Nondestructive evaluation (NDE) method capability is determined by a statistical flaw detection study called probability of detection (POD) study. In many instances, the NDE flaw detectability is given as a flaw size such as crack length. The flaw is either a crack or behaving like a crack in terms of affecting the structural integrity of the material. Although the crack size relates to structural integrity of the part, it may not be the only factor that affects the flaw detectability for most NDE methods. Crack length, depth, and opening may relate differently to the flaw detectability. Part surface roughness and cleanliness may affect the flaw detectability. The proposed flaw size parameter and the computer application described here give an alternative approach to conduct the POD studies and apply results of the POD study to reliably detect small flaws through better assessment of effect of interaction between various geometric parameters.

Many factors affect X-ray crack detectability. Therefore, crack depth-to-part- thickness ratio is not an accurate way of defining X-ray flaw detectability. Models that relate selected factors such as surface crack dimensions, orientation, part geometry, and technique parameters to film or image contrast for a simulated crack in a plate were developed. A geometric flaw size parameter is intended to be monotonically related to the POD under these conditions. This approach assumes that the flaw size parameter and technique requirements, including the detector response calibration characteristics, together provide a complete description of the X-ray flaw detection application, and provide sufficient information for assessment of the NDE capability, i.e., whether the flaw size under consideration can be reliably detected. A simulated contrast model could do the same assessment if correlations are established between the simulated contrast and the actual contrast, and the simulation contrast and the POD.

This work was done by Ajay Koshti of Johnson Space Center. MSC-25351-1