Delayed Cracking in an A514 T-1 Plate Girder
A shop is building a 900 mm deep plate girder. The flanges are 25 mm ASTM A514 Gr. B quenched-and-tempered plate; the web is 12 mm ASTM A36. The bay door is propped open for ventilation, so a steady draft crosses the fit-up, and the shop floor is at 12 °C. The flange-to-web joints are 8 mm fillets with tight fit-up. The crew ran them with the same 1.2 mm E71T-1 wire, 26 V, 240 A, 15 L/min CO2 and no preheat that they had used on an A36 job the week before. To widen the beads the welder used a pronounced weave. Magnetic particle inspection of the finished girder 48 hours later shows several short transverse cracks at the fillet toe on the flange side, 8–15 mm long, lying in the weld heat-affected zone. The girder is promised to the customer in six days, and the engineer asks you to review the whole situation before anyone starts grinding.
1. Explain the cracking mechanism and identify which decisions in the setup contributed to it. Why did the cracks appear 48 hours later instead of immediately?
Read the worked answer
Mechanism: hydrogen-assisted cold cracking in the HAZ.
- A514 Gr. B is a quenched-and-tempered low-alloy steel with a carbon equivalent around 0.55, so its HAZ transforms to martensite at almost any shop cooling rate.
- With no preheat and a 12 °C draft, the 25 mm flange pulls heat out fast, producing a hard, brittle HAZ at the fillet toe.
- The draft and any moisture on the plate or wire put hydrogen into the arc; the tight fit-up and small fillet geometry produce high restraint, so the toe is triaxially stressed.
- The under-matched E71T-1 deposit yields before the HAZ does, pushing strain into the hard zone.
Cracks appear late because hydrogen must diffuse through the lattice to the stress concentration, which takes hours to days at shop temperature. A 48-hour delay is the classic signature, and it is why inspection immediately after welding would have missed this. The wide weave also widened the austenitized zone and slowed cooling only slightly, so it did not help.
2. Specify the welding procedure you would require for this joint: consumable, preheat, interpass temperature, heat input, and whether post-weld heat treatment is appropriate.
Read the worked answer
Consumable: a matching-strength low-hydrogen filler, e.g. E110T5-K4 flux-cored wire or E11018-M covered electrode. E71T-1 is under-matched and was part of the problem.
Preheat: minimum 100 °C, measured on both sides of the joint and at least 75 mm either side of the weld centreline, held until the joint is complete. At 12 °C ambient that preheat also drives off surface moisture.
Interpass: cap at 200 °C. Above that you begin to temper the Q&T plate and lose strength and toughness in the HAZ.
Heat input: keep roughly 1.0–2.5 kJ/mm. For example 26 V × 240 A × 60 ÷ 300 mm/min ÷ 1000 = 1.25 kJ/mm, which is acceptable; stringer beads, no wide weaving.
PWHT: generally avoid it. A514 is tempered near 620 °C, and stress relief at conventional temperatures can soften the HAZ and drop properties below specification. If the contract mandates it, use a qualified procedure held below the mill tempering temperature and verify hardness and tensile results.
3. Give the repair sequence and inspection plan you would insist on, including how you would prevent a repeat.
Read the worked answer
- Excavate: carbon-arc gouge or grind each crack out, extending at least 50 mm past each visible end, then grind the gouged surface bright.
- Verify removal: magnetic particle test the excavation before any welding. If a trace remains, keep excavating; welding over the tip of a crack only moves it.
- Reweld: preheat 100 °C minimum, matching-strength low-hydrogen filler, heat input within the qualified range, stringer technique, and the same interpass cap.
- Re-inspect: MT the repaired area after cooling and again after 48 hours, since this is a delayed-cracking mechanism; also re-test the remaining fillets on the girder, because every one was made the same way.
- Prevent repeat: close the door or screen the work, store wire in a heated cabinet, train the crew on the difference between A36 and T-1, and issue a written WPS with preheat verification by contact thermometer.