PHYSOR 2012 – Advances in Reactor Physics – Linking Research, Industry, and Education

Knoxville, Tennessee, USA, April 15-20, 2012, on CD-ROM, American Nuclear Society, LaGrange Park, IL (2012)

COMPARISON OF A NUSCALE SMR CONCEPTUAL CORE DESIGN

USING CASMO5/SIMULATE5 AND MCNP5

Brandon Haugh
Studsvik Scandpower Inc.
1015 Ashes Drive, Suite 205, Wilmington, NC 28405
[email protected]

Ali Mohamed
NuScale Power Inc.
1100 NE Circle Blvd, Suite 350, Corvallis, OR 97330
[email protected]

ABSTRACT

A key issue during the initial startups of new Small Modular Reactors (SMRs) is the lack of
operational data for reactor model validation. To help better understand the accuracy of the
reactor analysis codes CASMO5 and SIMULATE5, higher order comparisons to MCNP5 have
been performed. These comparisons are for an initial core conceptual design of the NuScale
reactor. The data have been evaluated at Hot Zero Power (HZP) conditions. Comparisons of core
reactivity, fuel temperature coefficient (FTC), and moderator temperature coefficients (MTC) have
been performed.

Comparison results show good agreement between CASMO5/SIMULATE5 and MCNP5 for the
conceptual initial core design.

Key Words: SMR, CASMO5, SIMULATE5, NuScale, MCNP5

1 INTRODUCTION

The development of SMR’s presents unique issues for reactor vendors and potential operators. A
key issue addressed in this paper is the lack of operating history for SMRs. This lack of
operating history creates a challenge in validating predictions of neutronic parameters during the
startup of this new design.

The NuScale reactor is analyzed here using current state-of-the-art lattice physics code,
CASMO5 [1] and reactor simulator, SIMULATE5 [2]. The reference solution produced for
validation uses a higher order computational method. In this case it is the continuous energy
Monte Carlo code MCNP5 [3].

The neutronic parameters that are required to be validated can be parsed into two groups, global
parameters and local parameters.

Global reactor parameters are those that are derived from examining the state of the entire
reactor core. These include reactivity coefficients such as moderator temperature, fuel
 Brandon Haugh, Ali Mohamed

temperature (Doppler), and boron. As well as the reactivity associated with control rods. The
combinations of these parameters defines the operational characteristics of the reactor and are a
majority of the quantities measured during start-up physics testing.

Local reactor parameters are those calculated over small geometric regions at a given core state.
The parameters include those that are associated with the Specified Acceptable Fuel Design
Limits (SAFDLs). These include pin power quantities that are derived from the local reaction
rate distributions. The resulting integrated quantities provide fuel pin exposure, local linear heat
generation rate (Fq) and enthalpy rise hot channel factors (Fh) as examples.

As a first step this paper focuses on global neutronic parameters that are typically verified during
the start-up physics testing. These tests are a component of what is used to verify that core
design predictions are sufficiently accurate to allow the unit to operate within the domain
analyzed by the safety analysis.

In the future the next steps of this validation will focus on examining local reaction rate
distributions.

2 REACTOR OVERVIEW

The NuScale SMR is a scalable modular integral pressurized water reactor (iPWR) design. Each
reactor module has its own turbine generator set that is capable of producing 45MWe. A scalable
plant is capable of supporting 12 modules within a single reactor building for an overall capacity
of 540 MWe. The independent modularity of each reactor represents a reduced operational risk
for electricity generation and a new way of performing refueling and maintenance.

The passively safe design of the reactor uses natural circulation as the method to transfer heat
from the nuclear fuel to the steam generator as well as for the various safety systems. The
integral nature of the design allows for all of the primary components to be contained within the
main pressure vessel. These include the core, pressurizer, and steam generators. The control rod
drive mechanisms are fixed to the top of the reactor vessel.

The NuScale reactor employs several unique passive safety systems: the high-pressure steel
containment, the decay heat removal system (DHRS), and the containment heat removal system
(CHRS). The high-pressure containment houses and protects the nuclear steam supply system
(NSSS), control rod drive mechanisms, and associated safety-related systems from external
hazards. This containment is submerged in the reactor pool, which provides a passive heat sink
under loss of coolant accident (LOCA) conditions. The DHRS provides core cooling when
normal feedwater is not available, and removes residual heat generated by fission product decay
from the reactor core. The CHRS rapidly reduces the containment pressure and temperature and
maintains them at acceptably low levels for extended periods of time after a LOCA event.

The conceptual layout of a single module is shown visually in Figure 1.

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 2/12
Knoxville, Tennessee, USA April 15-20, 2012
 Comparison of NuScale SMR Conceptual Core Design Using CASMO5/SIMULATE5 and MCNP5

Figure 1: NuScale Passively Safe Reactor Concept

2.1 Reactor Core

The NuScale reactor core operates at a reduced power density, approximately 50% , compared to
a conventional PWR. Current cycle length design target is 24 months. Geometrically the core
consists of 37 fuel assemblies. The fuel assemblies are arranged to approximate a right cylinder
with seven assemblies comprising the widest dimension. The reactor core is controlled like a
conventional PWR with a combination of soluble boron and control rods. The design currently
contains 16 control rods arranged into two banks. Bank 1 is designated for reactivity control
during startup and operation. Bank 2 is designated as a shutdown bank for use when the reactor
receives a SCRAM signal. The lower right quarter core layout is shown below in Figure 2 where
the control rod positions are also indicated.

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 3/12
Knoxville, Tennessee, USA April 15-20, 2012
 Brandon Haugh, Ali Mohamed

Bank 1 Bank 2

Bank 1 Bank 2

Bank 2

Bank 2

Figure 2: NuScale Reactor Quarter Core Layout

2.2 Fuel Assemblies

The NuScale fuel assemblies are similar in design to conventional PWRs. The main difference is
a reduced active height of approximately 50% compared to a conventional assembly. The fuel
assembly lattice is 17x17 (Figure 3) with 264 fueled locations, 24 guide tubes, and one central
instrument tube. The fuel is conventional Zircaloy clad UO2 ceramic with enrichments below 5
w/o U235. An integral burnable absorber of Gadolina (Gd2O3) is used in some assemblies to
control beginning of cycle moderator temperature coefficient (MTC) and power peaking.

The initial conceptual core design consists of four different enrichments and five assembly types.
Of the five assembly types, two contain Gadolinia burnable absorber in some of the fuel pins.
All the fuel assemblies contain low enriched axial blankets to minimize leakage.

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 4/12
Knoxville, Tennessee, USA April 15-20, 2012
 Comparison of NuScale SMR Conceptual Core Design Using CASMO5/SIMULATE5 and MCNP5

Figure 3: NuScale 17x17 Fuel Lattice

2.3 Control Rods

The control rods are similar to a conventional PWR design where the 24 individual rodlets attach
to a central spider assembly and hub. The 24 rodlets are inserted into the guide tubes shown in
Figure 3. The central instrument tube is empty in assemblies containing control rods. The
control rod material for the conceptual design is Silver, Indium, Cadmium (AgInCd) alloy.

3 MODEL DESCRIPTION

There were two independent models generated for this benchmark. One created for the Monte
Carlo code MCNP5 and the second created for deterministic CASMO5/SIMULATE5. The key
assumptions and features are described below.

3.1 MCNP5 Model

The MCNP5 model of the NuScale reactor module comprised several units: the core, reflector,
barrel, and reactor pressure vessel. The core was modeled as an 11 x 11 lattice, and each core
lattice element is a 17x17 fuel/fuel-control assembly, an assembly (no fuel or control segment)
with only upper and lower core plates, a reflector or coolant. Figure 4 shows both plan and
elevation views of the MCNP5 model of one quarter of the NuScale reactor module.

The MCNP5 model of the 17x17 fuel assembly consisted of the lattice cells for fuel, control
rods, instrument and guide tubes, top nozzle, bottom nozzle, and spacer grids. The units that
make up the lattice cell/fuel pin were: top end plug, plenum, plenum spring, top blanket,
enriched zone, bottom blanket, bottom end plug, fuel-cladding gap, and cladding. The repeated-
structures capability of MCNP5 was used to build the lattice cells that comprise the UO2 or
2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 5/12
Knoxville, Tennessee, USA April 15-20, 2012
 Brandon Haugh, Ali Mohamed

UO2+Gd2O3 fuel elements at the beginning of cycle (BOC), control rod elements, instrument
tubes, and guide tubes. . Material compositions used in the MCNP5 model were matched to those
specified for the CASMO5 case matrix.

Reflective Boundary Fuel Assembly

at 45o With Control Rods

Reflector/Baffle

Upper Core Plate

Barrel
Spacer Grid

Fuel Assembly
With Control Rod
Guide Tubes

Downcomer Region
Coolant @ 541.7 K (515.4 oF)

Core Region
Coolant @ 541.7 K (515.4 oF) Lower Plenum Inlet

RPV Surveillance
Capsule
Liner

Reactor Pressure
Vessel

Reflective Boundary
at 0o Lower Core Plate

Figure 4: MCNP5 model of the NuScale reactor module shows plan and elevation views of
the core lattice, reflector, barrel, downcomer and pressure vessel. Temperature reflects Hot
Zero Power condition.

The MCNP5 code version 1.51 was used to model the NuScale reactor module core
configuration with continuous energy neutron cross sections data evaluations from ENDF/B-
VII.0 [5] libraries. The free gas scattering kernel model was chosen, except for hydrogen in light
water for which the appropriate S(,) data [3] at the appropriate temperatures were used to
account for chemical molecular binding and crystalline effects that become important as the
neutron’s energy becomes sufficiently low—below 4 eV. Doppler broadened cross sections were
processed, using the cross sections processing code Makxsf [4], at 541.7 K (515.4 oF) for HZP
isothermal calculations
For MTC calculations, the perturbed moderator temperature was processed at 552.8 K (535.4
o
F); for ITC calculations, cross sections for all materials in the model were were processed at
552.8 K (535.4 oF) for the perturbed condition. For FTC two perturbed calculations were
performed: one with fuel and cladding cross sections processed at 552.8 K (535.4 oF) and the
other with only the fuel cross sections processed at 552.8 K (535.4 oF). The average of the two
results was approximated to be the FTC best estimate value. Ideally cross sections for the

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 6/12
Knoxville, Tennessee, USA April 15-20, 2012
 Comparison of NuScale SMR Conceptual Core Design Using CASMO5/SIMULATE5 and MCNP5

cladding should be processed so as to reflect the true temperature profile since the FTC is very
sensitive to small temperature variations.
In the MCNP5 criticality calculations 80,000 starting fission neutrons were used, and 150 cycles
were skipped, to improve the spatial distribution of the source in the core. The criticality
calculations were run for a quarter core configuration with reflective boundary conditions at 0
and 45 degrees (see Figure 4). An initial guess of k-effective =1.0 was used, and a total of 2,000
k-effective cycles were run for all calculations.

3.2 CASMO5/SIMULATE5 Model

The CASMO5/SIMULATE5 model consisted of two main parts. These are the 2D lattice
segment models constructed in CASMO5 and the 3D core simulation model constructed in
SIMULATE5.

The recommended default input values for use in conventional PWRs were used. This was a
choice in this initial benchmark of the codes. This may be reexamined in the future if the
operating domain or characteristics of the NuScale reactor warrant it.

3.2.1 CASMO5

The lattice calculations were performed for any unique combination of fuel enrichment and
burnable absorber that may comprise an axial zone in a fuel assembly. For the NuScale core
design there are six unique lattice types. For each lattice a series of calculations were performed
to cover the potential operating domain of the reactor. In this model the spacer grids were
treated explicitly with branch calculations.

Three reflector segments were also run to determine the correct behavior at the boundaries of the
reactor. The three segments included a radial reflector to represent the core periphery, and a top
and bottom reflector. The radial reflector represented the explicit geometry of the boundary fuel
assembly and adjoining baffle, water, and barrel regions as needed. The top and bottom
reflectors were three zone regions that contain a homogenization of the appropriate materials
such as moderator, springs, cladding, endplugs, and nozzles.

The geometry was specified using cold dimensions and the thermal expansion model in
CASMO5 was deactivated to be consistent with the MCNP5 model. Additionally the resonance
upscatter model was turned off to be consistent with the MCNP5 modeling. The standard case
matrix, needed to generate all the cross sections and derivatives, was run. The cross section
library for CASMO5 was developed from ENDF/B-VII.0 [5].

3.2.2 SIMULATE5

The 3D simulator model was constructed using 25 axial zones and four radial zones per fuel
assembly. Each fuel assembly type was modeled and placed within the core model. The control
rods were modeled as the two banks shown in Figure 3. The spacer grids that were present in the

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 7/12
Knoxville, Tennessee, USA April 15-20, 2012
 Brandon Haugh, Ali Mohamed

active core region were modeled explicitly. The recommended default of four energy groups
was used in this calculation.

The thermal hydraulic detail in this model was minimal. The focus of the comparisons was on
isothermal zero power conditions.

4 VALIDATION RESULTS

To capture the data needed to simulate start-up physics testing several calculations were
performed. They can be easily divided into two categories: the first, core reactivity predictions,
the second, reactivity coefficients.

The core conditions for each calculation compared are shown in Table 1 and Table 2. All
calculations began at zero power isothermal conditions with All Rods Out (ARO) and a boron
concentration (BC) predicted by SIMULATE5. The resulting eigenvalue predictions and
coefficients were compared between the two codes.

Table 1: Core Reactivity Calculation Conditions

Calculation # Description
Power = 0.0%
TModerator = 515.4 °F, 541.7 K
ARO TFuel = 515.4 °F, 541.7 K
Boron Conc. = 1303 ppm
Rod Position = All Rods Out
Power = 0.0%
TModerator = 515.4 °F, 541.7 K
Bank 1@ TFuel = 515.4 °F, 541.7 K
50% Boron Conc. = 1303 ppm
Rod Position = Bank 1 @ 50% Withdrawn
Bank 2 @ 100% Withdrawn
Power = 0.0%
TModerator = 515.4 °F, 541.7 K
Bank 1@ TFuel = 515.4 °F, 541.7 K
0% Boron Conc. = 1303 ppm
Rod Position = Bank 1 @ 0% Withdrawn
Bank 2 @ 100% Withdrawn

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 8/12
Knoxville, Tennessee, USA April 15-20, 2012
 Comparison of NuScale SMR Conceptual Core Design Using CASMO5/SIMULATE5 and MCNP5

Table 2: Reactivity Coefficient Calculation Conditions

Coefficient Initial Condition Perturbed Condition
Power = 0.0% Power = 0.0%
TModerator = 515.4 °F, 541.7 K TModerator = 535.4 °F, 552.8 K
TFuel = 515.4 °F, 541.7 K TFuel = 515.4 °F, 541.7 K
MTC*
Boron Conc. = 1303 ppm Boron Conc. = 1303 ppm
Rod Position = All Rods Out Rod Position = All Rods Out

Power = 0.0% Power = 0.0%

TModerator = 515.4 °F, 541.7 K TModerator = 535.4 °F, 552.8 K
TFuel = 515.4 °F, 541.7 K TFuel = 535.4 °F, 552.8 K
ITC*
Boron Conc. = 1303 ppm Boron Conc. = 1303 ppm
Rod Position = All Rods Out Rod Position = All Rods Out

Power = 0.0% Power = 0.0%

TModerator = 515.4 °F, 541.7 K TModerator = 515.4 °F, 541.7 K
TFuel = 515.4 °F, 541.7 K TFuel = 535.4 °F, 552.8 K
FTC
Boron Conc. = 1303 ppm Boron Conc. = 1303 ppm
Rod Position = All Rods Out Rod Position = All Rods Out

Power = 0.0% Power = 0.0%

TModerator = 515.4 °F, 541.7 K TModerator = 515.4 °F, 541.7 K
TFuel = 515.4 °F, 541.7 K TFuel = 515.4 °F, 541.7 K
Boron
Boron Conc. = 1303 ppm Boron Conc. = 1313 ppm
Rod Position = All Rods Out Rod Position = All Rods Out

* For SIMULATE5 the ITC and MTC calculations are performed at the midpoints between
cross-section library temperatures (603.7, 567.4, 531.1, 469.1°F). This is done to
eliminate the cross-section interpolation noise. The calculated value presented in Table 4
is linearly interpolated from the two nearest calculation points. A linear interpolation was
deemed appropriate as the MTC behavior in that region displayed linear behavior.

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 9/12
Knoxville, Tennessee, USA April 15-20, 2012
 Brandon Haugh, Ali Mohamed

The core reactivity prediction results are shown in Table 3. The results compared between the
two code systems show very good agreement for rodded and unrodded conditions.

Table 3: Core Reactivity Calculation Results

Calculation CASMO5/SIMULATE5 MCNP5 Difference
ARO 0.99999 1.00014 ± 0.00005 -15 pcm
Bank 1@ 50 % 0.98940 0.98956 ± 0.00006 -16 pcm
Bank 1@ 0 % 0.97463 0.97493 ± 0.00006 -30 pcm

The reactivity coefficient prediction results are shown in Table 4. The results compare well
between the two codes.

The results for the reactivity perturbations in the model, boron concentration, ITC, MTC and
FTC, show very good agreement. However, these differences lie slightly outside the range of the
uncertainty in the Monte Carlo calculations. The root cause of these differences could be
attributed to several factors. One reason is how the temperature of the different cells (fuel,
moderator, etc.) within the MCNP model was treated. While thermal scattering law data is used
for the bound hydrogen in the water moderator. Free gas scattering is used to model the other
materials. The temperature of the free gas treatment for these materials is specified using the
temperature card (TMP), in units of MeV. In this study the TMP card is not used and the
temperatures default to 293K.

The MCNP5 calculations for some of the perturbations have small resulting reactivity changes.
Here the resulting the uncertainties in the MCNP5 results begin to contribute significantly
compared to the magnitude of the coefficients. For example in the MTC calculation with a
perturbation of 20°F the corresponding coefficient value is -2.50 pcm/°F. If we look at the range
of values based on the 99 percent keff (0.99949 to 0.99978) confidence interval for MTC, the
results range from -1.8 to -3.25 pcm/°F. While the first thought may be to perform a larger
perturbation there is a limit as water density becomes much less linear with the approach to
saturation.

Given some of these modeling choices and uncertainty limitations the results still agree well.

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 10/12
Knoxville, Tennessee, USA April 15-20, 2012
 Comparison of NuScale SMR Conceptual Core Design Using CASMO5/SIMULATE5 and MCNP5

Table 4: Reactivity Coefficient Calculation Results

Coefficient CASMO5/SIMULATE5 MCNP5 Difference

-2.50 pcm/°F -0.46 pcm/°F
(TModerator = 515.4°F, keff =
MTC -2.96* pcm/°F 1.00014 ± 0.00005; TModerator
= 535.4°F, keff = 0.99964 ±
0.00005)
-4.00 pcm/°F -0.62 pcm/°F
(TFuel, TModerator = 515.4°F,
ITC -4.62* pcm/°F keff = 1.00014 ± 0.00005;
TFuel, TModerator = 535.4°F, keff
= 0.99934 ± 0.00005)
-1.53** pcm/°F -0.12 pcm/°F
=-(1.4+1.65)/2
(TFuel = 515.4°F, keff =
1.00014 ± 0.00006; TFuel =
FTC -1.65 pcm/°F 535.4°F, TCladd = 535.4°F
keff = 0.99981 ± 0.00005);
TFuel = 535.4°F, TCladd =
515.4°F keff = 0.99986 ±
0.00005)
-11.91 pcm/ppm 0.88 pcm/ppm
Boron -11.03 pcm/ppm (1303 ppm, keff = 1.00014 ±
0.00005; 1313 ppm keff =
0.99895 ± 0.00005)

* Interpolated value see note from Table 2

** Average value of two FTC calculations: one with the fuel and cladding at 535.4°F and
the other the fuel is 535.4°F and cladding at 515.4°F.

5 CONCLUSIONS

The calculations presented form the first step in assessing state-of-the-art production core design
codes CASMO5 and SIMULATE5 for the NuScale SMR. The comparisons to the continuous
energy Monte Carlo code MCNP5 showed good agreement for the global core wide reactivity
parameters considered. Additional work is needed to quantify the local parameters and the
uncertainties over a wider range of core design parameters.

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 11/12
Knoxville, Tennessee, USA April 15-20, 2012
 Brandon Haugh, Ali Mohamed

6 REFERENCES

1. Joel Rhodes, Kord Smith, and Deokjung Lee, “CASMO5/CASMO5M User’s Manual,” SSP-
07/431 Rev. 0 (August 2007).
2. Tamer Bahadir, Sten-Örjan Lindahl, “Studsvik's Next Generation Nodal Code
SIMULATE5,” 2009 Advances in Fuel Management IV – Hilton Head Island, SC (2009).
3. X-5 Monte Carlo Team, MCNP5 - A General Monte Carlo N-Particle Transport Code,
Version 5, LA- UR-03-1987, Los Alamos National Laboratory, April 2003.
4. Forrest B. Brown, “The Makxsf Code with Doppler Broadening”, LA- UR-06-7002, Los
Alamos National Laboratory, Los Alamos, NM.
5. P. Oblozinsky and M Herman, “Special Issue on Evaluated Nuclear Data File ENDF/B-
VII.0,” Nuclear Data Sheet, Volume 107, Number 12 (December 2006).

2012 Advances in Reactor Physics – Linking Research, Industry, and Education (PHYSOR 2012), 12/12
Knoxville, Tennessee, USA April 15-20, 2012