The Indication That Was Signed Off
CJP T-joints (web to flange) in 1 in A572 Gr 50, AWS D1.1, requiring 100% UT for tension members. The NDT subcontractor's report shows shear-wave UT from the flange face. One record lists a linear indication 1.25 in long with its top 0.5 in below the flange face, disposed as 'acceptable, length less than one-third of material thickness'. UT was performed two hours after the last pass, with shop ambient at 40 deg F. A different joint was rejected and repaired; the subcontractor re-examined only the excavated repair area, using radiography, and passed it.
1. Evaluate the acceptability of the disposition recorded for the 1.25 in indication.
Read the worked answer
The disposition is not defensible, on two counts.
- Wrong criteria. The one-third-of-thickness length rule is a pressure-vessel style criterion, not the D1.1 UT acceptance basis. Under D1.1, cracks, incomplete fusion and incomplete penetration in CJP groove welds are rejectable regardless of measured length, with separate limits for slag and porosity by amplitude and length. A planar indication 0.5 in below the flange face in a T-joint, lying parallel to the flange, is the classic signature of lack of fusion at the flange. That is rejectable whether it is 1.25 in or 0.5 in.
- Who applied the criteria. The report does not show that the evaluator was working to the code the contract invokes.
Action: reject the disposition in writing, obtain the raw scan data rather than the summary, have a qualified Level II or III re-evaluate every recorded indication against D1.1 Clause 6, and re-scan where coverage or technique cannot be verified. The joint is not accepted pending that re-evaluation.
2. UT was run two hours after the last pass at 40 deg F ambient. Does that matter?
Read the worked answer
It matters as a technical risk even where the contract does not spell out a waiting period.
Thick-section CJP welds in A572 Gr 50 made with a hydrogen-bearing process can develop delayed cracking hours after the weld cools. A 40 deg F shop speeds cooling and raises the risk, especially where preheat and interpass control have been marginal. Examining two hours after completion can miss cracks that have not yet opened.
Practical action:
- Require re-examination of that joint after 24 to 48 hours, or as the engineer specifies, before final acceptance. Ideally the first UT is also delayed so the examination is not paid for twice.
- Pull the preheat and interpass records for the joint. Cold ambient plus marginal preheat is exactly the combination that calls for a delay.
- Write the delay into the inspection plan for the rest of the job so it is a rule, not an argument on this one joint.
3. One joint was rejected, repaired, and passed after RT of the repair area only. Assess that re-examination.
Read the worked answer
Two separate problems.
Method. Radiography is not a substitute for ultrasonic testing in a T-joint. The flaws of concern, lack of fusion between web and flange, are planar and parallel to the flange surface, which is the orientation UT was selected to detect. Shooting through a T-joint of that thickness will not reliably resolve a planar discontinuity lying in that plane. Re-examine by UT using the original technique.
Extent. Re-examining only the excavated area is thin coverage. The repair changes local restraint and heat input, and acceptance of the joint depends on a complete scan of the joint, not a patch scan. Re-examine the repaired region plus the length originally examined adjacent to it, as the code and engineer require.
Also verify before accepting anything: the repair was made under a qualified WPS with the specified preheat, the excavation was gouged and dressed to sound metal, and the repair welder is qualified.
Then step back. One reject plus a wrong acceptance criterion plus a wrong re-examination method is a subcontractor competence signal. Audit their written procedure and personnel certification before the next lot, and start trending reject rate by joint and by examiner.