Volume 77, Issue 4

This paper is dedicated to the latest version of TVS-M code – TVS-M 2007, which allows the neutron flux distribution inside fuel assemblies to be calculated without using the diffusion approximation. The new spatial calculation module PERST introduced in TVS-M code is based on the first collisions probability method and allows the scattering anisotropy to be accounted for. This paper presents some preliminary results calculated with the use of the new version of TVS-M code.

The experiments performed at the ZR-6 zero power critical reactor by the Temporary International Collective (TIC) and a burnup benchmark specified for depletion calculation of a VVER-440 assembly containing Gd burnable poison were used to qualify the APOLLO2.8-3.E (APOLLO2) code as a part of its ongoing validation activity. The work is part of the NURISP project, where KFKI AEKI undertook to develop and qualify some calculation schemes for hexagonal problems. Concerning the ZR-6 measurements, single cell, macro-cell and 2D calculations of selected regular and perturbed experiments are used for the validation. In the 2D cases, the radial leakage is also taken into account by the axial leakage represented by the measured axial buckling. Criticality parameter and reaction rate comparisons are presented. Although various sets of the experiments have been selected for the validation, good agreement of the measured and calculated parameters could be found by using the various options offered by APOLLO2. An additional mathematical benchmark – presented in the paper – also attests for the reliability of APOLLO2. All the test results prove the reliability of APOLLO2 for VVER core calculations.

DYN3D is a three-dimensional nodal code for steady-state and transient analyses of Light Water Reactors applicable for square and hexagonal fuel assembly geometries. Several versions of the DYN3D code are available including a multigroup diffusion and a simplified P3 (SP3) neutron transport option. The multi-group SP3 method based on a trigonal geometry was developed recently. This method is applicable to the analysis of reactor cores with hexagonal fuel assemblies and allows flexible mesh refinement. In this paper, the theoretical background for the SP3 method is briefly described. The consistency of the implementation of the trigonal SP3 methodology in DYN3D is demonstrated by means of a simplified VVER-440 core test example. The corresponding few-group homogenized cross sections and reference solutions were produced by the Monte Carlo code Serpent. The DYN3D results are in good agreement with the reference solutions.

A new analytical method is described to deal with the Leakage Environmental Effect for the cross section production. The method is based on the one dimensional analytical solution of the two-group diffusion equation for two adjacent, homogenized fuel assemblies. The high quality of the results for this highly efficient method is demonstrated for square fuel assemblies. In additional tests the limiting case for the transferability of the concept to the small hexagonal VVER-440 type fuel assemblies is shown.

The multi-group version of the DYN3D reactor dynamics code was used for calculations for a new concept of a boiling water reactor with tight lattice of fuel rods and reduced neutron moderation. For that purpose, a 5-group cross section library was prepared and connected to the DYN3D code. Comparison calculations with the steady-state finite-difference code ACADEM showed a very good agreement. The capability of the DYN3D multi-group code in modeling transients in boiling water reactors with tight fuel element lattices was demonstrated by the analysis of two reactivity accidents initiated by the ejection of one control rod and unauthorized withdrawal of a control rod bank from the reactor core. The corresponding analyses were performed for begin of cycle conditions, when the considered control rods are at their maximum insertion depth.

RUTA-70 model for simulations with the internally coupled codes DYN3D/ATHLET and DYN3D/RELAP5 was developed. A 3-D power distribution in the core is calculated by DYN3D with thermal-hydraulic feedback from the system codes. A steady-state corresponding to the full reactor power and an accident scenario initiated by failure of all primary coolant pumps were simulated with the DYN3D/ATHLET and DYN3D/RELAP5 coupled code systems to verify these codes. The compared coupled codes give close predictions for the initial and final states of the simulated accident but not for the transition between them. The observed deviations are explained by differences in the subcooled boiling models of the employed versions of ATHLET and RELAP5. Nevertheless, both simulations confirm a high level of the reactor inherent safety. The allowed safety margins were not reached.

A CFD model has been developed to investigate the coolant flow in the central tube of VVER-440 fuel assemblies. The model has been validated with measured data of Kurchatov Institute. With this model a peripheral and an inner fuel assembly of the core have been investigated and the mass flux and heatup ratios of the central tube flow have been determined. Based on this study, the outlet mass flux of the tube is 1.2 times higher than the inlet mass flux of the rod bundle and heat-up in the tube is 0.35 times lower than the heat-up in the rod bundle. The ratios are not sensitive to the operational conditions within the scope of these investigations. The results of these simulations can be used as boundary conditions for the central tube in the assembly head calculations.

In Loviisa VVER-440-type NPP the coolant outlet temperature of the hot subchannel is constantly monitored during the operation. According to the authority requirement the maximum subchannel outlet temperature must not exceed the saturation temperature. Coolant temperature distribution inside the fuel assembly is affected by the efficiency of the coolant mixing. In order to enhance the coolant mixing the fuel manufacturer is introducing the additional mixing vanes on the fuel bundle spacer grids. In the paper the effect of the different mixing vane modifications is studied with computational fluid dynamics (CFD) simulation. Goal of the modelling is to find vane modifications with which sufficient mixing is reached with acceptable increase in the spacer grid pressure loss. The results of the studies are discussed in the paper.

The research field focussing on the investigations and the analyses of postulated severe accidents is an important part of the nuclear safety. To maintain the safety barriers as long as possible and to retain the radioactivity within the airtight premises of the containment, to avoid or mitigate the consequences of such events and to assess the risk, thorough studies are needed. This paper is focused on the possibilities for accident management measures in case of severe accidents. The analyses performed with the integral code ASTEC cover two accident sequences which could lead to a severe accident: a small break loss of coolant accident and a station blackout. The discussed issues concern the main phenomena during the early and late in-vessel phase of the accidents, the time to core heat-up, the hydrogen production, the mass of corium in the reactor pressure vessel lower plenum and the failure of the reactor pressure vessel. Additionally, possible operator's actions and countermeasures in the preventive or mitigative domain are addressed.

The few-group cross section libraries, used by reactor dynamics codes, are affected by the spectral history effect – a dependence of fuel cross sections not only on burnup, but also on local spectral conditions during burnup. Neglecting this effect leads to an additional component of error in neutron-physical characteristics. Two solution approaches to this problem implemented in the reactor dynamic code DYN3D are described and compared in this paper: a cross section correction method based on239Pu concentration and separate cross sections treatment for each axial layer of reactor core. Steady-state and burnup characteristics of a PWR and a WWER-1000 cores, calculated by DYN3D with and without cross section corrections, are compared. An impact of the correction on transient calculations is studied for a control rod ejection example. Studies have shown a significant influence of spectral history on axial power and burnup distributions as well as on calculated cycle length. Two different correction methods have shown similar major effects.

Measurement of rod efficiency is an important part of nuclear reactor safety. In case of operating power reactors it is usual to measure all rod efficiency at beginning of cycle hot zero power condition. Measurement of all rod efficiency is technically complicated because the measured “dynamic” reactivity cannot be compared directly to calculated “static” value of control rod efficiency. At Paks Nuclear Power Plant a new method was established to solve this problem. A static computer code used for reload safety analysis was extended to calculate core kinetics as well. Using the developed code the rod drop measurement itself is simulated and the time dependence of flux is determined. In addition, using Monte Carlo calculations weight functions are determined to describe the detector signal response to neutron flux changes in the reactor core. Using this combined system the detector signal itself can be simulated to the actual cycle and condition. Thus the measurement directly can be compared to simulation and the result can be evaluated. The method is under introduction to NPP Paks measurement practice and the first results – which are shown in the paper, are very promising.

This paper reports on an investigation into fuel design choices of a pressurized water reactor operating in a self-sustainable Th- 233 U fuel cycle. In order to evaluate feasibility of this concept, two types of fuel assembly lattices were considered: square and hexagonal. The hexagonal lattice may offer some advantages over the square one. For example, the fertile blanket fuel can be packed more tightly reducing the blanket volume fraction in the core and potentially allowing to achieve higher core average power density. The calculations were carried out with Monte-Carlo based BGCore code system and the results were compared to those obtained with Serpent Monte-Carlo code and deterministic transport code BOXER. One of the major design challenges associated with the SB concept is high power peaking due to the high concentration of fissile material in the seed region. The second objective of this work is to estimate the maximum achievable core power density by evaluation of limiting thermal hydraulic parameters. The analysis showed that both fuel assembly designs have a potential of achieving net breeding. Although hexagonal lattice was found to be somewhat more favorable because it allows achieving higher power density, while having breeding performance comparable to the square lattice case.

Reutilisation of fissile nuclides after reprocessing in light water reactor VVER-440 with inert matrix fuel (IMF) is analysed. Development of transmutation combined fuel assembly, containing two pin types – with metal molybdenum + PuO 2 mixture and with classical UO 2 – is described. Static characteristics of five-year VVER-440 fuel cycle with IMF are given. Results of control rod ejection accident analysis are presented. Necessary improvements of transmutation combined fuel assembly with IMF for the reactor VVER-440 are mentioned.