Estimating POD of a screening technique for cracks about ferrous fasteners without fastener removal

Abstract

A study was conducted to estimate the Probability of Detection (POD) of a screening technique for samples representative of lap joints in wing structure of the P-3 Orion airframe containing EDM notches in the bottom of the top layer and the top of the second layer radiating from holes containing ferrous fasteners. It was assumed that the defects made up only a small fraction of the data (<= 5%) as would be normal in a damage tolerant design. A novel pulsed eddy current probe and robust cluster analysis were used to provide blind detection of the notches without removal of the ferrous fasteners. Detection could be confirmed using traditional bolt-hole eddy current after fastener removal at that site. Statistical simulations were used to explore the effects of sample size and number of defects in a sample. Using a nominal false call rate of 15% (actual ∼10%), the estimated a90/95 was better than 2.3 mm (0.09 inch) for these data sets. This represents a marginal probability as the results indicate that other variables affect the estimated POD. Sample statistics can be used to show that the assumptions of the technique (i.e. very small fraction of the fasteners have associated defects) are not violated and that the inspector performed the inspection in a suitable manner.