Fatigue crack at a crane bracket fillet-weld toe
A 20 mm S355J2 bracket is fillet-welded with a 6 mm leg, 200 mm long on both sides, to the flange of a crane runway girder. The bracket carries a cyclic load producing a nominal stress range of 90 MPa at the weld toe for 5×10^6 cycles. After three years, a 25 mm crack is found at the weld toe, initiating at a weld stop/start. The drawing called for a continuous fillet weld, but the fabricator made three passes with restarts and left a 2 mm plate mismatch at the bracket edge. The original fatigue assessment used detail category FAT 71, slope m=3, and 2×10^6 cycles, with no toe treatment. The client wants a repair that avoids recurrence and asks whether the original detail was adequate.
1. Check the original fatigue assessment and explain whether the detail was adequate for the stated loading.
Read the worked answer
For a slope of m=3, the allowable stress range at 5×10^6 cycles is 71 × (2/5)^(1/3) = 52.3 MPa. The applied stress range is 90 MPa, so the detail is overstressed by a factor of 90/52.3 = 1.72 in stress. The corresponding damage is (90/52.3)^3 ≈ 5.1, or a predicted life of 2×10^6 × (71/90)^3 ≈ 0.98×10^6 cycles. Failure before 5×10^6 cycles is therefore expected.
The crack initiation at a stop/start is consistent with the assessment: a weld toe restart is a stress concentration and a preferred fatigue initiation site, and the 2 mm mismatch adds secondary bending. The original FAT 71 detail was not adequate for 90 MPa at 5×10^6 cycles. The fatigue assessment should have used the actual detail, included the restart, and required either a lower stress range or a better detail category.
2. Propose a repair and a design/fabrication change that would prevent recurrence. Support the choice with a fatigue calculation.
Read the worked answer
Repair the crack by grinding out the full crack length to sound metal, confirming with magnetic particle inspection, and re-welding with a qualified procedure. Improve the toe by grinding or TIG dressing, and remove the stop/start by using continuous welding or by positioning restart positions away from the highest-stress region. Correct the 2 mm mismatch by alignment or by adding a transition.
For prevention, the required detail category at 90 MPa and 5×10^6 cycles is 90 / (2/5)^(1/3) = 122 MPa. A full-penetration butt weld ground flush, or an equivalent FAT 125 detail, would be just adequate. A fillet weld with toe treatment may reach a higher category, but that must be supported by qualified fatigue data. Alternatively, reduce the stress range to about 52 MPa by increasing the bracket or girder stiffness. Simply increasing the fillet leg from 6 mm to 8 mm does not change the toe detail category enough and is not a reliable fatigue fix.
3. What inspection and verification would you require after the repair and before returning the bracket to service?
Read the worked answer
After gouging or grinding, inspect the excavated area by magnetic particle testing to confirm the crack has been removed. After repair welding, perform magnetic particle testing or ultrasonic testing of the repaired toe and the adjacent heat-affected zone. Check the toe profile and remove any undercut or sharp restart. Confirm that the repair procedure is qualified for the thickness and steel, and that the welder is qualified.
Verify the alignment and mismatch correction, and record the final detail category used in the fatigue assessment. If toe grinding or TIG dressing is claimed, inspect the treated profile and confirm the process was applied consistently. A strain-gauge measurement or a revised fatigue calculation should be included in the technical file. Return to service only after the repair, toe treatment and revised assessment are accepted by the responsible engineer.