The need to quantify and predict the probability of detection (POD) of defects by ultrasonic nondestructive evaluation (NDE) techniques is driven by the growing importance of damage tolerant design and maintenance philosophies. Experimental demonstration programs, while costly, are useful in assessing NDE system reliability for existent inspection hardware and for “typical” test specimens. However, such an approach is not effective for prediction of optimal inspection protocols nor for improving the inspectability of components at the design stage. There is, thus, a need for computationally efficient analytical models for such applications.