Abstract
ASTM A709 Grade 50CR (50CR) is a low carbon (< 0.03 wt% C) structural steel with a ferritic-martensitic dual phase microstructure developed to address the corrosion issues associated with the use of traditional weathering steels, especially in environments involving combined prolonged wetness and chloride salts. The typical composition of 50CR contains 10.5 to 12.5 wt% chromium with a maximum of 1.0% silicon, 1.5% of nickel, and 1.5% manganese. When 50CR is produced by ArcelorMittal under the tradename Duracorr®, 0.20 – 0.35% molybdenum is also added. The fabrication of bridge girders from 50CR has largely relied on the use of 309L solid filler wire, which is an austenitic stainless steel. Prior to this research, heat inputs up to 55 kJ/in and maximum interpass temperatures up to 450 °F for 1” thick plate have been explored by the Oregon Department of Transportation.
The objective of this project was to determine the high end of heat inputs and maximum interpass temperatures that could be used during submerged are welding (SAW) of 50CR and to evaluate alternative filler wire materials for fabrication through the use of both 1/2” and 1” thick plates. Given the desire to produce 50CR bridge girders that are considered maintenance-free, microstructural analysis as well as mechanical testing was carried out to evaluate the proposed heat inputs in terms of failure within the lifetime of the bridge. A total of four austenitic stainless steel filler wires were used in this study, three of which were solid wires: 309L, 309LSi, and 316L, and one metal cored wire: 309LC. The heating profile was modeled using the Rosenthal equations while the fusion microstructure was modeled using the Schaeffler constitution diagram.
The following conclusions were drawn from this research. A range of austenitic stainless steel filler wires can be used to successfully weld 50CR; 309LC, 309LSi, and 316L were all found to be viable alternatives to the incumbent filler wire, 309L. Use of 309L filler wires for 1/2” plates results in welds which are borderline with respect to impact energy at the high heat input level of 75 kJ/in, over the maximum interpass temperature range investigated. Lower heat inputs (e.g., 50 kJ/in which was previously shown to be effective) are recommended. Similar results were obtained for welds in 1/2” plates produced with 309LSi and 316L. Heat inputs of up to 75 kJ/in at all maximum interpass temperatures explored (up to 450 °F) can be used to weld 1/2” plates with 309LC filler wire. Heat inputs of up to 90 kJ/in and interpass temperatures of up to 450 °F can be used for all four of the filler wires investigated (309L, 309LC, 309LSi, and 316L) for 1” thick plate. All of the resulting welds surpass the mechanical requirements of AASHTO/AWS D1.5 bridge welding code.
Interestingly, the metal cored wire 309LC out-performed the solid filler wires examined in this study with respect to production (higher deposition rates and fewer required passes) and mechanical properties (especially impact toughness) of the resulting welds. The enhanced toughness and lack of ductile to brittle transition in the fusion zone of 1/2” plates welded using 309LC filler wire appear to be due to the mitigation of large, aligned δ-ferrite grain formation during solidification. Such large, aligned grains otherwise serve as preferred crack paths in welds. This difference in solidification structure is hypothesized to be due to the difference in molten metal transfer modes suggested by literature.
Finally, tensile and Charpy V-notch tests to failure of the base material, 50CR, revealed a curious delamination cracking phenomenon akin to fracturing of wood. The delamination cracking of 50CR occurs due to the intersection of both microstructural and mechanical aspects, including stringers which appear to lower the through-thickness strength of the plate and a pronounced neck (with about 80% reduction in area) that causes the local stress state to transform from uniaxial to triaxial. Future research could focus on exploring the implications of this delamination phenomenon on other sorts of welds, such as T-joints. No cracks of any sort (including delamination cracks), however, were observed in the butt-welds created in this study.
