Delayed toe cracking in the HAZ of a thick S460ML node after an un-preheated repair
A 45 mm thick S460ML tubular node is welded with a basic flux-cored wire under M21 shielding gas, with 150 °C preheat and 250 °C maximum interpass temperature recorded on the WPS. An MPI indication in the root region of one brace was gouged out and re-welded the next morning at 12 °C ambient, without preheat, and the operator did not record interpass temperature because "it was only a small repair". Seventy-two hours later, MPI shows three short transverse indications at the weld toe in the coarse-grained HAZ, and a portable hardness survey gives 375 HV10 near the same location. The client's specification requires 100% UT 24 hours after welding, which had found nothing. The structure is due to be loaded in three weeks.
1. Explain the cracking mechanism and why the 24-hour UT found nothing. What is the significance of 375 HV10?
Read the worked answer
The pattern is hydrogen-assisted cold cracking. Three conditions coincide at one location: a crack-susceptible coarse-grained HAZ, hydrogen, and high restraint. The 45 mm section gives high restraint and fast heat extraction; the repair was made at 12 °C without preheat, so the weld cooled quickly, transformed to hard constituents, and hydrogen had no time or temperature to diffuse out. The indications sit transverse to the weld at the toe in the coarse-grained HAZ, which is the classic site, and the 72-hour delay is typical of delayed cracking.
375 HV10 is above the usual acceptance ceiling for this type of steel - 350 HV10 is the common limit in ISO 15614-1 and in EN 1090-2 hardness requirements. Hardness is not itself a crack, but it is physical evidence that cooling and hydrogen control were inadequate and that the HAZ is crack-susceptible. It justifies investigating the whole node, not only the repaired brace.
UT at 24 hours found nothing for two reasons. Hydrogen cracks need time to initiate and grow, so 24 hours is too early for a 45 mm high-strength joint; 48–72 hours is the usual delay. And tight, planar toe cracks are difficult for conventional UT, whereas MPI at the surface is well suited to them. Subsurface extensions need PAUT or TOFD, not a repeat of the same 24-hour scan.
2. What should have been done differently - preheat, hydrogen control and inspection timing - and how would you now establish the extent of cracking?
Read the worked answer
The repair should have been treated as a weld, not as a small job. The WPS preheat of 150 °C should have been applied and measured at the joint with a contact thermometer or thermal crayon, held through the repair, and the interpass temperature recorded. Consumables should have been low-hydrogen, stored and handled per the manufacturer, and a dehydrogenation hold (for example 250 °C for 2 hours) or post-weld heat treatment considered for a thick, restrained repair. "Small repair" is not a justification for omitting any of this, and the preheat and interpass values should appear in the record.
Inspection timing should have been agreed up front: for this section and strength, NDT after 48–72 hours, with surface methods (MT/PT) as well as volumetric.
To establish extent now, do not simply grind the visible toe. Clean and MPI the whole node, then apply PAUT from the accessible side. Excavate step by step with MPI at each stage to follow the crack tips, mapping length, depth and orientation. Where a crack runs under the cap or into the parent, excavate and verify by MT/PT, and run a hardness survey across the area. Then decide repair versus cut-out with the responsible engineer, and consider whether a fitness-for-service assessment is needed rather than repeated local repair.
3. Set out the repair and verification route you would require before the structure is loaded, and the acceptance criteria you would insist are defined in advance.
Read the worked answer
Excavate each crack to sound metal plus a margin beyond the tip (typically about 50 mm, confirmed by MT or PT at each stage), using a qualified repair WPS with the required preheat, low-hydrogen consumables, and controlled heat input so the t8/5 and resulting hardness stay within the qualified range. Apply a dehydrogenation treatment after welding. Check hardness (≤ 350 HV10) at the repaired area and in the HAZ.
Inspect after the agreed delay - 48 to 72 hours - with MT/PT over the full repaired region and UT/PAUT for the subsurface, not immediately after welding. Any further excavation re-starts the same sequence.
Define the acceptance criteria before work begins: which standard and acceptance level apply, whether the client's engineer must approve the repair procedure and each excavation, and the maximum number of repairs permitted at one location before replacement is required. Document all of it, because the missing preheat record and the "clear" 24-hour UT report are already quality-record findings against ISO 3834-2 and should be raised as nonconformities independent of the repair.