Methodology for Identifying Blunt Flaws Using Ultrasonic In-Service Inspection Data

This paper presents a methodology for identifying blunt flaws in piping using in-service data from ultrasonic inspection tools. The method is applicable to data obtained from predetermined inspection grids, or directly from bracelet or array type inspection tools. The flaw edge and size are identified by calculating the three-dimensional slope vector (i.e., magnitude and direction) of each point in the kriging interpolated wall thickness profile. The transition from a steeper slope to a relatively flat profile is then used by a search algorithm to determine the flaw edge. The method is applied to the flaw assessment of feeder piping in CANDU nuclear reactors. The results show that in addition to identifying blunt flaws, the developed methodology also provides a convenient way for characterizing the flaw dimensions for structural integrity assessment. The uncertainty in the results is mainly attributed to the signal loss and coverage error associated with the inspection data.