Bicharacteristic probability of detection of crack under multi-factor influences using alternating current field measurement technique

Alternating Current Field Measurement (ACFM) technique has been widely used in various fields such as oil & gas, aerospace areas. However, during the detection process, the probability of detection (POD) is influenced by multi-factors such as lift-off height, scanning angle and detection speed. In this paper, a bicharacteristic probability of detection (BPOD) model is developed for quantifying the crack detection reliability using integrated Bx and Bz signals in the ACFM technique. A multidimensional BPOD model is established to study the influence of multiple factors on the detection probability of cracks using ACFM technique. A Bayesian network-based method is proposed to reverse the hierarchical impact weights of influence factors on the BPOD of the ACFM technique. The results show that the BPOD model considering both Bx and Bz response signals can evaluate smaller defects than the conventional POD model. The multiple influencing factors can substantially decrease the BPOD of cracks. The hierarchical impact weights on BPOD for cracks are ranked as follows: lift-off height, detection speed, personnel involved in the detection process and scanning angle. This approach can retrospectively find the potential contributors to the missed detections based on the BPOD.