Impact of human and environmental factors on the probability of detection during NDT control by eddy currents

The Probability Of Detection (POD) curves are generally determined in laboratory conditions as the optimum case. These conditions are not the same as in a real in-service inspection. Build an experimental POD curve requires data from a wide range of inspector skills, defect types, test protocols, etc. This paper analyzes the impact of the maintenance conditions on the probability of defect detections in the High Frequency Eddy Currents (HFEC) method. After having described the experimental campaign on titanium specimens associated to HFEC method, we construct an experimental database based on three environments scenario (Laboratory, A321 and A400M aircrafts) and representative inspector pattern with different Non-Destructive Test (NDT) certification levels. Then, the POD using Berens methodology is provided for different data transformation. Comparisons between POD statistical parameters in different scenarios permit to quantify the impact of human and environmental factors. Some additional advices can be done to minimize the uncertainty on POD results. (C) 2018 Elsevier Ltd. All rights reserved.