Parameters for this problem are as follows:
With these parameters, the electromagnetic skin depth of penetration is about 3 mm. The conditions for using SF-X-CRK.EXE (Flaw Length >= 4 x Flaw Depth, Flaw Depth <= Electromagnetic Skin Depth) for simulating planar flaws are NOT met in this problem. An accurate theoretical prediction of the experimentally measured ET signal is NOT expected. This problem is outside the model's "regime of validity".
Experimental Data was transcribed from here:
http://www.compumag.org/jsite/images/stories/TEAM/problem15.pdf
To simulate the signal from a pancake impedance coil using SF-X-CRK.EXE, the transmit and receive coils are given the same dimensions with a centre-to-centre separation of 1E-10 m (1 angstrom).
Results of the comparison of SF-X-CRK.EXE calculated signal and the experimental signal are displayed in the charts below:
The results show that the SF-X-CRK.EXE calculated signal is not very accurate in predicting the experimental signal in this case (AS EXPECTED). However, the calculated signal is not a bad "ball-park" estimate.
Recalculating the signal from a slightly shallower planar flaw (4 mm vs. the real 5 mm depth) and a slightly shorter flaw (10 mm long vs the 12.6 mm real length) appears to generate a signal that closely matches the experimental signal from the real flaw. This indicates that when the flaw is somewhat deeper than a skin depth, and the flaw length to depth ratio is < 3, SF-X-CRK.EXE, over-estimates the effect of flaw depth on the signal and over-estimates the effect of eddy current flowing around the ends of the flaw producing larger and longer signal than it should.