In the early 2000s, ASME adopted Code Cases N-629 and N-631 [1–2], both of which permit the use of the Master Curve reference temperature (To) to define an reference temperature RTTo, as follows (in SI units, as are used throughout the paper):RTTo=To+19.4℃The Code Cases state that “this reference temperature … may be used as an alternative to [the] indexing reference temperature RTNDTfor the KIcand KIatoughness curves, as applicable, in Appendix A and Appendix G [of Section XI of the ASME Code].” KIa is now only used in Appendix A. The functional form of the ASME KIc and KIa curves dictate that the temperature separation between them remains constant irrespective of the degree of neutron radiation embrittlement, as quantified by ΔRTNDT or ΔRTTo. However, data collected from the literature and new data reported by Hein et al. show that radiation embrittlement brings the KIc and KIa curves closer together as embrittlement increases. As a result, current Code guidance will not produce a bounding KIa curve in all situations when RTTo is used as an reference temperature. To reconcile this issue, this paper summarizes available data and, on that basis, concludes that use of the following reference temperature will ensure that the ASME KIa curve bounds currently available KIa data:RTKIa=RTTo-19.4+44.97×exp⁡−0.00613×RTTo-19.4

The toughness requirements for the ferritic steels used to construct the primary pressure boundary of a nuclear power plant include both transition temperature metrics as well as upper-shelf metrics. These separate specifications for transition and upper shelf toughness find their origins in decisions made during the 1970s and 1980s, a time when there was much less empirical and theoretical knowledge concerning the relationship between these quantities. Currently, significant evidence exists to demonstrate a systematic relationship between transition and upper shelf toughness metrics for RPV-grade steels and weldments (e.g., the equations in draft Code Case N-830-1, empirical correlation between Charpy transition temperature and upper shelf metrics, etc.). This paper explores these relationships and demonstrates that, in many cases, the joint specification of transition temperature and upper shelf toughness values is redundant and, therefore, unnecessary.

Nonmandatory Appendices A [1] and G [2] of Section XI of the ASME Code use the KIc curve (indexed to the material reference transition temperature, RTNDT) in reactor pressure vessel (RPV) flaw evaluations, and for the purpose of establishing RPV pressure-temperature (P-T) limits. Neither of these appendices places an upper-limit on the KIc value that may be used in these assessments. Over the years, it has often been suggested by some of the members of the ASME Section XI Code committees that are responsible for maintaining Appendices A and G that there is a practical upper limit of 200 ksi√in (220 MPa√m) [4]. This upper limit is not well recognized by all users of the ASME Code, is not explicitly documented within the Code itself, and the one source known to the authors where it is defended [4] relies on data that is either in error, or is less than 220 MPa√m. However, as part of the NRC/industry pressurized thermal shock (PTS) reevaluation effort, empirical models were developed that propose common temperature dependencies for all ferritic steels operating on the upper shelf. These models relate the fracture toughness properties in the transition regime to those on the upper shelf and, combined with data for a wide variety of RPV steels and welds on which they are based, suggest that the practical upper limit of 220 MPa√m exceeds the upper shelf fracture toughness of most RPV steels by a considerable amount, especially for irradiated steels. In this paper, available models and data are used to propose upper bound limits of applicability on the KIc curve for use in ASME Code, Section XI, Nonmandatory Appendices A and G evaluations that are consistent with available data for RPV steels.

Efforts are under way to develop a theoretical, multi-scale model for the prediction of fracture toughness of ferritic steels in the ductile-to-brittle transition temperature (DBTT) region that accounts for temperature, irradiation, strain rate, and material condition (chemistry and heat treatment) effects. This new model is intended to address difficulties associated with existing empirically-derived models of the DBTT region that cannot be extrapolated to conditions for which data are unavailable. Dislocation distribution equations, derived from the theories of Yokobori et al., are incorporated to account for the local stress state prior to and following initiation of a microcrack from a second-phase particle. The new model is the basis for the DISlocation-based FRACture (DISFRAC) computer code being developed at the Oak Ridge National Laboratory (ORNL). The purpose of this code is to permit fracture safety assessments of ferritic structures with only tensile properties required as input. The primary motivation for the code is to assist in the prediction of radiation effects on nuclear reactor pressure vessels, in parallel with the EURATOM PERFORM 60 project. This paper begins with a brief overview of the strategy for implementing the new model into the DISFRAC computer code. The balance of the paper focuses on efforts to model the nucleation of a carbide particle crack near an existing macrocrack under applied load. The carbide microcrack initiation model applies dislocation mechanics to assess the stress intensity exerted on a stiff, elastic carbide particle embedded in an elastic-plastic ferrite matrix near a macrocrack tip. The paper derives and discusses the governing equations for the model; including (1) computation of a slip band dislocation pileup distribution by enforcing equilibrium with the macrocrack-induced elastic-plastic stress field, (2) calculation of the mode I stress intensity on the particle crack plane due to the dislocation pileup and (3) determination of the particle fracture toughness. Together, these calculations provide the basis for determining the applied load required to initiate particle fracture. This paper demonstrates how the prediction of particle fracture depends on various microstructure parameters.

The ASME Boiler and Pressure Vessel Code; Section XI provides Rules for inspection and fracture safety assessment of nuclear plant pressure boundary components. This Code provides methods for assessing the stresses and moments contributing to the forces available to drive crack growth in a component as described by stress intensity factors as well as the measures of material resistance to crack extension, measured by fracture toughness. Much of the current Code is based on linear elastic fracture mechanics methodologies developed 40 years ago [1], or more, at a time when drop weight tear tests [2] and Charpy V-notch impact tests [3] were the accepted standards used for characterizing a material’s resistance to brittle fracture. Ensuing research produced experimental methods to directly measure a material’s resistance to both brittle and ductile fracture. Data from such experiments provided the evidence supporting a suite of best estimate models describing fracture toughness behavior across a range of temperatures and strain rates. These models include cleavage crack initiation and crack arrest fracture toughness (KJc and KIa behavior, respectively) on the lower shelf and through transition, and also ductile crack initiation and crack growth resistance (JIc, J0.1, and J–R behavior) on the upper shelf. Best-estimate models provide a more accurate means of assessing a material’s expected behavior under all loading and temperature conditions; they also enable an explicit characterization of uncertainties. For these reasons, there is a growing advocacy within ASME Code groups for incorporating these best estimate toughness models into Sections III and XI of the Boiler and Pressure Vessel Code. The first direct implementation of the KJc best-estimate model in the ASME Code was in Code Case (CC) N-830, which was adopted by the ASME Code in 2014. N-830 states that the 5th percentile lower bound of the KJc Master Curve [4], indexed by T0, can be used as an alternative to the ASME RTNDT-indexed KIc curve in a flaw evaluation performed using Non-Mandatory Appendix A to Section XI. Since that time, work has progressed within the Working Group on Flaw Evaluation (WGFE) to further improve the CC. The proposed Revision 1 of CC N-830 incorporates a complete and self-consistent suite of models that completely describe the temperature dependence, scatter, and interdependencies (such as those resulting from irradiation or other hardening mechanisms) between all fracture toughness metrics (i.e., KJc, KIa, JIc, J0.1, and J–R) from the lower shelf through the upper shelf. By incorporating both a statistical characterization of fracture toughness as well as the ability to estimate a bounding curve at any percentile, the revised CC provides a consistent basis for the conduct of both conventional deterministic flaw evaluations as well as probabilistic evaluations that may be pursued in certain circumstances. Additionally, for the first time within ASME Section XI, both transition and upper shelf toughness properties are provided in a consistent manner in the same document, which provides the analyst an easy means to determine what fracture behavior (i.e., transition or upper shelf) can be expected for a particular set of conditions. The WGFE conducted round-robin assessments of the proposed CC N-830-R1 equations and their use in flaw evaluations, and is supporting documentation of the technical basis supporting the development and implementation of N-830-R1. This paper summarizes that technical basis report. A companion paper presented at this meeting describes the round-robin assessments.