The relationship of weld metal hardness residual stress and susceptible to stress corrosion cracking in hydrogen sulphide environment in A516 Grade 70 carbon steel shielded metal arc welded joint

Abstract

In the petrochemical industry, carbon steels exposed to H2S environments may be susceptible to stress-corrosion cracking. A tensile residual stress and high hardness increases the risk of cracking in H2S environments. NACE SP 0472 limits weld metal hardness to 210 HV10 (200 HB) and heat-affected zone hardness to 250 HV10 to prevent stress-corrosion cracking of carbon steel welds in H2S. The hardness is controlled by the weld thermal cycle or by a post-weld heat treatment. In this project, the effect of hardness on the susceptibility to stress-corrosion cracking was investigated by increasing electrode strength systematically and measuring residual stress in the weld metal in the as-welded state. Samples were manufactured from SA 516 Gr 70, a carbon steel used extensively in the petrochemical industry. Heavily clamped plates were welded to minimise distortion and to maximise residual stress. The weld metal hardness was increased by using E6013, E7018-1, E8018-B2 and E9018-B3 electrodes without a post-weld heat treatment. Mechanical tests included all-weld and transverse tensile tests, impact strength and hardness testing. As the nominal strength of the weld metal increased, the all-weld tensile strength increased from 512 to 829 MPa, while the yield strength increased from 443 to 659 MPa. The average weld metal hardness increased from 177 to 317 HV10. The transverse tensile strength was between 511-517 MPa, while the yield strength (in the transverse direction) was between 360 and 382 MPa. Residual stresses of the welded joint were measured by neutron diffraction in the through-thickness, longitudinal and transverse direction. The Von Mises theorem evaluated the principle residual stress. Results indicate that the residual stress in the weld metal may be up to 99% of the yield strength. For stress-corrosion cracking, the samples were submerged in the standard TM0177-2005 test solution for 30 days. The only sample to crack was the E9018, with an average weld metal hardness of 317 HV10. The study results were consistent with the NACE SP0472 specification and earlier publications.