Experimental investigations of fracture and fatigue crack growth of copper–steel joints arc welded using nickel-base filler

Abstract

This paper experimentally investigates fracture toughness and fatigue crack growth behaviour of mode-I cracked bi-metallic compact tension specimens. The specimens were made of weak-copper (UNSC11000) and strong-alloy steel (En31) arc-welded with nickel-filler. Residual stresses induced in constituent materials during welding, due to their different coefficients of thermal expansion (thermal mismatch) were measured at important locations in all the materials near the weld zone. Several notched specimens with positions of crack tip in critical zones of copper, steel and nickel-weld were subjected to fracture tests. The specimens were fatigue precracked prior to fracture testing. Bi-metallic specimens having an equal initial crack length were subjected to constant-amplitude fatigue crack growth rate tests until their fracture occurred. Residual stress intensity factor as a function of crack length was computed theoretically for inclusion in the analysis. Thermal and mechanical mismatches between constituent materials, the latter represented by elastic and plastic property differences, were strongly found to affect the behaviour of crack tip in the vicinity of the weld under both monotonic and fatigue load. The combined effect of “elastic–plastic mismatch” across the material interfaces over the crack tip was described by crack driving force, Jinterface. Under monotonic load (i) fracture toughness of bi-metallic specimens with crack in copper side facing the weld backed by steel was found to be higher than that of monolithic copper with crack arrest noticed in all the specimens when the crack tip was near the weld. Furthermore, Jinterface was found to be positive in this case which implied that the crack driving force, Jtip, was less than the remotely applied crack-driving force, Japplied. (ii) Fracture toughness of bi-metallic specimens with crack in steel side facing the weld backed by copper was found to be lesser than that of monolithic steel with the specimens failing quickly in the form of short brittle fracture within steel and the part of the weld. In addition, Jinterface was negative in this case with Jtip being more than Japplied. Under constant amplitude fatigue load (i) for crack growth from copper to steel side, crack growth rates in copper side retarded in the vicinity of the weld due to shielding caused by the stronger weld and back up steel. The crack deflected away from mode-I plane before penetrating into weld and back up steel. (ii) For crack growth from steel to copper side, crack growth rates in steel accelerated in the vicinity of the weld due to the amplification caused by the “elastic–plastic” property gradient from higher to lower. Furthermore, the crack penetrated through the weld and back up copper with no deflection. The results were also compared with those of monolithic metals subjected to similar fracture and fatigue crack growth rate tests. The shielding and amplification trends in the fatigue crack growth tests were not found to be altered by residual stresses. This implied that the combined effect of “elastic–plastic” mismatch was dominant over that of residual stresses in both directions of crack-transition.