Investigation of a Model-Assisted Approach to Probability of Detection Evaluation

Establishing the Limits of Validity of the Superposition of Experimental and Analytical Ultrasonic Responses for Simulating Imaging Data

“…calculated using large numbers of experimental results requiring many samples and trained technicians to perform the inspections, all at a great cost [5]. ere is a trend towards using model-based and model-assisted POD (MAPOD) curves to reduce this cost [6], however, this method requires simulations with a high degree of accuracy. While the principles behind simulating ultrasonic data and their resultant images has been known for several decades [7], [8], accurately simulating all sources of noise in real inspections proves di cult to achieve using analytical methods.…”

“…It is common practice to set a threshold amplitude, over which any pixel in the image can be considered a defect. For this paper, the threshold level has been set at 12.5 dB above the RMS noise value and this can also be seen in gure 6. is has been calculated to give a false call ratio of 1 in 1000 for the given image size and assuming the image pixel amplitude due to speckle follows a Rayleigh distribution.…”

Finite Element Modelling of a Reflection Differential Split-D Eddy Current Probe Scanning Surface Notches

“…A model developed by Nakagawa et al [19] for a split-D probe described the effect of electrical discharge machined (EDM) notch width on the probe's output signal, and simulation results were then validated with experimental tests. The primary objective in most of these modelling approaches is to reproduce the experimental results to the highest possible accuracy since a valid EC model can form the foundation of defect characterization (inversion), and reliability studies [22][23][24][25]. Therefore, it is of high importance to conduct a proper validation study to assess the rigorousness of each EC model [14,18,26].…”

Advanced ultrasonic “Probability of Detection” curves for designing in-service inspection intervals

“…For all these reasons, their definition process is very demanding in terms http://dx.doi.org/10.1016/j.ijfatigue.2015.07.018 0142-1123/Ó 2015 Elsevier Ltd. All rights reserved. of time and costs and, over the last years, experimental responses have begun to be partially substituted and integrated by numerical simulations [8][9][10][11][12].…”