The Prediction and Operational Experience of Stress Relaxation Cracking in Type 316H

Abstract This paper reviews the methods that have been developed by EDF UK for the prediction of stress relaxation cracking (otherwise known as reheat cracking) in Type 316H austenitic stainless steel weldments and other austenitic stainless steel weldments, for example Esshete 1250. These methods combine the prediction of (i) welding residual stresses, (ii) the prediction of stress relaxation using creep deformation models and (iii) the prediction of intergranular crack initiation using creep damage models. The validation of each of these three aspects of the overall approach will be described by comparison with (i) welding residual stress measurements made on mock-up weldments, (ii) uniaxial stress relaxation and creep deformation tests and (iii) the relaxation and subsequent failure of notched bar stress relaxation tests. The application of the entire methodology to both plant components and to welded laboratory features tests is used to demonstrate the validation and confidence in the overall method. The combination of modelling and testing reveals the critical factors that control the incidence of stress relaxation cracking in plant components. There are exacerbating factors such as the susceptibility of the grade and batch of austenitic stainless steel to relaxation cracking, which is affected by temperature and carbide precipitation. In addition, there are the primary controlling factors influencing relaxation cracking such as the magnitude and multiaxial state of stress of the welding residual stress and the sensitivity of the creep damage model to that multiaxial state of stress.