Assessment of Local Approach Methods for Predicting End-of-Life Toughness of RPV Steels

Abstract

Local approach methods are becoming increasingly popular as practical tools for cleavage fracture toughness prediction. Their application involves two distinct elements: calculation of "individual" probabilities of failure, dictated by the local mechanical fields; and summation of these failure probabilities to predict the probability of component failure. In this work, we demonstrate that development of the local approach methods to date has been essentially focused on improving the criterion for predicting local failure as a function of the local mechanical fields. Yet, the existing methods fail to predict with sufficient accuracy the effects of irradiation and defect geometry on fracture toughness when the calculations are based on a common set of model parameters. A possible reason for this, common to all methods, is found in the calculation of the cumulative failure probability, which is based on the weakestlink argument. We discuss the implications of the weakest-link assumption, identify those situations where it needs to be reconsidered, and propose future work that will increase our understanding for improving the calculation of global failure probability. INTRODUCTIONThe ability to predict changes in cleavage fracture toughness behaviour of ferritic RPV steels accounting for the effects of irradiation and defect geometry is important to safety assessment and life extension decisions in nuclear plant. Local approaches to cleavage fracture offer a promising methodology to help achieve this. In principle, these are based on current understanding of the micro-mechanisms involved in the cleavage failure phenomenon, such as the nucleation of microcracks at second-phase brittle particles, the propagation of such micro-cracks within grains and the propagation of a critical micro-crack leading to component failure. The scope of this work does not include local approaches that involve changes in

Cleavage Fracture Modelling for RPV Steels: Discrete Model for Collective Behaviour of Micro-Cracks

“…This size dependence has been demonstrated experimentally and formed the first proposal for local probability as a function of both mechanical fields and initiator size [8]. A lack of size dependence in the approaches examined in [7] leads to significantly different predictions for the positions of critical initiators in a given stress field from those seen experimentally. Therefore, a new micromechanically based expression for calculating local probability of failure as a function of initiator size and local mechanical fields was developed in [7].…”

“…The material studied is an RPV 22NiMoCr37 ferritic steel known as Euro Material A, also used as the subject case in [7], for which we have the mechanical and fracture toughness properties. The model is based on an experimentally determined density and distribution of cleavage initiating particles in Material A [12].…”

“…A lack of size dependence in the approaches examined in [7] leads to significantly different predictions for the positions of critical initiators in a given stress field from those seen experimentally. Therefore, a new micromechanically based expression for calculating local probability of failure as a function of initiator size and local mechanical fields was developed in [7]. The new model yielded better predictions for the positions of critical initiators than the previous methods when compared to limited published experimental data [9].…”

“…Existing methods fail to predict with sufficient accuracy and consistency the effects of temperature, irradiation and defect geometry on experimentally obtained fracture toughness values given common calculation parameters [7]. Furthermore, the methods are not able to predict the temperature dependence of fracture toughness in the ductile-to-brittle transition (DTBT) regime, despite disregarding constraint and irradiation effects [7].…”

“…Several local approach methods for assessing end-of-life fracture toughness of ferritic steels have been previously examined [7]. These approaches use different expressions to determine local probability of failure and a common methodology for accumulating such values into a global probability of failure.…”