Quantifying Probe Coverage Error in Feeder Thinning Assessment

Abstract

Estimating the actual minimum wall thickness of a feeder pipe is subject to uncertainties associated with the inspection data. The main objective of this study is to quantify the coverage error intrinsic to the 6-probe and 14-probe feeder inspection tools, based on the analysis of over 4,400 inspection scans from Darlington and Pickering reactors. To increase confidence in the results, a smaller subset of well characterized wall thickness profiles was selected as a reference for a simulation approach. The distribution of maximum probe coverage error was estimated in terms of numerous simulated probe passes over the reference profiles. The results of the simulation study show that the maximum probe coverage error ranges from 0 mm to over 0.2 mm in some cases, with the average errors equal to 0.042 mm and 0.065 mm for the 6-probe and 14- probe, respectively. The high variation in the results was due to a strong linear correlation of coverage error with the pattern of wall thinning, with lower values associated with general thinning and higher values applicable to more localized thinning profiles. The thinning patterns were characterized quantitatively using a patch area method, which allows the results to be used for uncertainty analysis and prediction of the actual minimum thickness in a general setting.