Objective

This project on Reactor Vessel embrittlement was identified in the very early stage (1991) of the TACIS and PHARE programmes for nuclear safety as a main topic.

The general objective of the project was to provide complementary data for the validation of the thermal annealing efficiency, as well as the embrittlement and re-embrittlement rate and kinetics of the VVER 440/230 RPV core materials.

The verification of the significance of direct characterisation on boat samples, taken from the inner surface of the non-cladded RPV wall, and the harmfulness of the remaining cut-outs can be stressed out as a second main objective. Beside the investigations on trepans taken from Novovoronezh unit’s 2 RPV, the direct characterisation of the operating Russian VVER 440/230 RPVs shall be seen as another main objective for the description of the most realistic situation of each plant, which is highly safety related.

According to the above-mentioned objectives the work plan was split into 7 main tasks, each of them related with the main objectives. Additional investigations were proposed under task 8 in order to assess the most critical core weld and to judge the quality of the neutron transport calculations.
KOLA 1 & 2 were not sampled in the frame of the project. These NPPs remain the only non-cladded VVER 440/230 units, which have not been investigated for that purpose.

Result

The contractor considers that situation as not satisfactory, regarding the related safety concern. Even if the materials properties are foreseen as showing higher toughness values, long term-operation should be based on specific measurements. Therefore, the recommendation for doing this sampling remains as a result of the project.

The feasibility of sampling at the inner surface of the non-cladded RPVs is demonstrated on the basis of local stress analyses. This allows systematic sampling of these RPVs at an extent, which appears comfortable for an extensive “direct characterisation”, at least for RPVs whose wall thickness is in compliance with the required dimensions.

Based on the chemical composition results, it can be assumed that the samples, which will be taken at the inner surface, are representative, indeed even conservative in terms of neutron embrittlement sensitivity. Therefore these samples are recognised to be very valuable for assessing safety cases.

A qualification programme of the Russian testing equipment was successfully implemented at the beginning of the project. It has provided for improvements on equipment (instrumentation of pendulums), on testing procedures, with respect to the international Q-A standards. A complementary round robin exercise, involving 4 laboratories (Siemens KWU, EDF, RRC Kurchatov Institute, CRISM Prometey) has been performed. It demonstrates that the conventional tests (tensile and Charpy V impact tests) are done in equivalent manner. For fracture mechanics testing a good co-operation between the main contractor and the local sub-contractor provided for assistance to the Russian laboratories in making their personnel familiar with the experimental procedures and evaluation techniques.

Sampling and testing of Novovoronezh 3 and 4 RPVs have been successfully performed. Tensile test results are looking consistent. Charpy V impact tests have been carried out at weld Nr 4 and 5 as well as on base metal. Prediction according to the Russian procedure appears to be rather conservative, at least for weld Nr 4, where the comparison was performed in detail. These results suggest that the re-embrittlement rate is appropriately addressed using the lateral shift. These analyses do not take into account any provision for uncertainties. The available results show evidence on the fact that weld Nr 4 is to be considered as the leading case for rest lifetime assessment. This statement is valid since the loads at welds Nr 4 and 5 are recognised to be at similar level.

The investigations performed at Novovoronezh 2 RPV trepans are unique and therefore valuable, even if the results are not as conclusive as expected in some areas. Therefore, a list of recommendations has been issued to address the main topics.
The chemical analyses were confirming the homogeneity of the base metal, whereas the weld showed a more heterogeneous picture, mainly do to the presence of the root. Consequently, that part of the samples was not considered for the investigations.

An extensive microstructure investigation programme has been carried out, involving 5 laboratories (Siemens KWU, EDF, AEA Technology, RRC Kurchatov Institute, CRISM Prometey). The metallographic examinations were almost confirming the expected microstructures and confirmed the particular sensitivity to grain boundary segregation of the weld metal. The advanced investigation techniques have been able by combining the various possibilities to identify radiation induced segregation and dislocation loops. The results are quite consistent with expected individual mechanisms and qualitatively of very high importance. Indeed, at this level of knowledge, it is not possible to account for physically based model proposals. Therefore, it is recommended to pursue these investigations in order to identify the relevant embrittlement mechanisms.

The mechanical tests performed in that tasks produced a broad quantity of data, whose consistency could not always be reached. The high neutron dose accumulated by the RPV at real irradiation conditions (up to 6,7 x 10-19 n/cm2) is expected to be the cause of some inconsistencies. This may be the case for the Charpy V results, which show at received specimens (AR) that the lateral expansion criteria is appearing dominant. This may be part of the explanation of the fact that the statistically based law, correlating sub-size and standard Charpy V results were found questionable in general and not conservative enough for high irradiated weld material. Two other inconsistencies shall be addressed at that stage on tensile results and trough wall-thickness Charpy V results.

Some tensile results show unexpected effects of annealing, mostly on base metal, which should be clarified later. On the other hand, the trough-wall Charpy V and neutron doses gradients were found inconsistent in many cases, and dependant on the leading criteria.

Therefore, the consistency of sub-size and standard Charpy V results was satisfactory in only a few cases. Fortunately, this appeared to be the case for weld material, at least at the inner surface, when all criteria are taken into account. This does compensate for any further validation. In the meantime, one should recommend to include a provision for compensating uncertainties in the safety analyses.
Concerning the neutron doses evaluation, it has been established that the calculated values are in good coincidence with the experimental ones, provided by RRC KI, at list at the RPV wall in front of the active core. Nevertheless, values evaluated by Siemens KWU, using the activity measurements provided by RRC KI for spectrum adjustment, are significantly higher than the calculated ones. Considering the results available it was decided to rely on the experimental neutron doses provided by RRC KI, which appear conservative for the evaluation of the irradiated material properties and accurate within the +/-20% scatter band, which is considered as technically reasonable.

The following recommendations have been issued in the project final report:

1. Sample the core weld of KOLA 1 & 2 RPVs, which was originally planned in the frame of this contract;
2. Re-irradiate spare metal from the core weld of Novovoronezh 3 & 4 RPVs;
3. Analyse the available data on the annealing behaviour of 15Ch2МFА type steel;
4. Reconsider the correlation between sub-size and standard Charpy V results for base metal;
5. Reconsider the correlation between sub-size and standard Charpy V results for weld metal;
6. Investigate the results’ inconsistencies and the uncertainties;
7. Assess an adequate K1c reference curve by further experimental investigations;
8. Pursue advanced microstructure investigations.