Thick-section SAW weld: PAUT indication missed by RT
In a 48 mm thick C-Mn steel pressure vessel, a double-V butt weld is examined by radiography and phased-array ultrasonic testing. RT, performed with Ir-192 at 2.5% sensitivity, shows no relevant indication and is accepted by the client's inspector. PAUT, using a 5 MHz linear array and a focused wedge, reports a planar indication 28 mm long and 3.5 mm high, 18–22 mm below the surface, oriented parallel to the fusion face. The weld was made by SAW with a narrow groove; the procedure was qualified with RT only. The vessel is due for hydrotest in 72 h. The fabrication specification accepts ISO 5817 level B, but the client's engineer asks for an engineering critical assessment (ECA) instead of immediate repair. The material yield strength is 355 MPa, design stress is 0.7 of yield, and the lowest design temperature is 0 °C.
1. Explain why the two NDT methods disagree and which result should govern the immediate disposition.
Read the worked answer
RT with Ir-192 is a volumetric method but is poor at detecting planar defects oriented normal to the beam, especially lack of sidewall fusion in a narrow-groove SAW weld. The PAUT indication at the fusion face is exactly the geometry RT can miss. Absence on RT is therefore not proof of soundness.
For immediate disposition, stop the hydrotest and treat the weld as containing a suspect planar flaw. Verify with a second PAUT operator and TOFD, or open the weld for MT/UT from the inside if accessible. The PAUT result should govern until resolved, but it must itself be validated because sizing and detection are operator-dependent.
2. Outline the ECA inputs and give a reasoned recommendation on repair versus acceptance.
Read the worked answer
ECA inputs: flaw height 3.5 mm, length 28 mm, location 18–22 mm below surface; membrane stress 0.7 yield = 248 MPa; yield 355 MPa; toughness from Charpy or CTOD; misalignment and residual stress; weld strength mismatch.
Under BS 7910 Level 2B, a lack-of-fusion flaw of this height in a 48 mm wall is a significant planar defect. It may be acceptable if fracture toughness is high and stress is low, but the conservative route is repair. I would not accept the flaw solely because RT missed it; require ECA by a competent fracture-mechanics engineer and client approval. Given the 72 h hydrotest deadline, repair is the safer schedule decision.
3. What verification NDT would you require after repair and before hydrotest?
Read the worked answer
After repair:
- remove the defect to sound metal;
- qualify the repair WPS, preheat as required, weld;
- perform PAUT/TOFD over the repaired area plus 100 mm each side;
- MT/PT of the cap and root;
- add RT if volumetric confirmation is required.
Repeat hydrotest only after NDT acceptance. Document the flaw, repair and ECA decision.