22nd Symposium of AER on VVER Reactor Physics and Reactor Safety

First the safety assessments made by Fortum due to Fukushima accident are discussed. These
include the EU stress test based on specifications defined by the European Nuclear Safety
Regulators Group. In addition, Finnish Ministry of the employment and the economy has
requested assessment, how Loviisa NPP can cope with natural events and loss of electrical
power. As a continuation of Fortum’s assessment to the ministry, Finnish Radiation and Nuclear
Safety Authority has sent a request for action plan to further improve the resilience of Loviisa
NPP against natural events and loss of electrical power. All these assessments have been
submitted to the authority in 2011. The methodology of the assessments is to use scenarios that
have been defined in a deterministic way. Multiple scenarios with different failure assumptions
are included.

Numerous fast reactor constructions have been appeared world-wide in pages of institutes?
reports. Most of them are benchmark exercise with complicated features. We shall examine 6
constructions: 5 sodium cooled and 1 lead cooled. Critical and dynamic parameters as well as
their burnup are calculated and compared with each other. The computational tool is composed
from NJOY, TRANSX, KENO and TIBSO programs.
Introduction
The basic composition of any fuel from the six cores is a typical composition of an LWR spent
fuel, that is the main fissionable isotope in each reactor is the Pu239. Thus the main goals of our
investigation are:
? To know whether a reactor can be operated with such composition
? To know whether the U238 left in the fuel can breed enough fissionables to maintain the
criticality
? To know whether these systems are applicable to incineration of higher actinides.
The examined six reactors are the following:
1. Sodium cooled fast breeder with thorium blanket (FBR) ¹
2. Lead cooled fast reactor (Pb) ²
3. Large carbide sodium cooled reactor (big-na) ³
4. Large oxide sodium cooled reactor (big-ox) ³
5. Medium metal sodium cooled reactor (midmet) ³
6. Medium oxide sodium cooled reactor (mid-ox) ³

The researches on the increase of achievable fuel burn-up in light water reactors with
so-called ?spectrum shift? had been carrying out in the world for a long time. The main idea
of the ?spectrum shift? is based on neutron spectrum shifting from the resonance 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
company, to changing water-uranium ratio with some mechanical equipment. Spectral shift
regulation with depleted uranium or zirconium displacers is presented in this paper for
uranium and MOX fuel loadings. New fuel assemblies with special guide tubes for displacers
were developed for these purposes.

The calculated assessment of the main neutron and thermo-physical characteristics of
VVER reactor with supercritical coolant pressure using the upgraded version of code package
SAPFIR&95RC has been performed. Calculation model of the VVER-SCP core is developed.
The version under consideration has a single-circuit coolant circulation system, heat up
amounts to more than 2500C, the power unit efficiency is more than 40 %.
Two versions of the reactor core lay-out were considered ? with a single-pass coolant
flow scheme and a two-pass coolant scheme. The design of standard VVER forms the basis of
VVER-SCP reactor core. There are differences ? fuel rod lattice is more tight, diameter of a
fuel rod is increased, jacketed FA is applied, MOX-fuel (mixture of the weapon-grade
plutonium and regenerated plutonium) is used.
The calculated analysis of maximum temperature of fuel, cladding, coolant, maximum
heat flux of a fuel rod and length of fuel cycle is carried out. Various versions of pressure
losses of throttling washers are considered.
VVER-SCP two-pass coolant core is divided into two parts of coolant flow with
approximately equal number of FA, they are a downcomer and a riser section. Such a
structure of the core will enable to decrease considerably a volumetric power peaking factor.
Fuel rod design and fuel matrix in the downcomer and riser section are similar.
The obtained neutron and thermo-physical characteristics of the core for a two-pass
coolant scheme (temperature of fuel, cladding, volumetric power peaking factor, maximum
linear heat flux of a fuel rod) are lower than in a single-pass coolant version. However, in
two-pass coolant scheme, coolant temperature coefficient of reactivity becomes positive in the
course of the fuel cycle unlike a single-pass coolant scheme. At the same time, a two-pass
coolant scheme does not realize the majority of the optimized solutions that have already been
used in a single-pass coolant scheme ? additional calculated analyses are needed to conduct
aimed at optimization of the core design.

Joint AER Working Group A on ?Improvement, extension and validation of
parameterized few-group libraries for VVER-440 and VVER-1000? and AER Working
Group B on ?Core design? twenty first meeting was hosted by VUJE, Inc in ModraHarm?nia, Slovakia during the period of April 16 to 18, 2012.
In total 13 participants from 7 member organizations took part in the meeting and 9
papers were presented.
Objectives of the meeting of WG A are: Issues connected with spectral calculations
and few-groups libraries preparation, their accuracy and validation.
Presentations were devoted to some aspects of transport and diffusion calculations,
and especially to the benchmark dealing with VVER-440 core periphery power tilt.
D. Sprinzl (co-authors V. Kr?sl, P. Mikol?? and J. ?varn?) presented ?Full-core
VVER-440 Pin Power Distribution Calculation Benchmark: MOBY-DICK Results?, Gy.
Hegyi (co-authors G. Hord?sy, Cs. Mar?czy and E. Temesv?ri) put ?Preliminary Solution of
the ‘VVER-440’ Pin Power Distribution Benchmark by KARATE and MCNP?, I. P?s (coauthors T. Park? and S. Patai-Szab?) provided information about ?Preliminary Results of
‘VVER-440’ benchmark by HELIOS/C-PORCA code?, S. Patai-Szab? spoke about ?Solution
of ‘FULL-CORE’ benchmark by HELIOS code? and G. Farka? (co-authors J. Ha???k, J.
L?ley, B. Vrban, M. Petriska, V. Sluge? and P. Urban) performed ?Critical Safety Analysis
of Spent Fuel Storage Pool for NPP Mochovce Using MCNP5?.
During the meeting were discussed questions connected with the benchmark solution
and, among others, future activities, which are also shortly described in the end of the paper.

THE RECOVERY PROCEDURE OF PIN-BY-PIN FIELDS OF POWER
DESTRIBUTION IN THE CORE OF VVER TYPE OF REACTOR FOR THE PROGRAM
BIPR-8. VERIFICATION CALCULATIONS.

Computer calculations are an important part of operation of any nuclear reactor. They are used
to predict properties of fuel assemblies during depletion in the core and they prove meeting of
all criteria for a safe operation. All calculations and their results are inevitably connected with
uncertainties. They can originate in numerical method applied, the calculated system model
developed by the user, and finally there are fundamental uncertainties in nuclear data. The
latter source of uncertainties is difficult to quantify and usually it is accounted for by some
conservative estimated increase of safety margins (by 5 % or more).
A new tool for quantitative determination of nuclear data uncertainties is being developed in
Oak Ridge National Laboratory. It is called TSUNAMI and it is distributed as a part of
SCALE package. This tool works with 1D and 2D deterministic calculations and 3D MonteCarlo 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 caused by nuclear data for the included nuclides. Contribution of each of the
nuclides is also calculated. 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.
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. Both fuel types have different conditions, e.g. burnable absorber
concentration and number of pins with burnable absorber, which influences uncertainties.
Also number of benchmark experiments for either of the elements will be compared and
resulting effect on calculation uncertainties will be stated.

Code complex for numerical analysis of demonstration gas cooled fast reactor ALLEGRO is
under preparation. Possible utilization of spectral code HELIOS 2.1 with broadened library is
analysed. Its verification starts by comparison with MCNPX 2.6.0. Calculations of
ALLEGRO subassemblies by both codes are presented in this paper. All of analyses in the
paper were inspired by an active participation at GoFastR project – component of FP7 EU.

At present time SSTC N&RS uses HELIOS code for few group cross section library
generation for WWER reactor cores calculations. There is actual problem of choice of
appropriate approach to implement cross section library into DYN3D code. Report provides a
survey of the used by SSTC approaches ? multidimensional table and polynomial
dependencies. Capabilities and possible extensions of each approach are described with
inherent advantages and disadvantages. Additionally the question of model development and
cross section preparation for WWER-1000 radial reflector with accounting of discontinuity
factors is covered. Brief results of calculations with the use of different approaches are
presented.
The work was performed in framework of orders SR 2511 and R0801504 (SR2611)
under German BMU support.

Because of difficulties with experimental validation of power distribution predicted by
macro-code on the pin level we decided to prepare calculation benchmark named ?FullCore?
for VVER-440. The submission of this benchmark was presented on the 21st Symposium of
AER in 2011.
The main task of this benchmark is to test the pin by pin power distribution in selected
fuel assemblies predicted by macro-codes that are used for neutron-physics calculations
especially for VVER-440 reactors.
The ?FullCore? benchmark is a 2D calculation benchmark based on the VVER-440
reactor core cold state geometry with taking into the account the geometry of explicit radial
reflector. The reference solution was calculated by MCNP code using Monte Carlo method
and sent to all participants. For comprehensiveness of reference results, the impact of different
Monte Carlo libraries and precision of calculation is studied.
In this paper we will in the first part present the results of this benchmark calculated
by MOBY-DICK macro-code.
In the second part a comparison of available macro-codes results of this benchmark
will be given.

Results of VVER-440 benchmark by HELIOS/C-PORCA code

DYN3D ² code was extended by capability of pin by pin power calculation. For
validation of given extension and prepared cross section library the calculations of pin power
distribution by DYN3D were performed for two AER benchmarks. First one ? ?MIDICORE?
WWER-1000 core periphery power distribution benchmark that was proposed on the 20th
Symposium of AER?¹. It?s a 2D calculation benchmark based on the WWER-1000 reactor
core cold state geometry with taking into account the geometry of explicit radial reflector.
HEXNEM1 and HEXNEM2 nodal methods are used in DYN3D for neutron flux calculation.
The second one ? AER benchmark for Khmelnitsky NPP, Unit 2, that was proposed by TUV
and SSTC NRC at 19th symposium ⁵. DYN3D calculations are compared with PERMAK
⁸ and DERAB [6, 7] data for FAs of 1st, 2nd, 3rd and 4th year operation.
The work was performed in framework of orders SR 2511 and R0801504 (SR2611)
under German BMU support.

Submitted paper is devoted to the problem of criticality for neutron transport equation.
Examination of the solution to this problem is going on in Lebesgue space L? and is based on
several main suppositions. These suppositions are sufficiently general and are in agreement
with spatial behaviour of both temperature and material atom densities as well as with
properties of all known models of neutron reactions with medium. The problem is
transformed onto the task to determine time asymptotic behaviour of solution to specific
initial-value one. The initial-value problem is solved numerically by Monte Carlo method.
and to this purpose, appropriate random process and random variable are constructed.

Accounting of feedbacks on the density and temperature of the coolant, as well as the
fuel temperature is important for VVER reactors, since it has a significant influence on the
neutron-physical characteristics of such systems. As a consequence, there is a need to
incorporate feedback into precision calculations of the power density of the VVER-1000 core.
In the present work the model of the core was used, divided into 10 layers by height.
The calculations were carried out by means of Russian Monte Carlo code MCU-PD.
Individual values of densities and temperatures of the coolant, as well as the fuel temperatures
were calculated in each layer of every fuel assembly. In addition, pin-to-pin power density
was calculated 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,
as well as the fuel temperature.

The isotopic composition of the control rods absorber is changing 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 of the
control rods the efficiency decreases due to the 10B isotope burn-up. In dysprosium part
dysprosium isotopes turn into other absorbing isotopes, so the absorbing properties of this part
decrease to a lesser degree.
For the estimation of the control rods absorber burn-up two models are developed:
VVER-1000 three-dimensional fuel assembly with control rods partially immersed (imitation
of the control rods? operation in the working group) and VVER-1000 three-dimensional fuel
assembly with control rods, located at the upper limit switch (imitation of the control rods?
operation in groups of the emergency shutdown system).

DYN3D is a three-dimensional nodal di?usion code for steady-state and transient analyses
of Light-Water Reactors with square and hexagonal fuel assembly geometries, applicable
also to innovative reactor concepts. Currently, several versions of the DYN3D code are
available including a multi-group di?usion and simpli?ed P3 (SP3 ) neutron transport
option.
In this work, both the di?usion and SP3 method in trigonal geometry are veri?ed by
means of ?ne-mesh homogenized reference solutions. Good agreement is observed for
eigenvalues, neutron ?ux and power distributions. Mesh re?nement studies show that the
trigonal nodal DYN3D methods converge well to the ?ne-mesh references. Furthermore,
an anisotropic-scattering benchmark problem was chosen to show the advantage of the
SP3 method over the di?usion approach.
1 INTRODUCTION
The reactor dynamics code DYN3D is a three-dimensional best-estimate tool for simulating steady states and transients of Light-Water Reactors (LWRs), which has been
developed at the Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Germany.? It comprises
a three-dimensional neutron-kinetics model based on nodal expansion methods, a thermalhydraulics core module with one- and two-phase coolant ?ow analysis capabilities, and a
fuel-rod model [1, 2]. Also burnup and poison-dynamics calculations can be performed
with DYN3D.
?

Several questions occur during the core design calculation of the Budapest Research Reactor
like keep the limit of the power peaking factor and set the neutron flux of the irradiation
channels inside a desired interval. These problems can be solved with an acceptable choice of
the load pattern of the nuclear fuel elements. To determine what kind of load patterns are
preferable for the above discussed requirements preliminary calculations have been made for
random configurations of load pattern. Large number of random configurations overlaps the
search space therefore the desired parameters as function of load pattern are examined. The
data processing of the previous calculations based on statistical tools like correlation which is
a good method for earn information about a large database with large number of variables.
The main deal of the calculations is finding roles for creating load pattern whose parameters
are inside the prescribed interval. The results will be used in the future for different
algorithms of optimization.

Review of AER Working Group B Meeting in Modra – Harm?nia, Slovakia is given.
Regular meeting of Core Design Group was organized by VUJE, Inc. and held at Modra pension Harm?nia, Slovakia, April 16?18, 2012, together with Working Group A. Presented
papers (see List of papers and List of participants) covered topics as follows.

This paper presents the current status of the development and VVER-440 fuel
assemblies (FA) modernization as well as VVER-440 nuclear fuel cycles update are
performed lately. The studied and the embedded innovations of the improved neutronics and
operational characteristics are coincided. Discussions were held in relation to the operation of
the reactor at 100% power in a steady state refueling. As an example of using the new
development we demonstrates a program of operation experience with RK-3 and some
numerical results are compared with measurements.
In spite of the fact that the VVER-440 has been present on the market for 40 years ago
it is still subjected to upgrading in order to improve the economic parameters of the fuel cycle
and increase the fuel reliability during its operation. In the process of VVER-440 gaining
operation experience ¹, numerous advancements of fuel assembly (FA) design
characteristics have been realized ². Achievement of these advancements required the
constant upgrading of existing FA designs to increase their service life, improve their
operational reliability and reactor safety and economics of fuel cycle (Table 1). A big effort
has been made in the reactor community to reach these goals. Below we show the effect of
these advancements and modifications on the neutronics and operational characteristics for
the initial and end of the modernization stages (Table 2 and Figure 1-6).
Now it is clear that improvement strategies refueling and development of the neutronic
codes helped to optimize the core load. This allowed to increase the fuel cycle economy and
to substantiate the overload (in-in-out) type scheme which provides to lower neutron leakage
from the core and to decrease the radiation load on the reactor vessel.
Advancement of the FA base design allowed to significantly improve fuel efficiency
by:
- using of zirconium spacer grids and reducing of FA shroud tube thickness resulting to the
increasing FA multiplication factor;
- grading of fuel enrichment in the cross-sectional plane of FA thus reducing a peaking
factor of linear fuel pin powers and ensuring by this a lightened realization of core loads
with reduced neutron leakage;
- increasing of initial enrichment of fuel feed and reducing of fresh fuel replaced during fuel
reloading;
- using of burnable gadolinium (Gd2O3) poison integrated with fuel ensuring desirable fuel
load arrangement and leading to improvement of load safety;
- raising the average fuel enrichment up to 4.87% and the application TVEGs;
- increasing the amount of fuel in the fuel elements by increasing the diameter of the fuel
pellets to 0,78 cm, excluding the central hole in tablets and replacing the cover on the nonshrouded design, increasing the spacing between fuel rods to 1,26 cm
- fuel assembly of the Third generation;

Presented results are contribution to the gradual evolution of VVER-440 fuel assembly.
Selected alternative variants of ?classical? assembly with the shroud and pin enrichment limit
5 % U235 (with one exception) are compared with recent assembly in exploitation – Gd2 with
mean enrichment 4,87% U235 (reference). Influence of in-assembly power peaking on
operation limit margins is evaluated.

Validation of macrocode ANDREA for VVER-440 reactors
Lenka Heraltova,
UJV Rez a.s., Husinec-Rez 130, 250 68 Rez, Czech Republic, email: hle@ujv.cz

CONSOLIDATED DATA FOR TASK 1 AND STATUS OF TASK2 OF THE
BENCHMARK FOR CORE BURNUP CALCULATIONS FOR A VVER-1000
REACTOR
T. L?tsch1), V. Khalimonchuk2), A. Kuchin2),
1): T?V S?D Industrie Service GmbH, Energy and Technology (IS-ET), Westendst. 199, 80686
Munich, Germany; 2): State Scientific and Technical Centre for Nuclear and Radiation Safety of
Ukraine (SSTC N&RS), Stusa st. 35-37, 03142 Kyiv, Ukraine

The paper discusses a probable significant global power field tilt caused by a random small
change in multiplying characteristics of fuel assemblies in a physically large reactor. The
paper proves that with random difference of multiplying characteristics in VVER-440 and
VVER-1000 cores, the power field tilt is higher in VVER-1000. The reaching of normal
power noticeably decreases the power field tilt, at that.

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.

An automatic loading pattern optimization tool called ALPOT was developed for Loviisa
VVER-440 reactors as a master’s thesis work. This paper discusses the methodology and the
performance of this tool.
The ALPOT code utilizes 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. The imitation utilizes heuristic rules but is also stochastic in nature. 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 search is guided so that the swapping of
fuel assemblies always starts at or near the most restrictive assembly. 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 neutronic analysis is always performed with the HEXBU-ELSI code package, which is
the standard 3D reactor core simulator at Fortum. Although, the tool was especially designed
for Loviisa reactors the basic ideas of the search algorithms could be utilized in any VVERreactors.
The performance of ALPOT was tested and validated extensively already in the development
phase. Furthermore, the tool has been used to search loading patterns for upcoming cycles of
Loviisa plants with good results. Depending on the strictness of the safety margins, finding a
feasible loading pattern takes from few minutes to couple of hours. The quality of the found
patterns is equal or better compared to the manually designed loading patterns.

One of the aims of the reactor physics group of SKODA JS, a.s. company is to develop a new
advanced optimization code called ATHENA. The code was presented last year and has
already been used for VVER type nuclear reactor fuel loading pattern of optimization but has
more general abilities. With this code an analysis of long cycles in VVER-440 reactor with
new Gd-2M based fuel was performed, using innovative Gd-2M+ and Gd-2X fuel, which are
currently under development by TVEL company as advanced fuel for VVER-440 reactors.
Optimization has allowed us not only to find long cycles, but also to evaluate the
effectiveness of new fuel types and to estimate options and feasibility of long cycles and
number of necessary fresh fuel assemblies.

Will be available later.

AER WORKING GROUP C ACTIVITY IN 2010

EXPERIENCES WITH APPLICATION OF GD-2N ASSEMBLIES IN PAKS NPP

Assembly power distribution Kqi and that of separate sections of assembly Kvi , j are
compared. The paper shows the different densities of distribution of calculated vs. verified
data both for all power density values and for separate ranges (for example, Kqi ? 1 or

Validation of VERETINA, a new nuclear reactor
analyzer system for VVER-440
J. P?les a, G. H?zi a,*, Cs. Horv?th a, J. V?gh a, I. P?s b, Z. K?lya b
a

The SCORPIO-VVER Core Monitoring System has proved since the first installation at
Dukovany NPP (CZ) in 1999 to be a valuable tool for the reactor operators and reactor
physicists. It is now operated on 4 units of Dukovany NPP and 2 units of Bohunice NPP (SK)
replacing the original Russian VK3 system. By both Czech and Slovak nuclear regulatory
bodies the SCORPIO-VVER was licensed as a Technical Specification Surveillance Tool.
Since it?s first installation, the development of SCORPIO-VVER system continues along with
the changes in VVER reactors operation. The system is being adapted according the utility
needs. With its flexible and modular framework successfully responses to the plant operating
needs and advances in nuclear fuel cycle strategies and fuel design. Modular framework
allows for easy modifications of the system and implementation of new methods in physical
modules. These facts have been confirmed by successful upgrades and more than 13 years of
reliable operation of the core monitoring system in two countries and six VVER-440 units.
This article describes the current status of the core surveillance system SCORPIO-VVER in at
Dukovany NPP in Czech Republic and at Bohunice NPP in Slovak Republic.

Method of accurate determination of VVER-440 reactor thermal power applying
ex-core detector signals

A coolant flow through the reactor is usually determined after annual outages at Slovak
NPP in two distinct ways. First method is determination on the basis of the secondary system
parameters – measurement of thermal balances. The value achieved by this method is used as
the input parameter in the Table of allowed reactor operation modes. The second method
draws from the primary system parameters – measurement of primary system hydraulic
characteristics.
Flow nozzles used for the measurement of feed water flow behind high pressure heaters
were replaced at both Bohunice Units during outages in 2008. The feed water flow behind
high pressure heaters is one of the main parameters used for the determination of coolant flow
through the reactor by the first method. Compared to the measurement executed during
previous fuel cycles, the calculated coolant flow through the reactor decreased considerably
after the change of flow nozzles. The imaginary change of coolant flow through the reactor at
Unit 3 was -1,6 %; and at Unit 4 -2,6 %. This change was not proved by the parallel
measurement of primary system hydraulic characteristics. Later it was found out that the
original flow nozzles used for 25 years were substantially deposited (original inner diameter
of the nozzles was reduced by about 0,6?0,9 mm). Therefore feed water flow measurement
was untrustworthy within the recent years.
On the findings stated above, Bohunice NPP has decided to increase coolant flow
through the reactor by changing the reactor coolant pumps impeller wheels. Modification of
impellers wheels is planned within years 2012 to 2014. The 4 impeller wheels at Unit 4 and 2
at Unit 3 were replaced during the outage in 2012. General contractor for Bohunice NPP is
VUJE, Inc. The supplier of impeller wheels is the Russian company CKBM St. Petersburg
through the Czech company SIGMA (as a subcontractors of VUJE, Inc.). The Division for
Preparation of NPP Operation is responsible for preparation, execution and evaluation of

In general a reactor is initially started up from a precondition stage (200?C) by
withdrawing control rods or by changing the boron acid concentration in primary circuit
coolant until the reactor is slightly supercritical, thus producing an exponentially increasing
neutron population on a very long period. As the neutron population increases, the fission
heating and thus the reactor temperature increases. This increase in temperature produces a
decrease in reactivity. That would lead to the slowdown of increasing the neutron population
and the reactor power should saturate at certain higher level. Generally, core loading patterns
are designed in such way that isothermal reactivity coefficient should already be negative at
the Cold Zero Power (CZP). For VVER 440 reactors CZP is equal to 200 ?C. To meet this
requirement could be problematic, particularly for the first core loading, when all fuel
assemblies are fresh. Due to the first criticality start-up of the Nuclear Power Plant (NPP)
Mochovce units 3 and 4 in the near future, detailed analyses of core parameters are required
by the Slovak Regulatory Authority to support safe operation of the nuclear facility. The
article introduces determination of the thermal reactivity coefficients, especially summary
(isothermal) and moderator (density) reactivity coefficients between 200?C and 260?C with
step of 2 ?C. The work presents calculated critical parameters, especially critical boron acid
concentrations at given coolant temperatures and position of the 6 th control assembly group.
Numerical iteration procedure was applied to calculate the critical parameters as a substitute
for a critical experiment. Geometrical and material models were created in compliance with
the reactor design and the first fuel loading of the NPP Mochovce 3 and 4. All mentioned
calculations were performed by computational code MCNP5 1.60 supported by NJOY99.364
microscopic sections processing system and our control scripts.

In design justification, considering a conservative approach, the lifetime of absorber
element (AE) is limited to 10 years, among them not more than 3 years as a part in the control
group. The main characteristics of radiation load of VVER-1000 AE affecting its lifetime are
boron burn-up and fast neutrons fluence (over 0,1 MeV) on the welded joint of AE end-piece.

The paper reports on Moscow final meeting of working group H ?Elaboration of
methodology of calculating the core design engineering factors?, which was held in
?Kurchatov Institute?, June 27-29. The meeting was attended by 18 physicists from Czech
Republic, Slovak Republic, Germany, Russia and the Ukraine.
Participants of working group ? coordinated the draft document which will be presented
on the twenty second symposium AER.
Name-list of the participants and titles of their presentations may be found in the paper,
which also gives brief information on the designation of presentations.

The macrocode qualification is key point of the redefinition of the engineering factor
methodological component. A hyperbolic dependence of the methodological component on
the power level is not strict requirement and from the point of some safety criteria, e.g. radial
fuel rod power limitation, can be simplified.
In the paper is presented development of the methodological component of volume
power distribution (engineering factor F0 ) like the prediction interval and trapezoidal width
of confidence of linear regression including simultaneous interval (Scheff? correction). The
adequate regression analysis was realized on comparisons of macrocode MOBY-DICK power
distribution with SPD measured power distributions for VVER core. There were determined
standard deviations calculated versus measured power distributions like time and spatially
varying uncertainty.
In reference to the above-mentioned standard deviations the true values of power
distribution were determined (according the Bayes? methodology) for nodal volume power
peaking K V . In this sense there were realized alternative evaluations of the methodological
components of engineering factor F0 .

In the design of VVER core loadings, the important parameters that characterise
reactor safety are the effectiveness of emergency protection and the worth of control rods. The
calculation uncertainties of those functionals are also important for the design. The paper
discusses the methods of accounting the uncertainties by comparing calculated and measured
data. The paper proves that the deviations between BIPR-7 calculated data and measured
data are distributed in accordance with close to normal law, while the parameters of
distribution (such as mean and standard deviations), which were determined for the available
comparison samples, are practically identical to those determined for the applicable general
samples.

Methodological ? q was calculated using two different sets of data. In the first case the
standard deviation was used of the differences between thermocouples (TC) temperature rises
calculated by the MOBY-DICK software and measured at the Dukovany NPP. In the second
case the standard deviation was used of the differences between relative fuel assembly power
values calculated directly by the MOBY-DICK software and reconstructed by the Scorpio
VVER measurement and control system. ? q calculated using TC temperature rises was 0.020
or lower for Units 1 ? 3 on Dukovany NPP a and 0.025 on Unit 4 when calculated in 60 deg.
geometry and 0.026 or lower for Units 1 ? 3 and 0.046 for Unit 4 when calculated in 360 deg.
geometry. Systematical deviation in average absolute differences between calculated and
measured TC temperature rises for the fuel assemblies 1 st year in the active zone was found.
The influence of the cleaning of the reactor shaft bottom and the throttling orifices on the
methodological ?q was analyzed for Unit 4 of Dukovany NPP.

Nowadays, there is a tendency to use best estimate plus uncertainty methods in the field of
nuclear energy. This implies the application of best estimate code systems and the
determination of the corresponding uncertainties. For the latter one an OECD NEA
benchmark was set up. The objective of the OECD NEA Uncertainty Analysis in BestEstimate Modeling (UAM) LWR benchmark is to determine the uncertainties of the coupled
reactor physics/thermal hydraulics LWR calculations at all stages.
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 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 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.
1. INTRODUCTION
In the frame of the OECD NEA UAM benchmark ¹ there has been a large activity to
determine the uncertainties of coupled reactor physics/thermal hydraulics LWR calculations
at all stages. In order to perform this large task, 3 phases were defined in the benchmark and
these phases include the survey of the uncertainties of the stand alone neutronics (Phase I), the
time dependent neutronics, stand alone thermo-hydraulics, the fuel behavior calculations
(Phase II) and the system phase, as well.
In this paper we concentrate on the determination of uncertainties of ?core physics? (Phase I,
Exercise I-3) and partially on the determination of uncertainties of the ?Time-Dependent
Neutronics? (Phase II, Exercise II-2). It is important to note that only the uncertainties of the

In the paper is represented the method of statistical account of uncertainties to obtain
estimation of DNBR and value of heat flux per rods length unit. The method is based on the
special developed FA model which includes mathematical models of energy field
calculations, FA thermomechanical deformations, thermohydraulics and technological
processes. FA thermohydraulics calculations carry out using cell by cell thermohydraulic code
SC-1 ¹. In these calculations were taken into account deviations in different parameters
Statistical properties of DNBR minimum value in FA and in specified rods, statistical
properties of heat flux from the surface of the fuel rods and coolant heating in specified cells
is determined using statistical simulation of large number of random parameters affecting this
values. It have been estimated the value of reducing factor for DNBR and this value was
analyzed for sensitivity to different factors: FA power, relative axial heat flux profiles, rods
energy distribution, the law of statistical distribution random parameters, the number of ?hot?
fuel assemblies. Also the estimation of engineering safety factors for coolant heating and local
heat flux have been obtained.

Summary on AER Group E Meeting 2012 held in conjunction with Seminar
focused on VVER PIE # 3958 data evaluation
Vladimir Chrapciak, VUJE, Inc., Slovakia, chrapciak@vuje.sk
L.Markov?, UJV Rez, Czech Republic, mar@nri.cz

Verification SCALE 6 and HELIOS for nuclide composition
calculation
Vladim?r Chrap?iak, Slovakia, VUJE a.s., vladimir.chrapciak@vuje.sk
Radoslav Zajac, Slovakia, VUJE a.s., radoslav.zajac@vuje.sk

The validation of depletion codes is an important task in spent fuel management,
especially for burnup credit application in criticality safety analysis of spent fuel facilities.
The calculations are performed for twelve samples with maximal, average and minimal
burnup from VVER-440 spent fuel assembly, based on #3958 ISTC PIE project. The fuel
assembly has initial enrichment 4.4% of

Criticality calculation for pools at reactor in Slovakia
Vladim?r Chrap?iak, Radoslav Zajac, Branislav Hatala
VUJE, Slovakia

Severe accident ? if no mitigation action is taken ? leads to core melt. The Fukushima
accident has triggered urgent discussions about the need for such type of investigation. 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.
The scaling ratio for reactor surface is 1:40 but the scale of the elevation is 1:1 to ensure
correct representation of the natural circulation processes. The heat flux is generated by
electric heaters. A large number of pressure, temperature, water level, void and flow
measurements allow tracking of the coolant behavior and surface temperature. In the
experiments, the area of the coolant channel can be adjusted in the critical section (where the
gap size is minimal).
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.

The paper gives the first comparison of the 7th three-dimensional AER dynamic benchmark
calculation results received with the codes RELAP5-3D? (?JV) and DYN3D/ATHLET
(HZDR). This benchmark was defined by A. Kotsarev, M. Lizorkin and R. Petrin at the 20th
AER Symposium in Hanassari (Finland). It is focused on investigation of transient behavior
in a VVER-440 nuclear power plant. Its initiating event is the opening of the main isolation
valve and re-connection of the loop with its main circulation pump in operation. A water slug
with lower temperature of 100 oC enters the core which leads to a reactor power excursion.
The activation of the reactor scram is caused by the corresponding power level signal. As a
result of the scram, the turbines are turned-off by closure of the turbine isolation valves. All
main circulation pumps remain in operation. Each participant had to use its own developed
input plant data deck and parameterized library of macroscopic cross sections. The VVER440/213 plant is at the end of the first fuel cycle and in hot full power conditions. The initial
reactor power tuning and burnup calculations were performed with the code DYN3D by both
participants. The first part of paper contains a brief description of RELAP5-3D?and
DYN3D/ATHLET system codes, their NPP input decks and reactor core models. The second
part shows the comparison of important global and local time dependent parameters. The
trends of main calculated parameters are very close. The calculated nuclear and thermalhydraulic phenomena like reactor power, cold leg mass flow rates, cold and hot leg coolant
temperatures, upper plenum pressure, pressurizer water level and steam generator powers
agree quite well. The differences in the time behavior of compared parameters are discussed
at the end of paper.

A new, multigroup version of the KIKO3D code allowing arbitrary number of energy
groups ? called KIKO3DMG ? was developed for calculation of the fast spectrum
reactors. From among the advanced new features, the possibility of using triangular
geometry and the automated cutting the triangles into further ones at many levels was
utilized for an attempt to reach a converged solution of the Second AER Kinetic
Benchmark Problem.
1. INTRODUCTION
From the 1990?s, several kinetic and dynamic benchmark problems have been defined
in the frame of the AER (Atomic Energy Research) cooperation. Most of them can be
classified into the group of the ?numerical exercise benchmarks? as far as
measurements are not supporting them and – at the same time – some modeling details
are entrusted to the participants. Although the most important ?integral? modeling
features are ?normalized? in the definitions (shut down margin, recriticality
temperature, advised thermal hydraulic correlations, etc.), in the strict sense the
abovementioned feature prevails starting from the third AER kinetic benchamark
concerning the not fully exact prescription of the cross sections or all the details of the
thermal hydraulic modeling. (An advantage of these characteristics must also
mentioned, namely this approach is representing the real situation concerning the
modeling uncertainties as far as the variance of the different solutions is regarded.)
The first and second kinetic benchmarks ¹ are exceptions from this point of view
because their definition makes it theoretically possible to have the converged
reference solution. In this sense, the two latter mentioned benchmarks can be regarded
as ?mathematical benchmarks? in close relation with their relative simplicity.
Nevertheless, the efforts for determining their converged reference solutions seem not
fully satisfactory, especially regarding the AER Benchmark Book
(?http://aerbench.kfki.hu/aerbench/ ?). A very important, large step in this direction
was made by N. Kolev, R. Lenain, and C. Magnaud calculating the second kinetic
benchmark by using the CRONOS fine-mesh method ². An extrapolated to zero
mesh size reference solution was recommended for the steady states and it revealed
significant deviations from the solutions obtained from the nodal codes. The
significant deviation raised certain doubts in some participants concerning the advised
reference solution. The goal of the present work is to contribute to the existence of the
reference solution of the second kinetic rod ejection benchmark in the AER
Benchmark Book.

In the present paper the possibility of performing a multiscale modeling of the hydraulics of a
reactor pressure vessel (RPV) using the system code ATHLET ¹ is considered. The main
focus of the work is the modeling of the active core. There are several length scales that can
be used in the modeling of RPV, namely 1) macroscale, or the set of fuel assemblies (sectors);
2) mesoscale or an individual fuel assembly; and 3) microscale or a pinwise modeling of one
or several fuel assemblies. In all above mentioned cases the modeling structures are
represented by the systems of parallel channels, hydraulically coupled with each other in the
transversal direction (in the absence of the walls). The modeling allows using both meso- and
microstructures and the object nodalization of the upper and lower part of the reactor uses
macro- and mesostructures. To improve the modeling of the flow in the transverse direction
the convection term in the ATHLET momentum equation is corrected to give the possibility
to simulate two- and three-dimensional flows.

Will be available later.

Will be available later.

Will be available later.

A diagnostics of initiating events, states of reactor systems and relevant safety
variables is important for safe operation of an NPP. On the one side one needs to correctly
interpret signals from measuring instrumentation. On the other side a fast classification of the
transients significantly improves the reliability of NPP operation.
This paper presents an original approach for detection of NPP equipment failures and
for classification of complex transient states of NPP. It is based on the use of artificial neural
networks. As a learning system to train the recognition of safety relevant processes the bestestimate system code ATHLET (GRS) is used.
The sequence and the principles of network-learning, as well as the justification for
selecting the type of neural network are presented. The paper contains two applications of the
approach: a) the integral and local estimates of recognition accuracy of latent and false
rejections in measuring channels and recommendations on the application of the implemented
methodology in automatic process control systems of NPP units; b) a robust classification of
the transients in German NPPs using the implementation of the ALADDIN tool developed in
the Halden Reactor Project.in the framework of ATLAS (GRS) environment.

In the paper, the basic results on working out of modules of thermomechanical calculation
for the analysis of the fuel rod behaviour in RIA type accidents according to requirements of
integrated code ATHLET/BIPR-VVER are stated.
Integrated code ATHLET/BIPR-VVER provides calculation of necessary neutronphysical and heat-hydraulic parameters representative fuel rods for the program of “mechanical”
fuel rod calculation (parameters of the stressed-deformed state and the mechanic of
destruction).
The module of an estimation of parameters of a preemergency condition representative
fuel rod contains a necessary set of correlations for their conservative estimation. It is necessary
to underline, that this module is under construction by a principle of the minimum sufficiency
for the emergency analysis, and for reception of more realistic estimation, the given module can
be replaced with the interface of the resource fuel rod code.
The description of total modules is presented.

Heat transfer in the gas gap of a fuel rod affects the behaviour of the reactor core also in the scale
of the whole reactor. In the VTT?s reactor dynamic codes, heat transfer in a gas gap is usually
characterized by some temperature dependent curve that is given as input data. Typically burnup
is taken into account in a rather simple way, e.g. by using different curves for fresh assemblies
and one, two and three year old fuel assemblies. Implementation of a detailed fuel model in the
reactor dynamics codes would not be a suf?cient way to model the gas gap. As the properties
of a fuel rod change remarkably during its normal reactor life, the initial state of the transient
must be somehow determined.
In this paper, the statistical version of the fuel performance code ENIGMA has been used to
determine the properties of a representative VVER-440 fuel rod at various points of its reactor
life. The aim was to ?nd such a way to model the gas gap properties, that could be utilized in
VTT?s reactor dynamic codes TRAB-3D and HEXTRAN.

The EU NURISP V1000-LR0-STAT benchmark is focused on three-dimensional VVER-1000
steady-states with assembly-averaged and fine-mesh power distributions. Calculation results
are compared with measurements from the ?JV ?e? experimental reactor LR-0, which
contains shortened fuel assemblies of the VVER-1000 power reactor. The calculations are
performed with the reactor dynamic nodal code DYN3D. The usual node-averaged approach
is applied in the first step. Then, an intra-nodal pin-wise power reconstruction method and its
combination with the pin-powers from the lattice code HELIOS zero-net-current fuelassembly calculations are applied. The calculations agree quite well with experiments.
Theoretically, the power distributions calculated by pin-wise reconstruction should show
better agreement with the measurement than the node-averaged values. This is confirmed in
selected cases.

The paper deals with CFD modeling and simulation of coolant flow in fuel assembly of
nuclear reactor VVER 440. Geometry of fuel assembly and thermocouple housing with
channel in upper core supporting plate have been created in 3D CAD system, where detailed
geometry model was simplified. The discretization of geometry was realized in ANSYS
ICEM CFD software, where blocking strategy was used. ANSYS CFX was chosen as a main
CFD software tool, where all analyses were performed. Also the flow of coolant in central
tube as well as heat transfer across the wall of central tube were considered in the model.
Only steady-state analyses are performed. Generated thermal power in fuel rods is prescribed
individually for each rod with axial distribution of power. In analyses, there is investigated the
influence of volume flow for defined thermal power on fuel assembly output temperature.
Also the influence of thermal power and volume flow on pressure drop is investigated.
Different turbulent models and mesh densities are examined.

Lately 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. With the calculations the
temperature distributions in the assembly heads are determined and the in-core thermocouple
signals are evaluated. The computed signals are compared to 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 burnups of
the fuel assemblies. These deviations are significant in the cases of low burnups and
negligible in the cases of higher burnups. This behavior is mainly caused by the three Gd fuel
rods near the assembly center.