1.2.1 Surveillance programmes
The accurate analysis of data coming from surveillance programmes (which are part of codes and standards in each country [Keim2012]) is a key factor for direct understanding of “real” materials in operating plants as their lead factor (2–18) with respect to the reactor pressure vessel (RPV) beltline is generally small [Debarberis1998]. Yet, for the better understanding of the irradiation embrittlement mechanism, systematic variation of different influence parameters is necessary. These variations cannot be realized by the standard surveillance programmes. Besides, geometry of the specimens and the irradiation capsules may influence the results when non-uniform radiation fields or temperature distribution throughout the specimen / capsule is present due to its size and location [PRIS2004]. Therefore material specimens are placed in material testing reactors or in suitable channels of commercial reactors to better control and to allow the variation of the irradiation conditions. This method relies on accelerated data obtained at fluence rates (mostly referred to as “flux”) greater than those present at the real component. The verification of effects related to the fluence and the flux is crucial for the proper interpretation of this accelerated data. In some cases, round-robin tests in the analysis of irradiated specimens are organised to validate results from different research groups [IAEA09] [GRETE2004] [EUR22282].
For the systematic study of the influence of certain alloying and accompanying elements, often model alloys are manufactured allowing a broader variation of chemical compositions than commercial steels. The portability of the results to commercial steels, however, must be critically reviewed [JRC24564].
1.2.2 Measurement of Irradiation Parameters
One key factor in irradiation experiments – beside neutron fluence and neutron spectrum – is the measurement of the irradiation temperature as exact as possible because the diffusion and therefore also the embrittlement process is highly dependent on temperature. Despite the thermal contact to the surrounding coolant, the specimen temperature may be significantly influenced by gamma heating. One common technique to monitor the maximum temperature during the irradiation period is the use of melting alloys in the surveillance capsule. Ballesteros et al. performed experimental investigations on the reliability of temperature measurement techniques in Kola-3 NPP (WWER 440/213). They used a more precise technique by inserting thermocouples inside the surveillance capsules which causes more experimental effort. As a result, the difference between the specimen temperature and the water inlet temperature as a well accessible variable was found to be around 5 °C which is considered acceptable without correction of the data. In addition, the positioning of the specimen inside a capsule may influence the irradiation temperature. Beyond the experimental part of the study, the authors developed 2D and 3D Finite Element (FEM) models, respectively, to evaluate the capsule temperature. They found good agreement between modelling and experimental results [JRC31142].
Another key factor in irradiation experiments is the measurement of the neutron fluence as exact as possible because the embrittlement process may depend on neutron fluence. One common technique is the use of activation monitors as dosimeters. Ballesteros et al. performed experimental investigations on the reliability of various activation monitors in Kola-3 NPP (WWER 440/213). Typically, monitors of Fe, Cu, and Nb are used. In the study of Ballesteros et al., Ni, Nb, Cu, Ti, and Fe monitors were compared, the Fe detector being most important. Differences between different detectors were less than 10 % except the Nb detector showing deviations up to 20 % [JRC31142]. During the REDOS project, reactor dosimetry was analysed and benchmarked using experimental and computational techniques. Special attention was paid to the improvement of neutron-gamma calculation methodologies and the accurate determination of radiation field parameters near and throughout the RPV wall. For this purpose, experimental techniques were improved or new ones developed. Results of RPV attenuation calculations revealed that using neutron energies above 0.5 MeV (as mostly used for WWER RPV materials) is more conservative than using neutron energies E > 1 MeV or dpa (as typically used for Western RPV) – at least for WWER RPV [EUR21771] [Ballesteros2004]. A detailed comparison between different neutron fluence scales and their conversion was performed during the MADAM project [Debarberis1998].
1.2.3 Specimen Examination Techniques
Irradiated specimens are typically analysed using destructive techniques like, e.g., Charpy impact or tensile tests. From the Charpy impact tests properties like the ductile-to-brittle transition temperature and the upper-shelf energy can be derived. They give helpful information for the integrity assessment of the RPV [Keim2012] [Horvath2005] [PISA2005] [EUR19961], yet a correlation to fracture toughness is still needed. Within the REFEREE project, different techniques (Charpy impact, static and dynamic fracture toughness) were applied to non-irradiated and irradiated 15Kh2MFA (WWER RPV material), 18MND5, JRQ (Western PWR RPV materials), and Magnox steel to compare the results gained by these techniques. In most cases, data gained by the different techniques were in accordance with correlations in codes and standards (ASTM International). Significant differences were found for the Magnox steel. In this latter case, the low number of available specimens and the high experimental data scatter may be one explanation. In the case of the 41 J Charpy-V transition temperatures, high scatter of data was found although the scatter in the mean correlation is satisfying [EUR19961].
When discussing long term operation beyond design life, more surveillance data irradiated to high values of fluence is necessary for the RPV. As the number of specimens is usually determined by the design life and specimens cannot be placed in retrospect in a larger amount into the surveillance channel of the NPP, other approaches had to be found. In this context, the RESQUE project investigated the feasibility of “reconstitution” of specimens from broken parts of tested specimens. E.g., broken halves of Charpy specimens from the irradiated material of interest were machined and welded to additional unirradiated material with similar properties (“reconstitution”) to gain standard specimens for further irradiation and testing. The first comparisons between reconstituted and compact specimens show promising results concerning mechanical properties with deviations only within the error margins of the testing technique. Thus, reconstitution is seen as one important aspect in future RPV surveillance [Walle2001].
In the same context of long term operation, non-destructive techniques are seen as another promising option to gain more surveillance data. The advantage of non-destructive techniques is that specimens are not wasted and can be re-inserted into surveillance channels or other irradiation facilities to further raise the neutron dose [GRETE2004]. In the AMES-NDT project, different non-destructive techniques were reviewed in an extensive bibliographic study and compared in a state-of-the-art report. Afterwards, reference samples were prepared and demonstration tests were performed [AMES-NDT2000]. Results of this comparison and implications for the following work were presented [AMES-NDT2000] [Dobmann2001]. During the subsequent GRETE project, the following non-destructive techniques were examined in more detail: STEAM (see below), 3MA (micro-magnetic technique based on multiple parameter regression algorithm), MBN (magnetic Barkhausen noise), ABIT (automated ball indenter), TEP (thermoelectric power), and HAS (Harmonic-Analysis-System). All techniques show promising results, but widespread use suffers from, e.g., few statistical data, highly variable and hard to characterise residual stresses, poor quality of results, or missing information about material details. In general, non-destructive techniques based on thermoelectric or mechanical properties as well as Barkhausen noise measurements show better results than the other ones [GRETE2004] [EUR22282].
A non-destructive technique for analysing microstructural changes upon neutron irradiation was developed and tested successfully: the Seebeck and Thomson effect on aged materials (STEAM). It is based on the measurement of the thermoelectric voltage occurring in the tested specimen due to Seebeck and Thomson effects combined to the Relative Seebeck Coefficient (RSC) [Acosta2001]. The RSC is characteristic of the material and its structure. The STEAM technique works well when analysing effects of copper and – with less success – nickel as alloying element while it fails when phosphorous is the element under investigation due to differences in microstructural changes caused by copper and nickel on the one hand and phosphorous on the other hand [JRC26177] [JRC30552]. REAM, a similar technique using the resistivity effect on aged materials is also presented by Acosta et al. but further validation is necessary [JRC30552]. In a comparison, STEAM was found to be more sensitive than REAM. A combination of both slightly improves the measurement results [JRC30867].
As the irradiation experiments are costly and time-consuming, huge effort was paid in the development of computational models predicting irradiation effects as a function of surrounding parameters like, e.g., temperature and neutron fluence (see, e.g., [EUR23742] and references therein). Beside the possibility to reduce experimental demand, models allow to extrapolate parameter values into parameter fields where no experiments are available yet. If these models are based on physical dependencies the extrapolation can be quite successful. Fundamental research in materials microstructural changes upon neutron irradiation is widely performed and a number of different models for this purpose have been set up. It is, however, often not easy to gather information about macroscopic changes of materials properties (i.e. irradiation embrittlement) from these models [Souidi2006]. As a final point, a virtual test reactor was developed to simulate irradiation effects in light water pressure vessel steels [Jumel2005].
A new positron lifetime spectrometer was manufactured within the PERFECT project to perform high-sensitivity positron annihilation spectroscopy measurements [Jardin2006]. These measurements – often together with accompanying Monte Carlo simulations (e.g., [Domain2004]) – give insight into microstructural changes during irradiation and hence help to understand the embrittlement mechanism. For further information see [EUR24455] and references therein.
Although the state-of-the-art at the end of the PERFECT project was already satisfying in many cases, open issues still remained. The experimental work on microstructural phenomena was continued in the LONGLIFE project and modelling work in the PERFORM60 project. For an overview see [Ortner2014] and [Malerba2014], respectively. In the course of the PERFORM60 project, modelling of irradiation effects in the FeNiMn and FeCuC systems was improved using ab initio and themordynamics calculations. While the FeCuC system could be treated satisfactorily, the FeNiMn systems was more challenging, so open issues in this system remain [PERFORM2013].