted.dtu.dk/staff/jaj/field/. The program operates using MATLAB which you may already use. There are probably other programs but I find this one very useful. Let me know if you have further questions. Paul ----------- Start Original Message ----------- : I can calculate the response of defects on the UT beam centre line using Ermolov, and I can calculate the response from a rough defect using Ogilvy. How do I calculate the response from a defect detected by the beam edges, ie a defect tilted slightly from the specular, but detected by the specular reflection from the beam edges ? : Any ideas ? : Don : ------------ End Original Message ------------

07:19 Feb-07-2004 tu R & D Chian Academy of Railway Sciences, China, Joined Dec 2002 ^{5}

Re: Calculation of UT signal response Dear Don

What is Ermolov and Ogilvy?Could you introduce it ? thank you! tu

----------- Start Original Message ----------- : I can calculate the response of defects on the UT beam centre line using Ermolov, and I can calculate the response from a rough defect using Ogilvy. How do I calculate the response from a defect detected by the beam edges, ie a defect tilted slightly from the specular, but detected by the specular reflection from the beam edges ? : Any ideas ? : Don : ------------ End Original Message ------------

Re: Calculation of UT signal response Does Field II include shear waves, and mode conversion?

Regards, Pim van Andel

----------- Start Original Message ----------- : Hi Don, : There are modeling programs available that calculate the response of a flaw anywhere in the sound field of a transducer. The flaw must be modeled as an array of points. The degree of difficulty will depend on the geometry of the flaw you wish to model. : One such program is FIELD II available at www.es.oersted.dtu.dk/staff/jaj/field/. The program operates using MATLAB which you may already use. There are probably other programs but I find this one very useful. : Let me know if you have further questions. : Paul : : I can calculate the response of defects on the UT beam centre line using Ermolov, and I can calculate the response from a rough defect using Ogilvy. How do I calculate the response from a defect detected by the beam edges, ie a defect tilted slightly from the specular, but detected by the specular reflection from the beam edges ? : : Any ideas ? : : Don : : ------------ End Original Message ------------

Re: Calculation of UT signal response ----------- Start Original Message ----------- : I can calculate the response of defects on the UT beam centre line using Ermolov, and I can calculate the response from a rough defect using Ogilvy. How do I calculate the response from a defect detected by the beam edges, ie a defect tilted slightly from the specular, but detected by the specular reflection from the beam edges ? : Any ideas ? : Don

Don

Interesting topic.

Thoughts from TWI:

"There are several relevant computer codes that can handle problems of this nature. For conventional UT of ferritic steel, we recommend the suite of semi-analytical models developed, and extensively validated, by the UK nuclear industry. These models use realistic pulse shapes and beam shapes (based on a Kirchhoff approximation), and allow general (3D) probe/flaw orientations.

In particular, PEDGE allows the prediction of both reflected and diffracted echoes from a smooth planar flaw in a wide range of pulse-echo configurations, based on Keller's Geometrical Theory of Diffraction. PKIRCH is a complementary code, based on a Kirchhoff approximation to the pulse-echo response (like Ermolov's far-field approximations). CORKIRCH is a similar Kirchhoff model for corner echoes from a smooth planar flaw close to a flat back wall. TILGEN simulates rough cracks (of a realistic morphology) as a series of interlocking triangular facets and TRANGLE predicts their pulse-echo response (again based on Kirchhoff theory). I guess you are using Ogilvy to estimate the average signal from a population of cracks of a given roughness, assuming that they scatter energy equally into a hemisphere. In practice, our experience suggests that it is not unusual for the signal from an individual rough crack to differ by ±6dB from this average value. Kirchhoff theory is, of course, only valid up to certain limits on the misorientation angle - the above programs set warning flags if these limits are exceeded (the same principle applies to other, less serious limitations identified during the validation work). We regularly run these codes under licence for our clients; run times are typically less than a minute per probe position (on a modern PC).

Other codes based on Kirchhoff theory have been written that simulate focussed beams, layered/anisotropic materials etc. But, as far as we know, there is relatively little validation evidence in the public domain to support these models. We are validating the CIVA model, developed by CEA.

Finally, for flaws having a dimension smaller than ~2 wavelengths (a special case not covered by the above, high-frequency theories), TWI has developed its own finite element code, primarily for use with guided waves (long range UT), which is now being developed for conventional UT. Initial results are in good agreement with PEDGE."

Gary : ------------ End Original Message ------------