The Reactor Pressure Vessel (RPV) is an essential component liable to limit the lifetime duration of PWR plants. The assessment of defects in RPV subjected to PTS transients made at a European level do not necessarily take into account the beneficial effect of load history (warm pre-stress WPS) on the resistance of RPV material regarding the risk of brittle failure. A 4-year European Research & Development program — SMILE — was successfully conducted between 2002 and 2005 as part of the 5th Framework of the European Atomic Energy Community (EURATOM). The objective of the SMILE project (‘Structural Margin Improvements in aged-embrittled RPV with Load history Effects’) was to provide sufficient evidence in order to demonstrate, to model and to validate the beneficial WPS effect in a RPV integrity assessment. Numerous experimental, analytical and numerical results have been obtained which confirm the beneficial effect of warm pre-stress on RPV steels, with an effective significant increase of the material resistance regarding the risk of brittle failure. In addition to SMILE, a new project dealing with WPS — NESC VII — has been launched in 2008 (linking with the European Network of Excellence NULIFE) with the participation of numerous international organizations (R&D, Utilities and Manufacturers). Based on experimental, analytical and numerical tasks, the project is focused on topics generally not covered by past experience on WPS: biaxiality of loading on large-scale specimens, effect of irradiation, applicability to intergranular fracture, modeling (including analytical and numerical models) … Among these tasks, some new novel WPS experiments are being conducted on large scale cruciform bend bar specimens in order to study the influence of biaxial loading on WPS effect, using a fully representative RPV steel (18MND5 steel similar to A533B steel). After a synthesis of main WPS results available from previous projects on representative RPV steels, a description of the NESC VII project is presented in this paper together with the corresponding organization, including the present status of the project.

For many design and ageing considerations fracture mechanics is needed to evaluate cracked components. The major parameters used are K and J. For that, the different codes (RSE-M appendix 5, RCC-MRx appendix A16, R6 rule, ASME B&PV Code Section XI, API, VERLIFE, Russian Code…) propose compendia of stress intensity factors, and for some of them compendia of limit loads for usual situations, in terms of component geometry, type of defect and loading conditions. The benchmark bench-KJ, proposed in the frame of the OECD/IAGE Group, aims to compare these different estimation schemes by comparison to reference analyses done by Finite Element Method, for representative cases (pipes and elbows, mechanical or/and thermal loadings, different type and size of cracks). The objective is to have a global comparison of the procedures but also of all independent elements as stress intensity factor or reference stress. The benchmark will cover simple cases with basic mechanical loads like pressure and bending up to complex load combinations and complex geometries (cylinders and elbows) including cladding or welds: these cases are classified into 6 tasks. Twenty-eigth partners are involved in this benchmark. This paper gives a global overview of the different tasks of the benchmark and presents the analysis of the results for the first task, devoted on the elastic stress intensity factor calculation (task 1).

The welded joints are particularly sensitive areas in the structures in terms of harmfulness of defects. Given the complexity of the problem (geometry poorly controlled, multi-material aspect, the potential influence of residual stresses), the tests are conducted based on pessimistic assumptions that can wrap all the uncertainties of the problem. In the case of a defect assessment, the considered toughness is deduced from conventional characterization tests with a crack in the welding, considering the current standards, ISO 12135 [1] or ASTM E-1820 [ 2] which are valid only for an homogeneous specimen. In 2010, a new standard ISO 16563 [3] was published to address the specificity of welded joints. If it covers some of the difficulties, it remains incomplete. In nuclear piping, welds have a mismatch M, ie the ratio between the yield strength of the weld metal and the base metal, usually greater than 1: this avoids any problem of strain localization at the junction and ensure that the stresses in the base metal are also easily supported by the welded joint. In this configuration, it turns out that for a given mechanical loading, a crack in the weld located generally has a solicitation, quantified by the parameter J, less (depending on the size of the junction) to those that would see the same crack located in the base metal. Unfortunately, this phenomenon exists also potentially for a characterization test, which would overestimate the true toughness of the welded joint. Plasticity that develops from the crack tip can quickly reach this interface and be affected. To evaluate this phenomenon, we considered two types of representative welded joint (PWR secondary loop ferritic weld and a 316 stainless steel weld) and performed a F.E. analysis of the multi-material CT specimen mechanical answer and on the η coefficient conventionally used to derive the plastic component of J from the area under the curve force-opening displacement.

In the framework of the NESC VII European project, a large experimental program has been dedicated to characterizing the Warm Pre-Stressing (WPS) effect in different testing configurations. One of the CEA (France) contributions to this project is the realization of five point bending tests on large cruciform specimens considering different WPS loading cycles. The five cruciform specimens, sponsored by EDF (France) and IRSN (France), are made of 18MND5 steel. Two of them have been tested on a same LCF (Load-Cool-Fracture) loading cycle and three others on a LCTF (Load-Cool-Transient-Fracture) loading cycle. The experimental results presented in this paper give a successful demonstration of the WPS effect in biaxial loading conditions either on a LCF or on a LCTF cycle. During the test interpretations, different models have then been tested and compared in order to evaluate their ability to predict the cleavage fracture in the case of different WPS loading cycles. They all provide very conservative predictions whatever loading cycle is concerned.

It is well known that the tearing resistance curve J–Δa is not a material property and that probably the energy dissipation rate is preferable to the integral J to characterize crack growth. The G parameter represents the energy dissipated in plasticity and fracture and, under certain conditions, this parameter could be used directly as a critical value for a criterion based on an energy release rate calculated near the crack tip to simulate propagation. Indeed, we see that a local energy release rate could be calculated near the crack tip to take account only of that portion of energy which participates in the fracture. We applied this approach to simulate the crack growth for CT specimens with and without side grooves in a 20 MnMoNi 55 ferritic steel, and compared the results with Rousselier’s model.