Characterizing Residual Stress Relaxation of 347AP Weldments Using Neutron Diffraction

Abstract Residual stress formation is an inherent feature of welding processes caused by localized heating, impacting weldment performance. One failure mode associated with welds is stress relief cracking (SRC). In the heat-affected zone (HAZ), elevated temperatures can cause precipitate dissolution. During reheating, whether from post-weld heat treatment (PWHT) or in-service conditions, relaxation occurs simultaneously with the formation of large carbides at grain boundaries. Reprecipitation creates a thin region along the grain perimeter that is deficient in alloying elements. Stress can be resolved by plastically deforming and eventually damaging these weakened regions. Significant research has been focused on understanding precipitate kinetics as the primary cause of failure. Nonetheless, an SRC-susceptible microstructure remains stable in the absence of stress. Therefore, understanding relaxation is necessary to produce mechanism-based lifetime predictions. Although residual stress can be modeled, experimental validation is essential to provide confidence in solutions. Neutron diffraction is generally regarded as the most reliable technique for measuring residual stress. 347H is a common creep-resistant austenitic stainless-steel used in the energy and chemical industries. Driven by the need to improve performance without significantly increasing costs, previous studies have shown that slight adjustments in base metal or filler metal compositions can affect cracking susceptibility. In this study, 347AP austenitic stainless-steel tubes were manually gas tungsten arc welded (GTAW) using both 347AP and 16-8-2 filler rods for comparison, and residual stress measurements were taken before and after PWHT. Measurements were made at the High Flux Isotope Reactor’s High Intensity Diffractometer for Residual Stress Analysis instrument beamline located at Oak Ridge National Laboratory. Although there is confidence neutron diffraction as a reliable residual stress characterization method, challenges still arise when determining reference lattice spacings (d0). Therefore, a plane stress assumption was investigated as a viable alternative to measured d0 fields used to calculate lattice microstrain and ultimately stress fields.