FEATURES OF XS GENERATION FOR RADIAL REFLECTOR OF THE VVER REACTOR IN SERPENT MONTE CARLO CODE

Krystsina Usheva1, Semen Kuten2, Arkadiy Khruschinsky2, Leonid Babichev2

1 Belarusian State University, Belarus

2 Institute for Nuclear Problems, Belarusian State University, Belarus

3 Joint Institute for Power and Nuclear Research - Sosny, National Academy of Sciences of Belarus, Belarus

The DYN3D is three-dimensional dynamic code used to calculate the dynamics processes in the nuclear core with quadratic or hexagonal fuel assembly geometry [1]. It is successfully applied to conservative estimations and best estimate calculations in safety analyses. The computer code DYN3D can be coupled with the thermohydraulic code (eg ATHLET [2]) to apply to the whole spectrum of operational and transient accidents, small and intermediate leaks, up to large breaks of coolant loops at PWRs/VVERs and BWRs.

The neutron kinetic model is based on the solution of the three-dimensional two-group neutron diffusion equation by nodal expansion methods. To use the DYN3D for NPP safety analysis it is necessary to generate the constants library for each fuel loading of the NPP reactor.

Serpent Monte Carlo reactor physics code [3] can be used as tools for generation of two-group constants library for DYN3D code. For generation of constants library using Serpent code it is necessary to carry out a very large number of simulations. To obtain reliable results in a reasonable time, it can use a supercomputer and parallel computing.

Detailed model of the radial reflector for the VVER reactor core is developed for calculations in Serpent (version 2.1.26) Monte Carlo code. Five types of the radial reflector cells can be considered taking into account design features of the baffle of VVER reactor. Two-group cross sections have been calculated for the two-layer reflector of DYN3D full core model.

The XS library of the reflector in the full core model contains values of the following characteristics, which are calculated in the two-group approximation: transport cross section (${\Sigma_{tr1,2}}$), absorption cross section (${\Sigma_{abs1,2}}$), group-transfer cross section (${\Sigma_{s12}}$), reflector discontinuity factor (RDF).

For correct work of the DYN3D two-group version the fast-to-thermal transfer cross-sections (${\Sigma_{s12}}$) were adjusted taking into account thermal-to-fast transfer (the heating of the thermal neutrons) (${\Sigma_{s21}}$):

$$\Sigma'_{s12} = \Sigma_{s12} - \Sigma_{s21} \frac{\Phi_2}{\Phi_1} \tag{1}$$

Reflector discontinuity factors were calculated using additional correction which taking to account for the influence of nearby FAs (Kord S. Smith, PHYSOR 2016) [4] :

$$RDF = \frac{DF_{ref}}{Df_{fuel}}ADF \tag{2}$$

Testing of the reflector XS library has been performed using typical first loading of VVER core based on four different FA types [5]. The library has been prepared for radial reflector as well as for each of individual FAs using Serpent. The two-group cross sections were calculated taking into account discontinuity factors at the "FA-FA", "FA-reflector" boundaries (so-called ADF and RDF factors, respectively).

Verification of the two-group XS's for reflector was been carried out by comparing the results of the full core calculations in DYN3D and Serpent codes. The results obtained by using the last code were considered as reference. The resulting XS libraries have been used in DYN3D for calculating the criticality and normalized power distribution in the reactor core during operation at zero power.

The results of the DYN3D calculations with the Serpent cross section sets agree very well with those of the Serpent full core MC calculations. The difference in $k_{eff}$ is 10,7 pcm for detailed models of the reflector, respectively. Maximum difference in the calculation of normalized power distribution does not exceed 5% in the peripheral layer of FAs, its average value is less than 2%.

Detailed model gives good agreements between reference full core calculation in Serpent and DYN3D using two-group XS's. The XS library generation method presented above can be used for developing the model of the radial reflector for Belarusian NPP with VVER-1200 reactor core.


[1] Rohde U., Kliem S., Grundmann U., Baier S., Bilodid Y., Duerigen S., Fridman E., Gommlich A., Grahn A., Holt L., Kozmenkov Y., Mittag S. The reactor dynamics code DYN3D – models, validation and applications / Progress in Nuclear Energy, 2016. №89. P. 170-190

[2] G.Lerchl, H.Austregesilo: The ATHLET Code Documentation Package, GRS-P-1 / Vol.1.: User's Manual, October, 1995;http://www.grs.de/en/computer-code-athlet

[3] Leppänen J. Serpent - a Continuous-energy Monte Carlo Reactor Physics Burnup Calculation Code. User's Manual / VTT Technical Research Centre of Finland, 2015. 164 p.

[4] Smith K.S. Nodal diffusion methods: understanding numerous unpublished details / PHYSOR 2016, Sun Valley, ID, USA, May 1-5, 2016. P. 1227-1241

[5] In-Core fuel Management Code Package Validation for WWERs. IAEA-TECDOC-847, 1995. 156 p.