Band 80, Heft 4

Defining precisely the burnup of the nuclear fuel is important from the point of view of core design calculations, safety analyses, criticality calculations (e. g. burnup credit calculations), etc. This paper deals with the uncertainties of MULTICELL calculations obtained by the solution of the OECD NEA UAM PWR pin cell burnup benchmark. In this assessment Monte-Carlo type statistical analyses are applied and the energy dependent covariance matrices of the cross-sections are taken into account. Additionally, the impact of the uncertainties of the fission yields is also considered. The target quantities are the burnup dependent uncertainties of the infinite multiplication factor, the two-group cross-sections, the reaction rates and the number densities of some isotopes up to the burnup of 60 MWd/kgU. In the paper the burnup dependent tendencies of the corresponding uncertainties and their sources are analyzed.

A brief description of the development history of engineering computer codes for VVER reactor calculations, which have been developed in the VVER physics department of NRC Kurchatov institute, since the end of 1950 s till now, are given in the report. The modern status of the basic codes of the Kurchatov institute for VVER fuel loadings design and safety analysis (TVS-M, KASKAD-BIPR-PERMAK, BIPR8, PERMAK-3D, SC1, TOPRA, ATHLET/BIPR-VVER, TIGR-1, etc.) is characterized from various points of view. Some specific problems which are shown during designing of fuel cycles, optimisation of fuel and technologies will be commented. They demand decisions within the complex of programs and, accordingly, demand the further development of a complex or working out and inclusion in a complex of new programs.

Up to now core calculations with Monte Carlo provided only average cross-sections of mesh cells for further use either in finite difference calculations or as benchmark ones for approximate spectral algorithms. Now MCU code is capable to handle functions, which may be interpreted as average diffusion coefficients. Subsequently the results of finite difference calculations with cells characteristic sets obtained in such a way can be compared with Monte Carlo results as benchmarks, giving reliable information on quality of production code under consideration. As an example of such analysis, the results of mesh calculations with 1-, 2-, 4-, 8- and 12 neutron groups of some model VVER fuel assembly are presented in comparison with the exact Monte Carlo solution. As a second example, an analysis is presented of water gap approximate enlargement between fuel assemblies, allowing VVER core region be covered by regular mesh.

Two types of calculations were made to compare BIPR-7A and MCU results for 3D full-scale models. First EPS (emergency protection system) efficiency and in-core power distributions were analyzed for an equilibrium fuel load of VVER-1000 assuming its operation within an 18-month cycle. Computations were performed without feedbacks and with fuel burnup distributed over the core. After 3D infinite lattices of full-scale VVER-1000 fuel assemblies (FA's) with uranium fuel 4.4 % enrichment and uranium-erbium fuel 4.4 % enrichment and Er 2 O 3 1 % wt were considered. Computations were performed with feedbacks and fuel burnup at the constant power level. For different time moments effective multiplication factor and power distribution were obtained. EPS efficiency and reactivity effects at chosen time moments were analyzed.

The different VVER-440 core fuel loadings subcriticality evaluations are performed during the start-up process by boron dilution or control assembly withdrawn by macrocode MOBY-DICK calculations. The dynamic reactivity and quasicritical reactivity are compared and sensitivity of reactivity prediction at the low boundary of start-up interval ( ρ = −0,01) has been provided on the basis of different modelling of ionization chamber (IC) response calculation. Special attention is paid to the impact of power distribution and spontaneous fission distribution form factor on IC response correction during control assembly movement. Precision and robustness of different corrections of IC signal processing in real core start-up processed IC signals was evaluated.

The paper discusses experimental and numerical investigations on intensification of thermal and mass exchange which were performed by National Research Centre “Kurchatov Institute” over the past years. Because of enhancing the power of active power plants, intensification of thermal and mass exchange investigations are performed by a number of Russian research and design organizations. Recently, many designs of heat mass transfer intensifier grids have been proposed. NRC “Kurchatov Institute” has accomplished a large scope of experimental investigations to study efficiency of intensifier grids of various types. The outcomes of experimental investigations can be used in verification of computational models and codes. On the basis of experimental data, we derived correlations to calculate coolant mixing and critical heat flux mixing in rod bundles equipped with intensifier grids. The acquired correlations were integrated in subchannel code SC-INT.

The Russian KORSAR/GP (hereinafter KORSAR) computer code was developed by a joint team from Alexandrov NITI and OKB “Gidropress” for VVER safety analysis and certified by the Rostechnadzor of Russia in 2009. The code functionality is based on a 1D two-fluid model for calculation of two-phase flows. A 3D CFD module in the KORSAR computer code is being developed by Alexandrov NITI for representing 3D effects in the downcomer and lower plenum during asymmetrical loop operation. The CFD module uses Cartesian grid method with cut cell approach. The paper presents a numerical algorithm for coupling 1D and 3D thermal-hydraulic modules in the KORSAR code. To ensure mass and energy balances at the interface, convective fluxes for 1D module are calculated as a sum of fluxes at corresponding boundary faces of 3D module. Mass balance condition is used to couple pressures in the modules. The combined pressure field is calculated by the multigrid method. The performance efficiency of the algorithm for coupling 1D and 3D modules was demonstrated by solving the benchmark problem of mixing cold and hot flows in a T-junction.

Nowadays the demands for accuracy of determining the weighted mean reactor power significantly increases due to implementing programs aimed at increasing installed power at nuclear power plants with WWER-1000 reactors. This accuracy strongly depends on the coolant temperature stratification in the primary circuit pipes, especially in the hot legs. Thus, the acute problem is to investigate the phenomenon on the basis of the full-scale tests experimental data and the calculated data of computational fluid dynamics simulations. The paper presents the full-scale experimental data on the coolant temperature distribution in the hot legs obtained via I&C systems, primarily via in-core monitoring system. These data were obtained at the pilot operation stage of Unit 4 Kalinin nuclear power plant in a stationary mode at nominal power level. To present the results of calculating simulation for the same mode CFD codes are used.

Homogeneous boron dilution scenarios in a VVER-440 reactor were analyzed using the coupled KIKO3D-ATHLET code. The scenarios are named “homogeneous” because of the very slow dilution caused by a rupture in the heat exchanger of the makeup system. In the case of the presented Beyond Design Basis Accident (BDBA) scenarios, it was supposed that after 30 min the operator doesn't stop the dilution, violating the prescriptions of the Limits and Conditions document. Without the presented analyses, a significant contribution of the homogeneous boron dilution to the Core Damage Frequency (CDF) had to be assumed in the Probabilistic Safety Analyses (PSA). According to the combined results of the presented deterministic and probabilistic analyses, the final conclusion is that boron dilution transients don't give significant contribution to the CDF for the investigated VVER-440 NPP.

A diluted water plug can form inside the primary coolant circuit if the coolant flow has stopped at least temporarily. The source of the clean water can be external or the fresh water can build up internally during boiling/condensing heat transfer mode, which can occur if the primary coolant inventory has decreased enough during an accident. If the flow restarts in the stagnant primary loop, the diluted water plug can enter the reactor core. During outages after the fresh fuel has been loaded and the temperature of the coolant is low, the dilution potential is the highest because the critical boron concentration is at the maximum. This paper examines the behaviour of the core as clean or diluted water plugs of different sizes enter the core during outages. The analysis were performed with the APROS 3D nodal core model of Loviisa VVER-440, which contains an own flow channel and 10 axial nodes for each fuel assembly. The wide-range cross section data was calculated with CASMO-4E. According to the results, the core can withstand even large pure water plugs without fuel failures on natural circulation. The analyses emphasize the importance of the simulation of the backflows inside the core when the reactor is on natural circulation.

A subcritical molten salt reactor is proposed for minor actinides (separated from spent fuel VVER-1000 light water reactor) incineration and for 233 U conversion from 232 Th. Here the subcritical molten salt reactor with fuel composition of heavy nuclide fluorides in molten LiF–NaF–KF salt and with external neutron source, based on 1 GeV proton accelerator and molten salt cooled tungsten target is considered. The paper presents the results of parametrical analysis of equilibrium nuclide composition of molten salt reactor with minor actinides feed in dependence of core dimensions, average neutron flux and external neutron source intensity. Reactor design is defined; requirements to external neutron source are posed; heavy nuclides equilibrium and fuel cycle main parameters are calculated.

A burn-up calculation of VVER's cores by Monte-Carlo code is complex process and requires large computational costs. This fact makes Monte-Carlo codes usage complicated for project and operating calculations. Previously prepared isotopic compositions are proposed to use for the Monte-Carlo code (MCU) calculations of different states of VVER's core with burnt fuel. Isotopic compositions are proposed to calculate by an approximation method. The approximation method is based on usage of a spectral functionality and reference isotopic compositions, that are calculated by engineering codes (TVS-M, PERMAK-A). The multiplication factors and power distributions of FA and VVER with infinite height are calculated in this work by the Monte-Carlo code MCU using earlier prepared isotopic compositions. The MCU calculation data were compared with the data which were obtained by engineering codes.

The effects of nuclear data covariance on important reactor parameters are investigated. The analyses are performed on the base of the OECD/NEA coolant transient Benchmark (K-3) on measured data at Kalinin-3 Nuclear Power Plant (NPP). For this purpose the GRS uncertainty and sensitivity software package XSUSA is applied to propagate uncertainties in nuclear data libraries to the full core coupled transient calculations. Moreover, based on the previous thermo-hydraulic studies a set of most important thermo-hydraulic parameters is chosen and added to the uncertain input vector. A statistically representative set of coupled ATHLET-PARCS code steady state calculations is analyzed and both integral and local output quantities are compared with the measurements available in the benchmark. The work is a step forward in establishing a “best-estimate calculations in combination with performing uncertainty analysis” methodology for coupled full core calculations.

Jacketless fuel assemblies change their form in the course of operation. Often they bow lengthwise. Primarily, these fuel assembly (FA) bows threaten to reduce the control rods' fall rate, but at the same time they change (e. g. increase) the amount of moderator in inter-assembly gaps, thus producing additional power surges. Gap sizes vary randomly and their impact is accounted for with the help of engineering margin factors. For VVER-1000, this account of engineering margin factors increases the fuel component of electricity generation cost by 3–5%, and a half of this increase is due to inter-assembly gap variations. This paper discusses the technique used to account for the impact produced by these gaps on fuel rod power; gives numerical values of sensitivity factors for power variations vs. gap sizes depending on the computational model assumed; and discusses the interference of gap effects and the account of power and coolant temperature feedbacks.