Delayed Transverse Cracks After a Thick-Section Repair
A pressure-vessel repair is being made on a 50 mm thick ASTM A516 Grade 70 shell. The repair groove is double-V, 60° included, and is welded with E7018 H4R, 4.0 mm diameter, using a WPS that specifies 20°C minimum ambient, no preheat, and no post-weld heat treatment. The ambient temperature is 8°C, and the plate is not heated. The root pass and fill passes are completed; the cap is ground flush. The joint is highly restrained by a ring stiffener 150 mm from the weld. After 48 hours, magnetic particle inspection reveals several transverse indications at the weld toe on the shell side, each 4–8 mm long. The welder says the electrodes were taken from a sealed container and used within 2 hours. The repair is scheduled for hydrotest in 3 days.
1. What cracking mechanism is most likely, and what evidence in the setup supports it?
Read the worked answer
Hydrogen-assisted cold cracking (delayed cracking) is the most likely mechanism. The repair is on 50 mm A516 Gr. 70, a thick, restrained joint; the WPS allows 20°C minimum ambient but no preheat, and the actual ambient is 8°C. The plate is not heated. E7018 H4R is a low-hydrogen electrode, but low hydrogen does not eliminate hydrogen; moisture, surface contamination, and atmospheric pickup still matter. The indications appeared after 48 hours, not immediately, and they are transverse at the weld toe-typical of hydrogen cracking under restraint. The 4–8 mm lengths are small and sharp, which fits cold cracking rather than hot cracking or solidification cracking.
2. What tests would you use to confirm the mechanism and characterize the cracks?
Read the worked answer
Start with surface NDT: MT is appropriate for surface and near-surface toe cracks in ferromagnetic steel; PT can supplement but is less effective for tight cracks. UT or TOFD can size the indications and check for subsurface extension. For confirmation, cut a section through a representative crack and do a macro etch, hardness traverse, and metallography. Look for a brittle, intergranular or transgranular path with no decarburization; check hardness in the heat-affected zone. Hydrogen analysis is not practical after the fact, but hardness and microstructure support the diagnosis. A destructive bend or nick-break test on a coupon from the same repair can reveal crack path if sampling is allowed.
3. What WPS/PQR changes would you require before resuming the repair?
Read the worked answer
Revise the WPS to require preheat based on thickness and carbon equivalent; for 50 mm A516 Gr. 70, a preheat of at least 100–150°C is typical, and the exact value should be supported by a qualified PQR or code table. Control interpass temperature and use dry, low-hydrogen electrodes with proper holding ovens and lot control. Add a post-weld heat treatment or a controlled cooling/hydrogen bake-out if the code or service requires it, and delay final inspection for at least 48 hours after the weld cools. Qualify the revised procedure with a new PQR that includes preheat, thickness, and PWHT variables, then re-inspect the repair with MT/UT before hydrotest.