Solving TEAM problem 8 (slot in a plate) on a PC with ELEKTRA

This note presents the solution of TEAM Problem 8, a problem in nondestructive testing with eddy currents. A differential probe consisting of two thin coils 15 mm apart produces a signal. The computation used a fairly coarse three-dimensional mesh: fewer than 8000 elements, fewer than 20,000 equations. Subtracting the signal from a differential probe over a plate with no flaw from the signal over a plate with a flaw eliminated the effects of mesh coarseness. Even though this process requires taking the difference of two already small differences, the solution for the effect of the flaw agrees well with the three sets of experimental results as the results agree with each other. With the use of this process, the results do not show the effects of the edge of the plate that complicate the experimental results. On a 90-MHz Pentium PC, each position of the probe required a 10-minute solution for the case with the flaw and another 10-minute solution without the flaw.