High-frequency shear wave on an austenitic pipe weld
A 316L austenitic stainless steel pipe butt weld, 88.9 mm outside diameter and 7.62 mm wall, is examined 24 hours after welding. The procedure uses a 5 MHz, 70° shear-wave probe on a curved shoe and a 2 mm side-drilled-hole DAC. The base metal produces grass up to 40% full screen height and many small echoes at 10-20 mm sound path. A 6 mm long response near the fusion face at 5 mm depth reaches DAC +2 dB when the probe is directed circumferentially one way; when reversed, it falls to DAC −6 dB. Coupling is good and the probe index is correct. The specification allows shear-wave or longitudinal-wave techniques but demands reliable detection of lack-of-fusion. The client asks whether the technique is fit for purpose and whether the response is a defect.
1. Is the 5 MHz shear-wave technique suitable for this austenitic weld? Justify your answer and propose an alternative.
Read the worked answer
The 5 MHz shear-wave technique is not demonstrated fit for purpose. Austenitic weld metal has large columnar grains; a 5 MHz shear wave has a short wavelength and is strongly scattered, causing high grass, beam skew and variable response. Grass up to 40% FSH with an indication only 2 dB above DAC gives poor signal-to-noise and is not reliable for lack-of-fusion detection. Before rejecting the technique, run a capability trial on a representative austenitic test piece containing known lack-of-fusion; require a clear response and signal-to-noise of at least 6 dB. If that fails, change to a lower frequency, about 2-2.5 MHz, and preferably a longitudinal-wave technique using dual-element or focused probes; phased array sectorial scanning can also help. Lower frequency reduces scattering but worsens resolution, so probe and focusing must suit 7.62 mm wall. Document the technique change and re-qualify the procedure.
2. How should the directional response be interpreted, and what further examination would you perform?
Read the worked answer
An 8 dB change on reversing the probe is too directional for random grain noise. It suggests a planar reflector whose orientation favours one beam direction, or beam skew in the austenitic weld. Do not classify it from one direction. Rotate the probe through 90° and map the response; scan the same volume from both sides of the weld and around the circumference; repeat with a lower-frequency longitudinal-wave probe. Check that the depth corresponds to the fusion face and measure length by the 6 dB drop method. If the indication remains planar, is reproducible from at least two directions and exceeds the recording level, treat it as lack of fusion: reject and repair. If it disappears at lower frequency and is not reproducible, it may be a grain-boundary reflection, but that conclusion needs the capability trial and a second technique, not amplitude alone.