Content uploaded by Blair Bromley
Author content
All content in this area was uploaded by Blair Bromley on Oct 25, 2018
Content may be subject to copyright.
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance UNRESTRICTED
Pittsburgh, Pennsylvania, USA, May 9-14, 2010, on CD-ROM, American Nuclear Society, CW-119190-CONF-009
LaGrange Park, IL (2010)
COMPARISON OF WIMS-AECL / DRAGON / RFSP AND MCNP RESULTS
WITH ZED-2 MEASUREMENTS FOR CONTROL DEVICE WORTH AND
REACTOR KINETICS
Jeremy Pencer, F. Choy Wong, Blair P. Bromley, Julian Atfield and Mike Zeller
Atomic Energy of Canada Limited (AECL)
Reactor and Radiation Physics Branch
Chalk River Laboratories,
Chalk River, ON, Canada, K0J 1J0
pencerj@aecl.ca; wongfc@aecl.ca; bromleyb@aecl.ca; atfieldj@aecl.ca; zellerm@aecl.ca
ABSTRACT
This paper summarizes comparisons between MCNP5 and WIMS-AECL / DRAGON / RFSP
calculations and experimental results obtained from the Zero Energy Deuterium (ZED-2) critical
facility at AECL Chalk River Laboratories. MCNP5 and WIMS-AECL / DRAGON / RFSP were
used to calculate reactivity worths for two reactivity devices, a mechanical zone controller (MZC)
and shut-off rod (SOR) in a lattice similar to that of the ACR-1000®. WIMS-AECL / DRAGON /
RFSP was also used to obtain kinetics parameters for a transient based on a rod drop of a ZED-2
SOR. ZED-2 experiments were performed using 43-element ACR Low Enriched Uranium (ACR-
LEU) fuel bundles with H2O- or air-cooled fuel bundles arranged in a 24-cm pitch square lattice.
Calculations with MCNP5 gave biases in device worths that were within 0.2 mk of measured
values, while WIMS-AECL / DRAGON / RFSP gave values that were within 0.3 mk of measured
values. Transient analyses using the CERBERUS module within RFSP yielded a total delayed
neutron fraction (β) that was within 4% of the value derived by point kinetics analysis of
experimental data. The corresponding delayed photo-neutron fraction (βphoto-neutron) from
CERBERUS was within 5% of that derived by point kinetics. This study has helped quantify the
agreement between calculation and measurement for codes that are used in the safety analysis of
the ACR-1000 reactor. Results demonstrate good agreement in code predictions.
Key Words: MCNP, WIMS-AECL, RFSP, ZED-2, experiments
1. INTRODUCTION
The ACR-1000 has been developed by Atomic Energy of Canada Limited (AECL) for both the Canadian
domestic and international markets [1]. This reactor is based on the CANDU® design with fuel bundles in
horizontal pressure tubes, on-line refuelling and heavy water moderation. The ACR-1000 design achieves
reduced capital costs by significantly lowering heavy water requirements through the use of light water as
the primary coolant and reducing the lattice spacing to 24 cm. For analysis of the ACR-1000, the
CANDU physics codes have been reviewed and enhanced [2]. An ongoing comprehensive program of
testing and analysis provides incremental validation of the capabilities and accuracy of these advanced
physics codes [3]. This paper summarizes the results of a validation study comparing calculation results
from MCNP5 and WIMS-AECL / DRAGON / RFSP with experimental results obtained from the ZED-2
critical facility, using fuel bundles similar to ACR Low Enriched Uranium (ACRsim-LEU) fuel bundles.
ACR® (Advanced CANDU Reactor) and CANDU® are registered trademarks of Atomic Energy of Canada Limited.
Jeremy Pencer, et al.
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 2/10
ACR fuel bundles have a number of features that distinguish them from conventional CANDU bundles,
notably enrichment of the fuel, an increase in the number of fuel pins (42 rather than 37 or 28) and the
presence of a neutron absorbing centre element containing oxides of zirconium, dysprosium and
gadolinium. ACRsim-LEU, the fuel used in this study, is very similar to ACR-1000 fuel. The use of light
water coolant and 24-cm square lattice pitch for these experiments are also ACR-1000 features. Two
types of reactivity control devices were tested in the ZED-2 critical facility, a Mechanical Zone Controller
(MZC) and Shut Off Rod (SOR). The MZC was composed of stainless steel, and the SOR was made of
boron carbide (B4C) clad in aluminium. Both are similar to reactivity control devices intended for use in
the ACR-1000.
A top-down view of the core layout is shown in Figure 1. Fifty-two aluminium ACR-type channels were
arranged in a 24-cm pitch square lattice. All experiments were performed using channels with 3 bundles
of ACRsim-LEU fuel at the bottoms of the channels supplemented with two bundles of CANFLEX®-LEU
on top. Since only a small portion of the CANFLEX-LEU bundles are below the critical moderator level,
the core reactivity is dominated by the ACRsim-LEU. The ACRsim-LEU fuel bundles each consist of a
large central burnable neutron absorber (BNA) pin made up of a mixture of dysprosium, gadolinium and
hafnium oxides, which is surrounded by 42 smaller pins, made up of 1.7 wt% 235U/U enriched uranium
oxide. The CANFLEX-LEU fuel bundles each consist of 8 inner fuel pins surrounded by 35 slightly
smaller fuel pins, all containing 0.95 wt% 235U/U enriched uranium oxide.
Three reactor physics phenomena were investigated: device-movement-induced-reactivity, flux
distribution in space and time, and prompt and delayed neutron kinetics. The experiments included
measurements of critical moderator heights, measurement of neutron flux distributions using activation of
copper wires, and measurement of transients induced by controlled reactivity insertions. The effects of
reactivity devices on core reactivity were inferred from measured changes to moderator critical height and
changes to reactor power versus time inferred from time dependent changes in neutron flux. The ZED-2
core under various operational conditions, including critical moderator height, was simulated using
MCNP5 [5] and WIMS-AECL / DRAGON / RFSP [6], [7], [8] in order to obtain calculated values for
reactivity (keff), normalized copper activation distributions, and changes to reactor power versus time.
Comparison of these calculated values with experiment contribute to validation of both MCNP5 and the
WIMS-AECL / DRAGON / RFSP code set [2]. Additional validation activities include the extension of
these results to ACR-1000 design and operation conditions using the methods described in [3] and [9].
2. EXPERIMENTS
2.1. Reactivity Worth of Control Devices and Flux Measurements
Critical moderator heights, hc, were measured for H2O-cooled lattices of ACR-type fuel assemblies alone,
or with either an SOR or MZC positioned intersecting the centre of the core. The SOR or MZC were
hung and placed orthogonal to the fuel channels (as shown in Figure 1). In the vertical direction (i.e.
along the length of the fuel channels), these devices were completely surrounded by ACRsim-LEU
bundles. Level Coefficient of Reactivity (LCR) measurements were also performed alone and with either
the SOR or MZC in place. In order to determine the LCR, a small positive transient was induced by a
perturbation, Δh, to the moderator height. Inverse point kinetics analyses were then performed to obtain
estimates of reactivity device worths. It is important to note that the reactivity worths derived from point
kinetics do not correspond to true experimental parameters, since their values depend on the
CANFLEX® (CANdu FLEXible) fuel is a registered trademark of AECL and KAERI.
Comparison of WIMS-AECL / DRAGON / RFSP and MCNP Results With ZED-2 Measurements: Control Device Worth and Reactor Kinetics
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 3/10
approximations inherent in the point kinetics method. Consequently, the reactivity worths derived from
point kinetics, while useful for comparison with MCNP and RFSP results, are not appropriate for use in
evaluation of the bias in reactivity worths of these codes.
In order to measure the flux distribution near the reactivity devices, experiments were performed with
copper wires spaced at 1 cm intervals in the North-South direction in the vicinity of either the SOR or
MZC (shown schematically in Figure 1). The foils were irradiated at 100 Watts and subsequently
retrieved and their activation measured.
2.2. Rod Drop
ZED-2 was operated at approximately 100 W for approximately two hours. The critical moderator level
was near the top of the ACRsim-LEU fuel. A shut down was initiated by dropping a single ZED-2 stand-
by absorber rod (SAR) into the centre of the lattice. This SAR (a hollow cadmium tube clad in
aluminium) was inserted parallel to the fuel channels. The change in neutron flux (i.e. reactor power)
versus time was recorded as the transient proceeded, using an ion chamber neutron detector. The total
delayed photo-neutron fraction was determined from point kinetics analysis of the transient.
A
CR-LEU
12 10 8 6 2 0 24 4 6 8 10 12
F
G
H
I
J
K
L
M
N
O
P
24.00 cm
SOR / MZC
Cu Wires
Figure 1. Top view of the ZED-2 lattice illustrating the core layout, reactivity control device
orientation and copper wire arrangement.
Jeremy Pencer, et al.
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 4/10
3. COMPUTER CODES AND COMPUTATIONAL MODELS
3.1. MCNP5
MCNP5, a general-purpose Monte Carlo N-Particle code, is a three-dimensional, stochastic, probabilistic,
neutron transport code [5]. The code treats an arbitrary three-dimensional configuration of materials in
geometric cells defined by surfaces. For the calculations presented here, MCNP5 was used in conjunction
with a nuclear data library based on ENDF/B-VI.
The MCNP5 models were precise representations of the actual experimental set-up, including the fuel
bundles, aluminium fuel channels, bottom and radial D2O reflectors, aluminium calandria tank, bottom
and radial graphite reflector, outer neutron and gamma-ray shielding, various air gaps, etc. Runs were
performed with 80 million neutron histories, which typically gave statistical uncertainties of less than
±0.1 mk in the calculation of keff. For the flux distribution cases, each MCNP case was run using 600
million neutron histories, in order to obtain acceptable statistical uncertainties in the flux tallies.
3.2. WIMS-AECL 3.1
WIMS-AECL 3.1 is a two-dimensional multi-group deterministic lattice physics code that solves the
integral neutron transport equation using collision probabilities [6]. In addition, the code can take into
account leakage due to finite reactor dimensions, using input bucklings, and can model heterogeneous
effects via its multi-cell capability. For this study, WIMS-AECL was used in conjunction with an 89-
group nuclear data library derived from ENDF/B-VI. For the results presented here, WIMS-AECL was
used to produce homogenized two-group cross sections for use in RFSP core calculations.
Fuel channel models in WIMS-AECL consist of two-dimensional slices through the channels,
perpendicular to the channel axes. A number of approximations were performed in order to obtain two-
dimensional representations of fuel bundles. These approximations include smearing of the fuel density,
to account for volume lost due to the dish depth of the fuel pellets, and smearing of the cladding density
in order to account for the reactivity contribution from material at the ends of the fuel bundles.
Two-group homogenized cross sections for use in RFSP were determined based on calculations using
both single-cell and multi-cell WIMS-AECL models. Single-cell models were used for fuel channels in
the interior of the x-y plane of the core, while channels that were adjacent to the radial D2O reflector were
modelled using multi-cell models that accounted for the heterogeneous environment around the channel
of interest [8].
3.3. DRAGON
DRAGON is a deterministic computer code for multi-group neutron transport calculations, which is
capable of simulating neutron transport in 1-d, 2-d, and 3-d geometries [7]. It also provides options to use
a number of different methods for calculation of collision probabilities, by taking advantage of the
symmetries of various geometries. In conjunction with WIMS-AECL via the “side-step” method,
DRAGON was used to calculate macroscopic, incremental two-group cross sections for reactivity
devices, for use in full core simulations with RFSP.
Comparison of WIMS-AECL / DRAGON / RFSP and MCNP Results With ZED-2 Measurements: Control Device Worth and Reactor Kinetics
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 5/10
In the “side-step” method, WIMS-AECL is used to generate 89-group homogenized macroscopic cross
sections for fuel, calandria and pressure tubes, and reactivity devices. These cross sections are then used
in a DRAGON model, which consists of two fuel channels and the interceding reactivity control device.
In the DRAGON model, the fuel channel is represented approximately by a channel lumen, consisting of
a homogenized cylindrical region of fuel and coolant and a surrounding annulus, consisting of the
pressure tube and calandria tube, and interceding gas annulus. The reactivity device materials are
represented explicitly. For a particular reactivity device, two DRAGON calculations are performed, one
with the reactivity device in place and a second with the device removed. Incremental cross sections for
use in RFSP are calculated from the difference between the two-group homogenized cross sections
calculated with the device in place and removed.
3.4. RFSP
RFSP is a neutron-diffusion Reactor Fuelling Simulation Program, which solves the three-dimensional,
two-group neutron diffusion equations using the finite difference method in a Cartesian coordinate system
[8]. The geometry is defined by sets of Cartesian meshes, where each mesh is assigned a set of two-group
cross sections and corresponds to one point in the two-group flux solution of the diffusion equation. In
this study, RFSP was used to evaluate reactivity levels and their changes. The homogenized two-group
cross sections used in RFSP were evaluated (as described above) from WIMS-AECL calculations of
single-cell and multi-cell lattice physics models, and from DRAGON for reactivity devices.
Because RFSP is a deterministic, two-group diffusion code with a Cartesian coordinate system, a number
of approximations and simplifying assumptions were made in the modelling of the ZED-2 experiments.
Each fuel channel was represented in the x-y plane by a 2×2 arrangement of sub-cells, or meshes, with
equal sub-cell mesh spacings in the x and y directions (dx=dy=12 cm). Finer Cartesian mesh spacings
were used around the radial interface between the D2O reflector and calandria, in order to represent better
the circular boundaries of the D2O, calandria and radial graphite reflector. The RFSP model of ZED-2
extended to the outermost physical boundary of the radial and bottom graphite reflectors. In the radial
direction, a vacuum boundary condition was imposed. At the upper axial boundary, the RSFP model
extended to the top of the moderator level. The lower axial boundary coincided with the bottom of the
graphite axial reflector. Above the moderator and below the graphite axial reflector, vacuum boundary
conditions were also imposed.
The rod drop transient was modelled using the CERBERUS module within RFSP [10]. Cross section data
used in the CERBERUS calculation were prepared in the same way as for the static RFSP calculations,
using WIMS-AECL and DRAGON, as described in Section 3.2 and Section 3.3, respectively. The
starting point for the transient calculation corresponds to ZED-2 at a power of approximately 100 W, with
the ZED-2 SAR suspended above the core. During the insertion of the ZED-2 SAR, 100 milli-second
calculation time steps were employed. After full insertion of the device, these time steps were increased
to 1-second intervals.
4. RESULTS
4.1. Reactivity Calculations
Results of reactivity calculations for H2O-cooled lattices of ACR fuel assemblies alone, or with either an
SOR or MZC positioned at the centre of the core are listed in Table I and Table II, for MCNP and
Jeremy Pencer, et al.
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 6/10
WIMS-AECL / DRAGON / RFSP, respectively. MCNP gives negative biases in keff in the range of
between –10.4 mk and –10.5 mk with and without reactivity devices in place, with uncertainties (1 σ) of
approximately ±0.1 mk for all cases. The code combination WIMS-AECL / DRAGON / RFSP also gives
negative biases in keff, ranging between –4.1 mk and –4.4 mk, with uncertainties of approximately
±0.2 mk for all cases. The large negative biases in keff result primarily from the use of the ENDF/B-VI
library [11].
Table I. keff, keff bias and uncertainties from MCNP
Device keff σ keff keff bias
(mk) σ keff bias
(mk)
None 0.98960 ±0.00007 -10.51 ±0.07
MZC 0.98967 ±0.00007 -10.44 ±0.07
SOR 0.98975 ±0.00007 -10.36 ±0.07
Table II. keff, keff bias and uncertainties from WIMS-AECL / DRAGON / RFSP
Device keff σ keff keff bias
(mk) σ keff bias
(mk)
None 0.99563 ±0.00018 -4.37 ±0.18
MZC 0.99588 ±0.00017 -4.12 ±0.17
SOR 0.99549 ±0.00018 -4.51 ±0.18
4.2 Reactivity Worth of Control Devices
As discussed above in Section 3.1 and Section 3.4, MCNP and RFSP were used to calculate keff for a
critical H2O-cooled lattice with either an MZC or SOR reactivity control device in place. In order to
determine the reactivity worth of the MZC or SOR, the control device was removed from the model and
keff was recalculated using the same critical height. The reactivity worth,
ρ
, can then be determined from
()
(
)
()( )
outdevicehkindevicehk
outdevicehkindevicehk
ceffceff
ceffceff
−×−
−
−
−
=,,
,,
ρ
(2)
where hc refers to the critical height with the device in place, and “device–in” and “device–out” refer to
models with the reactivity device either in place or removed, respectively.
As shown in Table 3, MCNP predicts the worth of the MZC to be –5.9 mk with a bias of +0.1 mk
±0.3 mk. The worth of the SOR is predicted by MCNP to be –6.1 mk with a bias of +0.2 mk ±0.3 mk.
RFSP predicts similar values for the worths and biases of the reactivity devices, with the worth of the
MZC predicted to be –5.8 mk with a bias of -0.3 mk ±0.3 mk, and the worth of the SOR to be -6.7 mk
with a bias of -0.1 mk ±0.3 mk. The corresponding reactivity worths based on reactor period
measurements of LCRs are -5.5 mk ± 0.5 mk and -5.9 mk ± 0.5 mk, for the MZC and SOR, respectively.
Comparison of WIMS-AECL / DRAGON / RFSP and MCNP Results With ZED-2 Measurements: Control Device Worth and Reactor Kinetics
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 7/10
Table III. Reactivity worth, bias and uncertainties from MCNP and WIMS-AECL / DRAGON /
RFSP
Device Worth
(mk) Βias (mk) Bias
Uncertainty
(mk)
MCNP
MZC -5.9 +0.1 ±0.3
SOR -6.1 +0.2 ±0.3
WIMS-AECL / DRAGON / RFSP
MZC -5.8 -0.3 ±0.3
SOR -6.7 -0.1 ±0.3
Point Kinetics Analysis
MZC -5.5 ± 0.5 na na
SOR -5.9 ± 0.5 na na
4.3. Fine Structure Flux Map Distributions
The simulations for the flux mapping evaluation were run with neutron-flux tallies at the same positions
of the foil locations. The copper wires were explicitly modelled, although they were not used for
performing the tallies of the activation rates. The very small volume of these activation wires would
require an enormous number of neutron histories in order to keep the statistical uncertainties in the
calculated activation rates to an acceptably low level. So, larger tally volumes surrounding the regions of
the activation wires were used instead. Since the foil activities are dependent on the flux distributions,
comparison between MCNP and the experiment for the foil activity distributions gave an estimate of the
accuracy with which MCNP can predict the flux distributions.
The Cu activity comparisons for the case with the MZC are shown in Figure 2 and those for the SOR are
shown in Figure 3. The error bars plotted in the figures correspond to a 1% uncertainty for the
experimental values and the average uncertainty value of the sample points for the MCNP values. For
these fine-structure flux maps, the Cu activities along the 0 axis as calculated by MCNP agree very well
with the experimental values to within approximately 1% relative difference. The corresponding Root
Mean Square (RMS) errors for the copper activation are less than 2% for both the SOR and MZC. The
slight asymmetry in the flux plots with respect to the centre line is a result of small asymmetries in the
ZED-2 core (likely related to the locations of the moderator dump lines and detector spaces in the base of
the graphite axial reflector). The fact that the flux asymmetry in the experimental data is reproduced in
the MCNP calculation shows that the MCNP ZED-2 model faithfully reproduces this asymmetry.
Jeremy Pencer, et al.
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 8/10
0.6
0.7
0.8
0.9
1.0
1.1
-15.0 -10.0 -5.0 0.0 5.0 10.0 15.0
Distance From MZC Centreline (cm)
Normalized Cu Activation
Experimental
MCNP-Cu Activity
Figure 2 Normalized distributions of Cu activity at 0 axis north-south for MZC.
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
-15.0 -10.0 -5.0 0.0 5.0 10.0 15.0
Distance from SOR Centreline (cm)
Normalized Cu Activation
Experimental
MCNP-Cu Activity
Figure 3 Normalized distributions of Cu activity at 0 axis north-south for SOR.
4.4. Transient Analysis
The CERBERUS module of RFSP is used for the simulation of transients, such as the rod drop
experiment considered in this paper. Two parameters were examined: the delayed neutron fraction and
the time dependence of the neutron flux. The total delayed neutron fraction as determined by
Comparison of WIMS-AECL / DRAGON / RFSP and MCNP Results With ZED-2 Measurements: Control Device Worth and Reactor Kinetics
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 9/10
CERBERUS was 7.4966e-3 and delayed photo neutron component was 1.3080e-4. The total delayed
neutron fraction was within 4% of the value obtained by point kinetics analysis, 7.2310e-3, while the
delayed photo neutron fraction was within 4% of the point kinetics value, 1.2491e-4. Both the
CERBERUS calculation and point kinetics analysis were based on the delayed neutron data of Blachot
[12], Baumann [13] and delayed photo-neutron yields based on calculations of Lone and Jones [14].
The measured time-dependent flux and corresponding CERBERUS calculation are shown in Figure 4.
During the first 6 seconds of the rod drop, the CERBERUS power calculation closely followed the
experimental data. However, at intermediate times, CERBERUS underpredicted the power by
approximately 1% total power or approximately 10% relative difference, followed by an over prediction
at long times.
0.1%
1.0%
10.0%
100.0%
10 100 1000
time (s)
% Total Power
Experiment
RFSP
Figure 4: Comparison of calculated to measured values of % total power.
5. CONCLUSIONS
In this study, calculation results from MCNP5 and WIMS-AECL / DRAGON / RFSP were compared with
ZED-2 measurements obtained from a critical configuration of H2O-cooled ACRsim-LEU fuel arranged
in a D2O-moderated, 24-cm pitch square lattice alone, or with an MZC or SOR inserted in the core.
Transient calculations were also performed using the CERBERUS module of RFSP, in order to model a
rod drop experiment performed using a ZED-2 SAR. The biases in device reactivity worths and their
uncertainties obtained from both MCNP and WIMS-AECL / DRAGON / RFSP were small, with
magnitudes less than 0.5 mk for both codes and both devices. Transient analysis using the CERBERUS
module within RFSP yielded a total delayed neutron fraction (β) that was within 4% of the value derived
by point kinetics analysis of experimental data. The corresponding delayed photo-neutron fraction
(βphoto-neutron) from CERBERUS was within 5% of that derived by point kinetics. Overall, these results
demonstrate good agreement between measurement and calculation and have helped to quantify the
accuracy of reactor physics codes for use in ACR-1000 reactor analysis.
Jeremy Pencer, et al.
PHYSOR 2010 – Advances in Reactor Physics to Power the Nuclear Renaissance
Pittsburgh, Pennsylvania, USA, May 9-14, 2010 10/10
ACKNOWLEDGMENTS
The authors gratefully acknowledge the help and assistance provided by fellow staff at AECL including:
Alex Trottier, Peter Schwanke, Tony Liang, Elisabeth Varin, Jingliang Hu, Jinchao Mao, Wei Shen, David
Jenkins, Bruce Wilkin, Fred Adams, Jerry McPhee, Greg Morin, Dimitar Altiparmakov, Jim Sullivan,
Peter Boczar, Dave Wren, and Michele Kubota.
REFERENCES
1. J. Hopwood, “The AECL ACR-1000 Reactor,” Proceedings of PHYSOR-2006, Vancouver, British
Columbia, Canada, September 10-14 (2006).
2. A. Buijs, “The Design Process and Simulation Tools Used in the Neutronic Design of the ACR-1000
Core,” Proceedings of the 23rd CNS Nuclear Simulation Symposium, Ottawa, Ontario Canada, Nov 2-
4 (2008).
3. B. P. Bromley, F. P. Adams, M. B. Zeller, D. G. Watts, B. V. Shukhman, and J. Pencer, “Validation of
MCNP and WIMS-AECL / DRAGON / RFSP for ACR-1000 applications,” PHYSOR 2008,
International Conference on the Physics of Reactors, Nuclear Power: A Sustainable Resource,
Interlaken, Switzerland, September 14 –19 (2008).
4. M. B. Zeller “Physics Experiments in the ZED-2 Reactor in Support of the Advanced CANDU
Reactor,” Proceedings of the 25th Annual Conference of the Canadian Nuclear Society, Toronto,
Ontario, Canada, June 6-9 (2004).
5. X-5 Monte Carlo Team, “MCNP – A General Monte Carlo N-Particle Transport Code, Version 5,”
LA-UR-03-1987, Los Alamos National Laboratory, Los Alamos, New Mexico, April (2003).
6. D. V. Altiparmakov, “New Capabilities of the Lattice Code WIMS-AECL,” PHYSOR 2008,
International Conference on the Physics of Reactors, Nuclear Power: A Sustainable Resource,
Interlaken, Switzerland, September 14 –19 (2008).
7. R. S. Davis, “Qualification of DRAGON for Reactivity Device Calculation in the Industry Standard
Toolset,” Proceedings of the 25th Annual CNS Conference, Toronto, Ontario, Canada, June 6-9 (2004).
8. W. Shen, D.A. Jenkins, A. Connolly, I. Attieh, D. Buss, T. Liang, and S. Hart, “Benchmarking of
WIMS-AECL / RFSP Multi-cell Methodology with MCNP for ACR-1000 Full-Core Calculations,”
PHYSOR 2008, International Conference on the Physics of Reactors, Nuclear Power: A Sustainable
Resource, Interlaken, Switzerland, September 14 –19 (2008).
9. B. T. Rearden, M.L. Williams and J. E. Horwedel, “Advances in the TSUNAMI Sensitivity and
Uncertainty Analysis Codes Beyond SCALE 5,” Trans. Am. Nucl. Soc., Volume 92, pp. 760-762
(2005).
10. J. V. Donnelly and E. M. Nichita, “Verification of Two-Group CERBERUS for a Loss-of-Coolant
Analysis in a Simplified Reactor Model,” 20th Annual CNS Conference, Montreal, Quebec, Canada,
May 30 – June 2 (1999).
11. K. S. Kozier, “Energy and Spatial Dependence of MCNP Simulations For ZED-2 Critical
Experiments,” 23rd CNS Nuclear Simulation Symposium, Ottawa (2008).
12. J. Blachot, M. C. Brady, A. Filip, R. W. Mills and D. R. Weaver. “Status of Delayed Neutron Data –
1990,” NEACRP-L-323 / NEANDC-299”U”, OECD Nuclear Energy Agency (1990).
13. N. P. Baumann, R. L. Currie and D. J. Pellerin. “Production of Delayed Photo-neutrons from U235 in
Heavy Water Moderator,” Trans. Am. Nucl. Soc., pp 499-500 (1973).
14. M. A. Lone and R. T. Jones, “Delayed Photo-neutron Yields in D2O from Fission in 235U, 238U, 239Pu
and 241Pu,” Proceedings of the 21st Conference of the Canadian Nuclear Society, Toronto (2000).
