Volume 78, Issue 4

This paper presents the current status of works performed to date to further upgrade fuel assemblies and fuel cycle of WWER-440 reactors, and also considers some innovations – both envisaged and implemented in practice – intended to improve their neutronic and operating parameters. Reactors were assumed to operate at 100% power in a steady-state refueling mode. Trial operation of third-generation fuel assemblies (FA-3), including a comparison between some calculated and measured results, is described as an example illustrating these new developments.

Problems related to use of Erbium as burnable poison for VVER are discussed. Comparison is made between neutronics characteristics of Uranium-Gadolinium and Uranium-Erbium fuel cycles. The study shows that use of Erbium as burnable poison allows decreasing the peaking factor in the core. Meanwhile residual Erbium at the end of the fuel cycle makes it necessary to increase fuel enrichment. There is made the conclusion of prospects of using Erbium as burnable poison for VVER.

The isotopic composition of the control rods absorber changes under the neutron flux influence, so the control rods efficiency can decrease. In the VVER-1000 control rods boron carbide and dysprosium titanate are used as absorbing materials. In boric part the efficiency decreases due to the 10 B isotope burn-up. Dysprosium isotopes turn into other absorbing isotopes, so the absorbing properties of dysprosium part decrease to a lesser degree. Also the control rod's shells may be deformed as a consequence of boron carbide radiation swelling. This fact should be considered in substantiation of control rods durability. For the estimation of the control rods absorber burn-up two models are developed: VVER-1000 3-D fuel assembly with control rods partially immersed (imitation of the control rods operation in the working group) and VVER-1000 3-D fuel assembly with control rods, located at the upper limit switch (imitation of the control rods operation in groups of the emergency shutdown system).

The researches on the increase of achievable fuel burn-up in light water reactors (LWR) with so-called “spectrum shift” have been carried out in the world for a long time. The main idea of the “spectrum shift” is based on neutron spectrum shifting from the resonance energy region at the beginning of the cycle to the thermal region at the end of the cycle. There are many different ways of such regulation in the core – starting from coolant density variation during reactor cycle, to changing water-uranium ratio with some mechanical equipment. Spectrum shift regulation with depleted uranium or zirconium displacers is presented in this paper for uranium and mixed uranium and plutonium oxide (MOX) fuel loadings. New fuel assemblies with special guide tubes for displacers were developed for these purposes.

An automatic loading pattern optimization tool called ALPOT has been developed for Loviisa VVER-440 reactors. The ALPOT code utilizes combination of three different optimization methods. The first method is the imitation of the equilibrium pattern that is the optimized pattern in case the cycle length and the operation conditions are constant and the same shuffling pattern is repeated from cycle to cycle. In practice, the algorithm imitates assemblies’ operation year distribution of the equilibrium pattern stochastically. The function of the imitation algorithm is to provide initial patterns quickly for the next optimization phase, which is performed either with the stochastic guided binary search algorithm or the deterministic burnup kernel method depending on the choice of the user. The former is a modified version of the standard binary search. The standard version goes through all possible swaps of the assemblies and chooses the best swap at each iteration round. The guided version chooses one assembly, tries to swap it with every other possible assembly and performs the best swap at each iteration round. The search is guided so that the algorithm chooses the assemblies at or near the most restrictive fuel assembly first. The kernel method creates burnup kernel functions to estimate burnup variations that are required to achieve desired changes in the power distribution of the reactor. The idea of the kernel method is first determine the optimal burnup distribution that minimizes the maximum relative assembly power using the created kernel functions and a common solver routine. Then, the burnups of the available fuel assemblies are matched with the obtained burnup distribution.

In this paper the uncertainties of the neutronic calculations at core level – originating from the uncertainties of the basic nuclear data – are presented. The investigations have been made for a VVER-1000 core (Kozloduy-6) defined in the frame of the OECD NEA UAM benchmark. In the first part of the paper, the uncertainties of the effective multiplication factor, the assembly-wise radial power distribution, the axial power distribution and the rod worth are shown. After that the preliminary evaluation of the uncertainties of the neutron kinetic calculations are presented for a rod movement transient at HZP (Hot Zero Power) state, where the uncertainties of the time dependent core and assembly powers and the dynamic reactivity were evaluated. In both cases, we will see that the most important quantities –at core level and at HZP state – have a considerable uncertainty which is originating from the uncertainties of the basic cross section library in these investigations.

The reactor dynamics code DYN3D is a three-dimensional best-estimate tool for simulating steady states and transients of light-water reactors and innovative reactor designs. An overview of the DYN3D features is provided. This paper further focuses on the recently developed trigonal-geometry diffusion model DYN3D-TRIDIF including a description of the underlying nodal approach and the characteristics of trigonal geometries. Via a mesh refinement study by means of a VVER-1000-type core benchmark using a fine-mesh diffusion reference solution, DYN3D-TRIDIF shows spatial convergence. Furthermore, the performance of DYN3D-TRIDIF is verified by means of a single-assembly problem on pin-cell level. Good agreement between DYN3D-TRIDIF and the detailed-geometry transport reference is achieved with an average deviation in power of less than 1%.

Computer calculations important for safe operation of nuclear reactors are inherently affected by uncertainties. These originate mostly in numerical method applied, the calculated system model developed by the user, and finally there are fundamental uncertainties in nuclear data. This analysis is aimed at comparison of uncertainties in nuclear data affecting model of VVER-1000 fuel assembly containing either gadolinium or erbium as a burnable absorber using TSUNAMI-2D. This tool for quantitative determination of nuclear data uncertainties is being developed in Oak Ridge National Laboratory (USA). This tool works with 1D and 2D deterministic calculations and 3D Monte-Carlo calculations and processes the results. Output of the calculation is a complete characterization of the solved system with an overall uncertainty of multiplication factor determination. Furthermore, different system responses can be defined (e.g. pin power peaking) and their sensitivity to nuclear data in dependence on neutron energy is determined.

During reactor start-up and operation, one of the most important parameters, which determine reactor safety, are the effectiveness of reactor shutdown system and the worth of working group control rods. The assessment of the calculation uncertainty of these functional is also important for the design of reactor loadings. The paper considers methods for estimating uncertainties by comparing calculated and measured data. The paper shows that the distribution of the deviations of calculated (BIPR-7A) and measured data is close to normal, and the characteristics of the distribution (average and standard deviations), which are determined for a final sample, are close to the parameters. It was shown, that the worth and effectiveness of control rod groups can be assessed by using the characteristics of finite samples.

Full-scale three-dimensional computer model describing neutron-physical processes that occur in the VVER-1200 reactor core involves detailed FA geometry description. However, the detailed description significantly complicates input data, increases calculation time and error probability. Consequently, creation of simplified models of the design elements becomes necessary. In present work, simplified geometrical models of FA design elements are suggested. Calculations of simplified models are compared to calculations of exact models that completely correspond to the working general view drafts provided by OKB “Hydropress”. It is demonstrated that in many problems the use of simplified description of FA elements geometry is possible.

This paper considers two interrelated issues and – proceeding from the conditional probability concept and Bayes’ statistical theorem – suggests a universal method to process measured data with account of prior information available. To assure reliable reactor operation, at the design stage it is suggested to account for errors of subsequent control measurements while assessing the engineering margin factors for reactor parameters to be measured.

Accounting of feedbacks on the density and temperature of the coolant, the fuel temperature, as well as on the equilibrium concentration of xenon is important for VVER reactors, since it has a significant influence on the neutron-physical characteristics of such systems. Therefore, there is a need to incorporate feedbacks into precision calculations of the power density of the VVER-1000 core. In the present work, we used a model of the core, divided into 30 layers by height. The calculations were carried out by means of Russian Monte Carlo code MCU (Monte-Carlo Universal). Individual values of densities and temperatures of the coolant, the fuel temperatures, as well as the equilibrium concentration of xenon were calculated in each layer of every fuel assembly. In addition, power density was calculated pin-by-pin in each layer. The obtained results were compared to the results obtained by means of the certified BIPR-7A code. It is shown that within a reasonable computational time using currently available hardware resources one can obtain the three-dimensional power density with an acceptable accuracy and taking into account the feedbacks on the density and temperature of the coolant, the fuel temperature, as well as the equilibrium concentration of xenon.

A CFD model for the head part of the VVER-440 fuel assemblies with Gd burnable poison has been developed. Using this model calculations are performed for some recent Gd assemblies of the Paks NPP with the code ANSYS CFX. The temperature distributions in the assembly heads are determined and the in-core thermocouple signals are evaluated. The computed signals are compared to the in-core measured data of Paks NPP. The study shows that the deviations between the in-core thermocouple signals and the cross-sectional average temperatures at the measurement level depend on the burn-ups of the fuel assemblies. These deviations are significant in the cases of low burn-ups and negligible in the cases of higher burn-ups. This behavior is caused by the depletion of Gd from the three fuel rods near the assembly center.

Usually, reactor safety analyses are based on the traditional conservative deterministic approach. An alternative approach to reactor safety analyses uses best estimate computer codes in combination with quantification of uncertainties in model and plant parameters. The German Gesellschaft für Anlagen- und Reaktorsicherheit (GRS) has applied the methodology of uncertainty analysis based on the Wilks’ approach. According to this approach, the number of calculations required to perform a probabilistic estimation of an output parameter with a given coverage/confidence level (e.g., 95%/95%) does not depend on the number of input uncertainty parameters. In this paper, the results of an uncertainty analysis for a Large Break LOCA scenario for a generic German PWR of the Konvoi design are reported. All calculations were performed using the system code ATHLET Mod 2.2 Cycle A. The EXCEL-integrated software SUSA developed by GRS was used for generation of the uncertainty parameter vectors, their incorporation into ATHLET input files and for statistical evaluation of the results.

Severe accident – if no mitigation action is taken – leads to core melt. An effective severe accident management strategy can be the external reactor pressure vessel cooling for corium localization and stabilization. For some time discussion was going on, whether the in-vessel retention can be applied for the VVER-440 type reactors. It had to be demonstrated that the available space between the reactor vessel and biological protection allows sufficient cooling to keep the melted core in the vessel, without the reactor pressure vessel losing its integrity. In order to demonstrate the feasibility of the concept an experimental facility was realized in Hungary. The facility called Cooling Effectiveness on the Reactor External Surface (CERES) is modeling the vessel external surface and the biological protection of Paks NPP. A model of the CERES facility for the ATHLET TH system code was developed. The results of the ATHLET calculation agree well with the measurements showing that the vessel cooling can be insured for a long time in a VVER-440 reactor.

Fukushima related questions are currently being addressed at Gesellschaft für Anlagen- und Reaktorsicherheit (GRS) mbH within several research projects funded by the Federal Ministry for Environment, Nature Conservation and Nuclear Safety (BMU) and Federal Ministry of Economics (BMWi). In the following section first results of selected issues are presented.