Finite Element Modelling of Fracture Under Combined Primary and Secondary Loads Supporting Updates to R6

Abstract This paper provides an overview of a series of finite element analyses performed to help provide confidence in proposed changes to R6 Section II.6 regarding the treatment of secondary stresses in fracture assessments. Secondary stresses are defined as displacement-controlled stresses such as thermal and weld residual stresses that do not contribute to plastic collapse whereas primary stresses arise from an applied load such as pressure. The impact of secondary stresses on fracture is crucial to an assessment and is commonly one of the driving inputs to evaluating limiting conditions such as limiting defect size. Various work programmes considering how secondary stresses contribute to fracture, in isolation or combined with a primary load, have led to a refinement of the methods within Section II.6 of R6. The work outlined in this paper presents the results of a suite of finite element analyses to allow specific focus to a number of factors related to the treatment of combined loading to be examined with the updates to R6 being applied. These analyses and the comparisons in this paper allows areas to be considered such as the calculation of elastic-plastic secondary stress intensity factor, levels of elastic follow-up under secondary loads in isolation or when combined with other loads, estimates of the normalised secondary reference stress term used in R6, and to compare against some potential new implementations within the R6 procedure that define the interaction of primary and secondary stresses. More realistic methods for the treatment of secondary stress would lead to the fracture assessment of components being undertaken with improved margins that better reflect reality, potentially avoiding the need to deploy in-service-inspection or allowing less onerous inspections during manufacture offering wider benefits. Improved methods for treatment of secondary stresses will also help contribute to the development of probabilistic fracture models since they require a better formulation of realistic (i.e. mean) values of the input parameters. The results in the paper show these new methods provide a bounding but more realistic assessment of combined loading and provide benefit to their inclusion. Inclusion of such methods to R6 then makes it easier to justify the less onerous methods in assessments.