Probabilistic treatment of pipe girth weld residual stress in fracture assessment

This study examines the influence of probabilistic treatment of transverse residual stress in a structural integrity assessment. Pipe girth weld measurement data gathered from numerous studies was statistically interpreted to provide realistic through-thickness stress distributions for use in fracture assessments. The database measurements are shown to follow a normal distribution which is capable of evaluating a probabilistic interpretation of weld residual stress within a fracture assessment. A comparison between deterministic upper bounds, provided by various standards, and probabilistic interpretation was undertaken. The presented results propose a 42% reduction in the estimated probability of failure of the case study using a probabilistic assumption of residual stress. Direct implications include a more realistic model for treatment in fracture assessments which can in turn improve acceptance criteria and avoid unnecessary weld repairs.