Testing of Neutron Data Libraries in Application to Reactor Pressure Vessel Dosimetry

The fast neutron induced embrittlement of the reactor pressure vessel (RPV) is the main cause limiting the lifetime of many types of nuclear reactors, especially of the VVER type, produc-ing most of the nuclear energy in Russia and seven other countries. Uncertainty estimations show that the largest contribution to the uncertainty of calculated fluences is the uncertainty of the neutron cross section data. In the present work, different aspects of uncertainties of neutron fluence determination connected with the neutron data for transport calculations, data preparation methods and activation dosimetry cross sections have been investigated by com-parison of calculations with data from different libraries and by comparison of calculated and measured activation rates. The main object of these investigations was the VVER-1000 type reactor Balakovo-3. The tested libraries were the ENDF/B-VI based 47n/20g-group library BUGLE-96, the Russian 299n/15g-group library ABBN-93 (v. 99.01) and two modifications of the ABBN library: ABBN/B-VI and ABBN/JEF-2.2. In these modifications the cross sec-tion data of Fe, Cr and Ni are replaced by data based on the evaluated nuclear data libraries ENDF/B-VI and JEF-2.2, respectively. This allows to investigate the influence of the data of the elements most critical for the RPV fluence. Although thermal neutrons contribute only negligible to the embrittlement they are a strong source of gamma irradiation, which is proba-bly contributing noticeable to RPV embrittlement, as newer results showed. Therefore, the influence of different treatment of neutron data in the thermal energy region was investigated by calculations with and without upscattering in the region below 5 eV, as well as with and without consideration of core heterogenity effects. The neutron transport calculations were performed by the discrete ordinate code DORT using a 3D synthesis of (r-*)-, (r-z)- and r-calculations. Results obtained with the Monte Carlo codes MCNP and TRAMO with ENDF/B-VI data, were used to validate the DORT-results. The results of calculations at ex-vessel positions were compared with reference data of the Interlaboratory Experiment at Balakovo-3. The best agreement between results of calculations (C) and experiments (E) was obtained for BUGLE-96 and ABBN/B-VI, and 3D TRAMO results for all used reactions: 237Np(n,f), 93Nb(n,n'), 238U(n,f), 58Ni(n,p), 54Fe(n,p), 46Ti(n,p) and 63Cu(n,a). Further, the in-fluence of using different dosimetry cross sections, from IRDF-90v2, JENDL/D-99, RRDF-98 and a new Russian evaluation for 237Np(n,f), was investigated. The systematic calcula-tional underestimation of 237Np(n,f) reaction rates in case of IRDF-90v2, observed also by other investigators, can be removed using JENDL/D-99 or the new Russian evaluation. A similar underestimation for 93Nb(n,n') is reduced largely if JENDL/D-99 data are used. To evaluate the influence of cross sections of individual elements and of water on neutron and gamma fluence calculations, a simple two-zone cylindrical model was defined. The inner zone simulates the reactor core, the outer consists of H2O or Fe or Cr or Ni. In addition to DORT calculations with different libraries MCNP calculations with ENDF/B-VI were per-formed, to test the influence of point-wise data representation. The found discrepancies should be investigated further.