Comparison of Experimental and Model Based POD in a Simplified Eddy Current Procedure

The probability of detection (POD) of a well controlled, automated eddy current procedure is evaluated in a numerical model and compared to experiments. The procedure is applied in laboratory environment with a single absolute probe which is positioned for raster scan over flat surfaces containing fatigue cracks in Ti-6Al-4V. Data resulting from the experiments and the computations are treated with the classical approach to render the POD curves. The procedure provides a well defined basis for a comparison between a simulated and an experimentally based POD assessment. Finite element analysis is used to model the eddy current method and a simplified fatigue crack model is introduced. The numerical computations show good agreement with the experimental result.