Accuracy of ultrasonic flaw sizing techniques for reactor pressure vessels: Final report

The accuracy of several ultrasonic flaw sizing techniques was measured using thick-walled mock-ups simulating typical reactor pressure vessel weld and clad configurations. The evaluated techniques were: backward-scattering tip-diffraction, time-of-flight diffraction (TOFD), dB-drop (with three different amplitude thresholds), and large-diameter focused probes. For the amplitude-based dB-drop technique, amplitude thresholds of 6 dB below peak, 50 percent DAC, and 20 percent DAC were used, and the effect of correcting for the spreading of the ultrasonic beam was also evaluated. The mock-ups contained intentional flaws representing three major classes of defects: cracks under the cladding, embedded defects in welds, and lack-of-fusion defects in nozzle-to-shell welds. Approximately 200 measurements were made. The results showed that for sizing near-surface and embedded defects in thick welds, TOFD and the backward-scattering tip-diffraction techniques were far more accurate than the amplitude-based techniques. Beam-spread corrections reduced the mean sizing errors for the 20 percent DAC sizing method, but made negligible reduction in the scatter of the data. Even after beam-spread corrections were made, the sizing errors of the amplitude-based method remained much greater than the tip-diffraction measurement error. For small-bore nozzles, focused probe measurements were substantially more accurate than unfocused probe measurements. 21 refs., 43 figs., 2 tabs.