IIW IWI-C study materials
IIW IWI-C study resources

Fatigue crack at an eccentric diaphragm-to-web fillet

IIW IWI-C · Put your knowledge to work in context.

All case studies

Fatigue crack at an eccentric diaphragm-to-web fillet

A 20-year-old road bridge has a 12 mm thick diaphragm plate fillet-welded to the web of a 1500 mm deep plate girder. The connection was designed as a non-load-bearing detail, but a recent inspection found a 35 mm long toe crack at the diaphragm-to-web weld, starting at the weld toe on the web side. Measured geometry: fillet leg 5.5 mm against 8 mm specified, undercut 0.6 mm deep along 40% of the weld length, and the diaphragm is offset by 9 mm from the web centreline, creating eccentricity. The weld toe radius is approximately 0.2 mm. The bridge carries 30,000 vehicles per day with 12% heavy goods vehicles. The engineer must decide between (a) grinding the toe and re-welding the crack, (b) adding a cover plate, or (c) removing the diaphragm and redesigning the connection. The crack has not reached the web thickness; the web is 14 mm thick.

1. Diagnose why the crack initiated at this location using fatigue and stress-concentration reasoning.

Read the worked answer

The crack is a fatigue crack. The detail was assumed non-load-bearing, but the diaphragm is offset 9 mm from the web centreline, so the weld is eccentrically loaded and the web sees out-of-plane bending. The fillet leg is 5.5 mm instead of 8 mm, reducing the throat to about 3.9 mm, so the stress is higher than assumed. The 0.2 mm toe radius and 0.6 mm undercut create a sharp notch; undercut acts as a crack-like flaw. Fatigue cracks initiate at the toe where the stress concentration is highest. The 35 mm surface length shows the detail has been cycling for years. The crack has not reached the web thickness, but surface length is not depth, so the through-thickness extent is unknown. The root cause is a combination of poor fit-up, eccentricity and poor weld geometry, not just a weld defect. Re-welding the crack without removing the eccentricity will likely lead to recurrence. The detail should not be assumed to have the fatigue category of a simple non-load-bearing fillet; the effective category is lower because of the secondary bending and undercut.

2. Recommend a repair option and explain whether toe grinding can restore the original fatigue life.

Read the worked answer

Toe grinding and re-welding can remove the crack and improve the toe radius, but only if the crack is fully removed and the eccentricity is corrected. Grinding alone cannot fix the 9 mm offset. A cover plate adds new welds and may simply move the problem. Removing the diaphragm and redesigning the connection is the most reliable option, for example a flexible connection with a cope hole or a bolted detail. If the diaphragm is functionally required, use a new connection with a larger leg, a smooth transition and no eccentricity. Toe grinding improves the toe radius and can raise the fatigue category, but it does not remove the secondary bending from the offset. Therefore the fatigue life can be improved but not necessarily restored to the original design category. A fatigue assessment to EN 1993-1-9 should use the measured geometry and traffic data, and the repair detail should be assessed as a new detail category.

3. What inspection and monitoring would you require after the repair?

Read the worked answer

Before re-welding, gouge out the crack and verify removal by magnetic particle or penetrant testing, and by ultrasonic testing or phased array to confirm the crack depth has been removed. After welding, perform magnetic particle or penetrant testing at the repaired toe, and ultrasonic testing if the repair is partial-penetration or if the crack may have entered the web. Inspection should be after a suitable delay to allow any hydrogen-assisted cracking to appear. Record the final weld profile, toe radius and leg length. For monitoring, establish a periodic inspection interval based on the fatigue assessment and the consequence of failure; pay attention to the repaired detail and adjacent welds. If the diaphragm is retained, check for out-of-plane movement or new cracking at the connection. Document the repair, the inspection results and the revised fatigue assessment in the bridge inspection file.