Content uploaded by Blair Bromley
Author content
All content in this area was uploaded by Blair Bromley on Oct 16, 2019
Content may be subject to copyright.
UNRESTRICTED / ILLIMITÉ
COMPARISON OF MCNP AND
WIMSAECL/RFSP CALCULATIONS
AGAINST CRITICAL HEAVY WATER
EXPERIMENTS IN ZED2
WITH CANFLEXLVRF AND
CANFLEXLEU FUELS
Blair P. Bromley, D. G. Watts, J. Pencer,
M. Zeller, and Y. Dweiri
AECL –Chalk River Laboratories
M&C 2009, Saratoga Springs, NY
May 4, 2009
UNRESTRICTED / ILLIMITÉ 2
Introduction (1)
•ACR1000 is nextgeneration reactor
•Evolutionary design from CANDU6 reactor (PHWR).
•Pressure tube, heavywater moderated, H2Ocoolant.
•Modular design, reduced capital costs, enhanced safety.
•Canadian and international markets.
•Physics Toolset for Design/Safety Analyses
–WIMSAECL 3.1 2D transport code, lattice physics
–RFSP 304 / 3.5 3D diffusion model, core physics
–DRAGON 3.04Bb 3D supercell transport, control devices
–MCNP5: Benchmark comparisons
•8 Postulated Accident Scenarios / 15 Reactor Physics Phenomena
–All scenarios and phenomena must be analyzed and addressed.
–Code biases and uncertainties for phenomena must be determined.
UNRESTRICTED / ILLIMITÉ 3
Introduction (2)
•Validation of codes required for ACR1000 application
•Follow standards CSA N286.799, ANS19.32005
•Determine biases () and uncertainties () for reactivity coefficients.
•Determine systematic and random errors in flux/power distributions.
•Validation / Experimental Data Base
•ZED2 critical experiments
•Fuel / lattices similar to ACR1000.
•Mixed lattice CANFLEXLVRF / CANFLEXLEU substitution experiments.
•24cm pitch, H2O, air and checkerboard cooling.
•Fluxmap measurements
•Data to address PH01 (coolant void reactivity) and PH14 (flux prediction)
•Use for validating MCNP, WIMSAECL, and RFSP
UNRESTRICTED / ILLIMITÉ 4
ACR Physics Toolset
•WIMSAECL –Lattice Physics
–WIMSAECL 3.1, E65LIB nuclear data library (ENDF/BVI.5)
–2D Integral Neutron Transport, up to 89 groups
–Models heterogeneous lattice cell in fine detail
–Improved resonance treatment / selfshielding, multicell capability.
–Generates homogenized, 2group data for RFSP
–Generates data for use in DRAGON
•RFSP 304 / 3.5 –Core Analysis
–3D, Diffusion Theory, 2 groups, statics and dynamics
–Models core flux and power distributions, core burnup, transients
•DRAGON 3.04Bb –Supercell Calculations
–3D Multigroup Neutron Transport; incremental cross sections
–Models effect of reactivity devices between channels
UNRESTRICTED / ILLIMITÉ 5
ACR Physics Toolset
•MCNP5 –Benchmarking
–Version 1.30/1.40 with ENDF/BVI Release 8 nuclear data library
–3D static probabilistic / stochastic neutron transport
–Dualpurpose use
–Analysis to complement WIMSAECL/DRAGON/RFSP calculations.
–Benchmarking and extension of validation data base to ACR1000
design and operational conditions.
UNRESTRICTED / ILLIMITÉ 6
Governing Physics Phenomena
•15 reactor physics phenomena that arise during progression of
postulated accident scenarios in ACR1000
–PH01: CoolantDensityChange Induced Reactivity
–Addressed in this study
–PH02: CoolantTemperatureChange Induced Reactivity
–PH03: ModeratorDensityChange Induced Reactivity
–PH04: ModeratorTemperatureChange Induced Reactivity
–PH05: ModeratorPoisonChange Induced Reactivity
–PH06: ModeratorPurityChange Induced Reactivity
–PH07: FuelTemperatureChange Induced Reactivity
–PH08: FuelIsotopicCompositionChange Induced Reactivity
UNRESTRICTED / ILLIMITÉ 7
Governing Physics Phenomena
•Continued
–PH09: RefuellingInduced Reactivity
–PH11: DeviceMovement Induced Reactivity
–PH12: Prompt/Delayed Neutron Kinetics
–PH13: FluxDetector Response
–PH14: Flux and Power Density Distribution in Space and Time
–Addressed in this study.
–PH15: LatticeGeometryDistortion Reactivity Effects
–PH17: Moderator Level Change Induced Reactivity
–PH10: FuelStringRelocationInduced Reactivity (CANDU6)
–PH16: CoolantPurityChangeInduced Reactivity (CANDU6)
UNRESTRICTED / ILLIMITÉ 8
Experimental Data Base:
ZED2 Critical Facility
•Tanktype critical facility, 3.3 m diameter & depth
–Moderator height adjusted to control criticality and power.
–Power level ~ 100  200 Watts.
UNRESTRICTED / ILLIMITÉ 9
Test Fuels  CANFLEX
•CANFLEXTM (CANdu FLEXible Fueling –AECL/KAERI)
–43element fuel bundle (~50 cm long)
–Central 7 pins larger
–Slightly enriched UO2
•CANFLEXLEU (aka CANFLEXSEU)
–0.95 wt% 235U/U (low enriched uranium)
•CANFLEXRU
–0.96 wt% 235U/U (recovered uranium)
–almost identical to CANFLEXLEU
•CANFLEXLVRF (Low Void Reactivity Fuel)
–1.0 wt% 235U/U (outer 42 pins)
–15 wt% Dy/U poison central pin
–Dysprosia in natural uranium dioxide.
UNRESTRICTED / ILLIMITÉ 10
Coolant / Voiding Experiments
•Mixedlattice Substitution Experiment
–Room temperature (~25C)
–24cm square lattice pitch
–Aluminum pressure tube / calandria tube.
–12 central channels (substitution region)
–CANFLEXLVRF (3) + CANFLEXLEU (2)
–40 reference channels
–22 x CANFLEXLEU (5)
–16 x CANFLEXLEU (4) + CANFLEXRU (1)
•All 52 channels with same coolant
–H2O (cooled), or Air (voided)
•Measurements
–Mixed lattice results
–Critical moderator heights
–Flux distributions
–Copper foil (Cu63) activation
UNRESTRICTED / ILLIMITÉ 11
Checkerboard Cooling Experiments
•H2O and air coolants
–Alternating channels
–40 Reference channels
–CANFLEXLEU
–12 Test channels
–CANFLEXLVRF
•Measurements
–Mixed lattice results
–Critical moderator heights
–Flux distributions
–Axial, radial
–Copper foil (Cu63) activation
Light Water Coolant
Air Coolant
Light Water Coolant
Air Coolant
UNRESTRICTED / ILLIMITÉ 12
Critical Height Measurements
•Critical Heights:
–Aircooled: Hc ~ 154 cm
–Checkerboard: Hc ~ 174 cm
–H2Ocooled: Hc ~ 222 cm
•Exposed fuel:
–45 cm to 110 cm of unmoderated fuel
–Affects extrapolation distance; reactivity contribution.
Case
#
Coolant
Moderator
Purity
(wt%
D2O)
Core
Temperature
(C)
Critical
Height
(cm)
Approx. #
Unmoderated
Bundles
7
air
99.125
25.16
154.25
2.2
14
CHKB
99.119
25.57
173.93
1.8
21
H2O
99.104
25.55
221.51
0.9
UNRESTRICTED / ILLIMITÉ 13
FluxMap Measurements
•Cufoil activation
•Axial and radial distributions
UNRESTRICTED / ILLIMITÉ 14
MCNP Model of ZED2
•MCNP5 Version 1.30/1.40, ENDF/BVI.8
•Explicit modeling of entire ZED2 reactor
•30 million to 300 million histories
–keff < 0.1 mk (10 pcm)
UNRESTRICTED / ILLIMITÉ 15
WIMSAECL Models
•WIMSAECL 3.1
–89group ENDF/BVI.5, VI.7
–Collision probabilities.
•Multicell models
–2x5 for core/reflector cells
–1x1 for interior (single coolant)
–2x1 for interior (checkerboard)
•Reflector / aluminum
–Annular supercell model.
•Data Collapse
–Spatially homogenized over lattice cells.
–2group cross sections.
–Use data in RFSP core models.
CANFLEX

LVRF Channels
with CANFLEX

LEU
CANFLEX

LEU Channels
CANFLEX

LEU Channels
with CANFLEX

RU
CANFLEX

RU
CANFLEX

LEU in
Channels
with CANFLEX

P
P
P
P
CANFLEX

LVRF 1
P
P
P
P
CANFLEX

LEU 1
R
R
V
V
P
P
P
P
P
P
P
P
P
P
CANFLEX

LEU 2
CANFLEX

LEU 3
R
R
V
V
P
P
P
P
P
P
P
P
P
P
CANFLEX

RU 1
P
Denotes periodic boundary conditions
R
Denotes reflecting boundary conditions
V
Denotes void boundary conditions
UNRESTRICTED / ILLIMITÉ 16
RFSP Model of ZED2
•RFSP 304 (3D, 2group diffusion, finite difference, Cartesian)
–Model to extent of graphite reflector outer boundary; void (vacuum) beyond.
–Upper axial boundary condition for moderator level.
–Void BC, or use experimentallybased extrapolation distance (from flux data)
–Special test models set up for unmoderated fuel region.
–Use DRAGON (rz) models to get 2group xsec for unmoderated fuel
24 cm
Mesh line
Lattice line
Graphite reflector
Aluminium calandria
D2O reflector
24 cm
Mesh line
Lattice line
Graphite reflector
Aluminium calandria
D2O reflector
Moderator
Level
UNRESTRICTED / ILLIMITÉ 17
keff Results –Mixed Lattice
•keff biases
–MCNP: –7 mk to –10 mk, more negative with cooling.
–WIMSAECL/DRAGON/RFSP; choice of BC
–Vacuum: 3 mk to –5 mk (neglects reactivity effect of unmod fuel)
–Exp. BC: 0.5 mk to –3.5 mk (extrapolation distance slightly corrects)
–Explicit modeling of unmoderated fuel: 1.5 mk to +3 mk
–Likely overestimate of reactivity effect of unmod fuel.
–Sharp change in diffusion at interfaces; strong anisotropy.
–More challenging to achieve convergence on flux solution.
keff
Model
Air
Ckbd
H2O
CVR
(mk)
CBCVR
(mk)
Uncertainty
(mk)
MCNP
0.99344
0.99229
0.99019
+3.3
+2.1
0.2
RFSP  Vacuum BC
0.99455
0.99746
0.99615
1.6
+1.3
0.2
RFSP  Exp. BC
0.99648
0.99944
0.99683
0.4
+2.6
0.2
RFSP + Unmod Fuel
1.00108
1.00307
0.99864
+2.4
+4.4
0.2
UNRESTRICTED / ILLIMITÉ 18
CVR / CBCVR Results
Mixed Lattice
•Coolant Void Reactivity (full and checkerboard)
–CVR = (kvoid  kcool)/(kvoid kcool) , CBCVR = (kckbd –kcool)/(kckbd kcool)
–MCNP: CBCVR is less than CVR
–WIMSAECL/DRAGON/RFSP: results depend on choice of upper axial BC.
–CVR ~ 1 mk to 5 mk below MCNP result.
–CBCVR is ~ 2 mk to 3 mk above CVR, comparable to MCNP.
–Development of alternative methods to account for effect of unmoderated fuel
–Use of adjusted moderator heights under consideration.
–Unmoderated fuel represented by increasing moderator height.
keff
Model
Air
Ckbd
H2O
CVR
(mk)
CBCVR
(mk)
Uncertainty
(mk)
MCNP
0.99344
0.99229
0.99019
+3.3
+2.1
0.2
RFSP  Vacuum BC
0.99455
0.99746
0.99615
1.6
+1.3
0.2
RFSP  Exp. BC
0.99648
0.99944
0.99683
0.4
+2.6
0.2
RFSP + Unmod Fuel
1.00108
1.00307
0.99864
+2.4
+4.4
0.2
UNRESTRICTED / ILLIMITÉ 19
Axial Flux Distributions
H2Ocooled Mixed Lattice
•MCNP within 1%, RFSP within 2%
–Slightly larger errors for RFSP near top of core.
H2OCooled CANFLEXLVRF in CANFLEXLEU, 25 C, K2W
0.00
0.20
0.40
0.60
0.80
1.00
1.20
050 100 150 200 250
Axial Position (cm)
Cu Foil Activation
(Normalized to K0)
Experiment
MCNP
WIMSAECL/RFSP
UNRESTRICTED / ILLIMITÉ 20
Radial Flux Distributions
H2Ocooled Mixed Lattice
•MCNP within 1%, RFSP within 3%
–Slightly larger local errors in reflector region.
H2OCooled CANFLEXLVRF in CANFLEXLEU 25 C, z=100 cm
0.40
0.50
0.60
0.70
0.80
0.90
1.00
1.10
150 100 50 0 50 100 150
Radial Position (cm)
Cu Foil Activation
(Normalized to K0)
Experiment
MCNP
WIMSAECL/RFSP
UNRESTRICTED / ILLIMITÉ 21
Axial Flux Distributions
Aircooled Mixed Lattice
•MCNP within 1%, RFSP within 3%
–Slightly larger errors for RFSP near top/bottom of core.
AirCooled CANFLEXLVRF in CANFLEXLEU, 25 C, K2W
0.00
0.20
0.40
0.60
0.80
1.00
1.20
020 40 60 80 100 120 140 160
Axial Position (cm)
Cu Foil Activation
(Normalized to K0)
Experiment
MCNP
WIMSAECL/RFSP
UNRESTRICTED / ILLIMITÉ 22
Radial Flux Distributions
Aircooled Mixed Lattice
•MCNP within 1%, RFSP within 3%
–Slightly larger local errors for RFSP near core/reflector interface.
AirCooled CANFLEXLVRF in CANFLEXLEU, 25 C, z=70 cm
0.50
0.60
0.70
0.80
0.90
1.00
1.10
150 100 50 0 50 100 150
Radial Position (cm)
Cu Foil Activation
(Normalized to K0)
Experiment
MCNP
WIMSAECL/RFSP
UNRESTRICTED / ILLIMITÉ 23
Axial Flux Distributions
Checkerboard Mixed Lattice
•MCNP usually within 1%, RFSP within –5% to +2%
–Neutron spectrum changes between checkerboard lattice cells.
CheckerboardCooled CANFLEXLVRF in CANFLEXLEU, 25 C, K2W
0.00
0.20
0.40
0.60
0.80
1.00
1.20
020 40 60 80 100 120 140 160 180
Axial Position (cm)
Cu Foil Activation
(Normalized to K0)
Experiment
MCNP
WIMSAECL/RFSP
UNRESTRICTED / ILLIMITÉ 24
Radial Flux Distributions
Checkerboard Mixed Lattice
•MCNP within 1%, RFSP within 3%
–Slightly larger local errors for RFSP near core/reflector interface.
CheckerboardCooled CANFLEXLVRF in CANFLEXLEU
25 C, z=70 cm
0.50
0.60
0.70
0.80
0.90
1.00
1.10
150 100 50 0 50 100 150
Radial Position (cm)
Cu Foil Activation
(Normalized to K0)
Experiment
MCNP
WIMSAECL/RFSP
UNRESTRICTED / ILLIMITÉ 25
Substitution Analysis
•Neutron Production Correction Factor (NPCF)
–Adjusts Monte Carlo weight of neutron; force keff=1.000; get for ref, test.
Full Core
Reference Lattice
Reference Lattice +
Test Fuel
Large Core of Test Fuel
(Bare, Reflected, or Driven)
Adjust:
1) NPCFref
Until:
keff=1.000
B2MCNP should match B2exp
(r,z)MCNP should match (r,z)exp
Adjust:
1) NPCFtest
Until:
keff=1.000
Adjust:
1) Size
2) NPCFbooster (if necessary)
Until:
keff=1.000
B2ref
B2test
UNRESTRICTED / ILLIMITÉ 26
MCNP Bare Lattices of
CANFLEXLVRF
•Repeated substitution experiments
–Reduces uncertainty in NPCFtest
–For a pure singlefuel lattice, keff ~ 1 / NPCF
UNRESTRICTED / ILLIMITÉ 27
CANFLEXLVRF
Bare Core Dimensions / Bucklings
•Bare cores with with 3 different heights
–4bundle, 5bundle, 10bundle high cores (vary axial, radial leakage)
–Radius adjusted (with NPCFtest) applied so that keff=1.000000.00005
–MCNP tallies of flux curve fitted
–(r,z) = A0cos((zzmax))J0(r), B2= 2 + 2.
–Use for WIMSAECL standalone validation to calculate keff
–Slight change in total buckling (~0.1 m2) with core aspect ratio.
Coolant
Critical Dimensions
Buckling
Hc (cm)
Rc (cm)
B2z (m2)
B2r (m2)
B2 (m2)
B2 (m2)
Air
198.04
252.30
2.369
0.882
3.250
0.015
Air
247.55
178.34
1.535
1.759
3.294
0.015
Air
495.10
136.78
0.395
2.971
3.366
0.015
H2O
198.04
366.50
2.390
0.422
2.812
0.013
H2O
247.55
207.00
1.547
1.321
2.868
0.013
H2O
495.10
148.62
0.393
2.539
2.932
0.013
UNRESTRICTED / ILLIMITÉ 28
keff Calculations for
CANFLEXLVRF Bare Cores
•keff biases:
–MCNP: –4 mk (voided) to –7 mk (cooled), 0.4 mk
–RFSP bias: –4 mk to –6.5 mk, 0.4 mk (varies with aspect ratio),
–Diffusion coefficients isotropic
–WIMSAECL: –6 mk to –8 mk, 0.6 mk (varies with aspect ratio)
–Anisotropy accounted approximately by Benoist diffusion coefficients
Coolant
Critical Dimensions
MCNP
WIMSAECL/RFSP
WIMSAECL
Hc (cm)
Rc (cm)
keff
keff
keff
keff
keff
keff
Air
198.04
252.30
0.99584
±0.00035
0.99612
±0.00035
0.99334
±0.00058
Air
247.55
178.34
0.99584
±0.00036
0.99503
±0.00036
0.99300
±0.00059
Air
495.10
136.78
0.99584
±0.00037
0.99346
±0.00037
0.99217
±0.00058
H2O
198.04
366.50
0.99320
±0.00025
0.99598
±0.00025
0.99282
±0.00039
H2O
247.55
207.00
0.99320
±0.00026
0.99501
±0.00026
0.99230
±0.00039
H2O
495.10
148.62
0.99320
±0.00027
0.99377
±0.00027
0.99184
±0.00039
UNRESTRICTED / ILLIMITÉ 29
CVR Biases for CANFLEXLVRF
•Coolant Void Reactivity (CVR) Bias
–(kvoid –kcool)/ (kvoid x kcool)
–MCNP: +2.7 mk 0.4 mk (no variation with core aspect ratio)
–WIMSAECL: +0.3 mk to +0.7 mk 0.7 mk (slight variation)
–Neutron anisotropy and leakage effects.
–RFSP: 0.3 mk to +0.1 mk (slight variation)
–Within uncertainties of standalone WIMSAECL results, consistent
–2 mk to 3 mk below MCNP results
MCNP
WIMSAECL/RFSP
WIMSAECL
Hc (cm)
CVR
(mk)
CVR
(mk)
CVR
(mk)
CVR
(mk)
CVR
(mk)
CVR
(mk)
198.04
+2.7
0.4
+0.1
0.4
+0.5
0.7
247.55
+2.7
0.4
+0.0
0.4
+0.7
0.7
495.10
+2.7
0.4
0.3
0.4
+0.3
0.7
UNRESTRICTED / ILLIMITÉ 30
Conclusions
•ACRtype experiments performed in ZED2
–24cm pitch, H2O, air, checkerboard coolants
–CANFLEXLEU / CANFLEXLVRF mixed lattice substitution experiments
–Critical height and flux map measurements
–Substitution analysis to isolate properties of CANFLEXLVRF
–Bare core sizes, bucklings determined
•Validation of MCNP, WIMSAECL, WIMSAECL/RFSP
–Mixed lattice CVR Bias < +3.3 mk, CBCVR bias < +2.1 mk
–RFSP results comparable to MCNP, depending on upper BC.
–Determination of adjusted moderator heights would probably help.
–Excellent agreement on flux distributions
–MCNP within ~1%, RFSP within ~3%
–Isolated CANFLEXLVRF properties:
–MCNP CVR Bias ~ +2.7 mk
–WIMSAECL and RFSP comparable to each other, 2 mk to 3 mk below MCNP.
UNRESTRICTED / ILLIMITÉ 31
Acknowledgements
•Jerry McPhee, Greg Morin  ZED2 Experiments
•Danila Roubtsov, Dimitar Altiparmakov –WIMSAECL
•Peter Schwanke, Wei Shen –RFSP
•Tony Liang  WIMS Utilities
•Fred Adams –MCNP
•Jingliang Hu, Michaela Ovanes –ACR Physics
•James Sullivan, Peter Boczar, Dave Wren  Management
•Michele Kubota, Diane Heideman –Administrative Assistants
UNRESTRICTED / ILLIMITÉ 32
References
•1. M.B. Zeller, “Physics Experiments in the ZED2 Reactor in
Support of the Advanced CANDU Reactor,” Proceedings of 25th
CNS Annual Conference, Toronto, June 69 (2004).
•2. B.P. Bromley et al., “Validation of MCNP and WIMS
AECL/DRAGON/RFSP for ACR1000 Applications,” Proceedings of
PHYSOR 2008, Interlaken, Switzerland, September 1419 (2008).
•3. X5 Monte Carlo Team, MCNP –A General Monte Carlo
NParticle Transport Code, Version 5, LAUR031987, Los Alamos,
NM, April (2003).
•4. D.V. Altiparmakov, “New Capabilities of the Lattice Code
WIMSAECL,” Proceedings of PHYSOR 2008, Interlaken,
Switzerland, September 1419 (2008).
UNRESTRICTED / ILLIMITÉ 33
References
•5. T. Liang et al., “Improvement and Qualification of WIMS
Utilities,” Proceedings of 29th CNS Annual Conference, Toronto,
June 14 (2008).
•6. W. Shen et al., “Benchmarking of WIMSAECL/RFSP Multicell
Methodology with MCNP for ACR1000 FullCore Calculations,”
Proceedings of PHYSOR 2008, Interlaken, Switzerland, September
1419 (2008).
•7. R.D. Mosteller, “ENDF/BVII.0, ENDF/BVI, JEFF3.1, and
JENDL3.3 Results for the MCNP Criticality Validation Suite and
Other Criticality Benchmarks,” Proceedings of PHYSOR 2008,
Interlaken, Switzerland, September 1419 (2008).
•8. IAEA, Directory of Nuclear Reactors, Volume V, pp. 223224,
Vienna, (1964).
UNRESTRICTED / ILLIMITÉ 34
References
•9.D.G. Watts et al., “Comparison of MCNP Calculations
Against Measurements in Moderator Temperature Experiments
with CANFLEXLEU in ZED2,” Proceedings of 29th CNS Annual
Conference, Toronto, June 14 (2008).
•10. B.P. Bromley et al., “WIMSAECL Validation for 43Element
BruceB Low Void Reactivity Fuel,” Proceedings of 25th CNS
Annual Conference, Toronto, June 69 (2004).
•11. R.S. Davis, “Qualification of DRAGON for Reactivity Device
Calculations in the Industry Standard Toolset,” Proceedings of
25th CNS Annual Conference, Toronto, June 69 (2004).
•12. Y. Dweiri et al., “Comparison of BareLattice Calculations
Using MCNP Against Measurements with CANFLEXLEU in ZED
2,” Proceedings of 29th CNS Annual Conference, Toronto, June 1
4 (2008).
UNRESTRICTED / ILLIMITÉ 35
References
•13. B.P. Bromley et al., “Development and Testing of a MCNP
Based Method for the Analysis of Substitution Experiments,”
Transactions of the American Nuclear Society, Volume 97, pp. 711
712 (2007).
•14. B.T. Rearden, M.L. Williams, and J.E. Horwedel, “Advances in
the TSUNAMI sensitivity and uncertainty analysis codes beyond
SCALE 5,” Transactions of the American Nuclear Society, Volume
92, pp. 760762. (2005)
UNRESTRICTED / ILLIMITÉ 36
ACR Validation Approach (1)
•Validate MCNP against ZED2 experiments
–New approach (alternative to deterministic codes)
•Validate WIMSAECL and WIMSAECL/DRAGON/RFSP against ZED2.
–Traditional approach, supplementary.
•TSUNAMI Analysis of ZED2 experiments / ACR1000
–Analyze broad range of ACRtype experiments in ZED2 (and others)
–Determine sensitivities and uncertainties.
–Adjust cross sections to eliminate / minimize biases for MCNP.
–Apply adjustments to ACR1000 model (TSUNAMI / KENO).
–Determine reactivity coefficients in ACR1000 with and without adjustments.
–Get biases (k) which can be applied to MCNP for ACR1000.
–Propagate remaining irreducible uncertainties (k) .
UNRESTRICTED / ILLIMITÉ 37
ACR Validation Approach (2)
•CodetoCode Comparisons at ACR1000 (WIMSAECL/DRAGON/RFSP vs. MCNP)
–Simulations of ACR1000 at design and operating conditions.
–Evaluation of reactivity coefficients (dkeff/dxi, xi= Tcool, Tfuel, Tmod, ppmGd, etc.)
–Determine codetocode biases (RFSPMCNP)
–kCodetoCode (ACR1000) = kWIMSAECL/DRAGON/RFSP (ACR1000)  k MCNP(ACR1000)
•Combine MCNP Biases with CodetoCode Biases
–kWIMSAECL/DRAGON/RFSP (ACR1000) = kCodetoCode (ACR1000) + kMCNP (ACR1000)
–ANS Standard, ANSI/ANS19.32005, supports approach.
•Supplementary / Complimentary Data
–Direct validation of WIMSAECL/DRAGON/RFSP against ZED2 experiments
–Eg. reactivity device measurements, transient measurements.
–WIMSAECL validation against other measurements
–Burnup/irradiation measurements for fuels in other reactors (NRU, CANDU)
UNRESTRICTED / ILLIMITÉ 39
CANDU Reactor Technology
•D2O Moderator (~70C, low pressure) in calandria.
•D2O Coolant (~10 MPa, 250C–310C)
•Pressure Tubes, Calandria Tubes
•28.58cm square lattice pitch
•Natural uranium fuel (UO2) in bundles
–37element (CANDU6, Bruce, Darlington)
–28element (Pickering)
•Burnup ~ 7,500 MWthday/tonne.
•OnLine Refueling (8 to 12 bundles per day)
•Two independent shutdown systems.
UNRESTRICTED / ILLIMITÉ 40
ACR1000
•Evolution from CANDU6 design
–Pressure tubes (520 channels)
–Heavy water moderator at low temp. (~70C)
–Short fuel bundles (~50 cm) –online refuelling.
–Multiple shutdown systems.
–Balanceofplant similar, but higher steam P, T.
–Higher coolant pressure/temperatures
–3187 MWth / 1085 MWe(net) , 34% net efficiency.
–CANFLEXACR (43element),
–24cm pitch, negative CVR
UNRESTRICTED / ILLIMITÉ 41
ACR1000
Plant Layout
UNRESTRICTED / ILLIMITÉ 42
ACR1000
•Special features
–Light water coolant (11 MPa, 319°C)
–Reduced capital costs.
–CANFLEXACR Fuel Bundle
–43element design; enhanced heat transfer.
–Enriched fuel (2 wt% to 3 wt%), central absorbing pin (Dy).
–20,000 MWd/t burnup (nominal), extend with experience.
–Tighter lattice pitch; larger calandria tubes.
–More compact core; smaller reactor.
–Negative coolant void reactivity.
UNRESTRICTED / ILLIMITÉ 43
ACR1000
•Modular construction, competitive design
–Lower capital costs.
–Local fabrication of components.
–Economical electricity.
UNRESTRICTED / ILLIMITÉ 44
ACR1000
•Layout of Plant
UNRESTRICTED / ILLIMITÉ 45
ACR1000
•Comparison with CANDU6, Darlington
UNRESTRICTED / ILLIMITÉ 46
ACR1000
•Comparison with CANDU6, Darlington
UNRESTRICTED / ILLIMITÉ 47
Compare CANDU and ACR Lattices
•ACR1000 lattice pitch tighter than CANDU6
CANDU 6
Lattice Pitch = 28.58 cm
ACR
Lattice Pitch = 24.0 cm
UNRESTRICTED / ILLIMITÉ 48
CANDU 6 / ACR Fuel Comparison
•Switch to 43element bundle
–Enriched fuel for higher burnups
–Lower linear ratings
–Enhanced heat transfer
–Large central pin with neutron absorbers
UNRESTRICTED / ILLIMITÉ 49
Compare
CANDU6, Darlington, ACR Cores
•> 65% reduction in D2O / MWe
CANDU6 Darlington ACR1000
Number of Channels 380 480 520
Reactor Core Diameter (m) 7.6 8.5 7.6
Lattice Pitch (mm) 286 286 240
Volume of D2O in Moderator (m3)265 312 235
Volume of D2O in HTS ( m3)192 280 0
Total Volume D2O (m3)466 602 240
UNRESTRICTED / ILLIMITÉ 50
Coolant Void Reactivity (CVR)
•Key parameter for safety analysis of ACR1000
–CVR = (keff(cool=0) –keff(cool)) / (keff(cool=0) keff(cool))
•How well can this be predicted?
–CVR Bias = CVRpredicted  CVRactual
•How to determine CVR Bias
–Perform critical lattice experiments (keff = 1.000)
–Test fuel cooled (kcool), test fuel voided (kvoid).
–Obtain B2critical B2cool, B2void
–Evaluate kvoid(B2void), kcool(B2cool) from lattice physics code WIMSAECL
–CVR Bias = (kvoidkcool)/kvoid/kcool with B2cool, B2void
–Ideally: kvoid = 1.000, kcool = 1.000 CVR bias = 0.0
–CVR Bias > 0 code prediction higher than reality, and vice versa
22
infinity
effective B M 1
k
k
UNRESTRICTED / ILLIMITÉ 51
Objective of Validation
•For each reactor physics phenomena
–Coefficients of reactivity (keff/dxi)
–xi= cool, Tcool, mod, Tmod, ppmGd, wt%D2O, Tfuel, etc.
–Flux and power distributions P(x,y,z,t), (x,y,z,t)
•Need to determine:
–Biases ()
–Uncertainties ()
–%RMS Errors (flux distributions)
•Separate effects –single phenomena
•Integral effects –simultaneous phenomena
•Use data to validate:
–WIMSAECL (standalone)
–Use buckling (B2) data, PIE data
–WIMSAECL/DRAGON/RFSP and MCNP
–Simulate entire critical experiment or reactor
UNRESTRICTED / ILLIMITÉ 52
Applicability of Data
•ZED2 data
–ACR physics behavior simulated as closely as possible.
–Lattice design, geometry, material compositions, operational
conditions closely duplicated.
–TSUNAMI analyses done to assess similarity of data and for
sensitivity / uncertainty analysis.
–Uncertainties in calculations due to xsec uncertainties.
–Extension of MCNP validation results at ZED2 to ACR1000.
UNRESTRICTED / ILLIMITÉ 53
Experimental Accuracy
•ZED2 experiments designed to minimize .
–keff < 3 mk (absolute value of keff)
–(keff) < 0.3 mk (differential changes in keff)
–Sufficient for reactivity changes, coefficients.
–Substitution experiments (test fuel in reference lattice)
–Repeated experiments to minimize uncertainties.
–Foil measurements < 1% (typical)
•Biases and uncertainties
–Taken into consideration for ACR design/safety analyses.
UNRESTRICTED / ILLIMITÉ 54
ZED2 FullCore Flux Maps
•CANFLEXLEU with H2O, or air.
•24cm square lattice pitch.
•Buckling determined from curve fits
of Cufoil flux maps
RA DIA L FLU X S HAPE
DISTAN CE FRO M L AT TIC E CEN TR E (c m)
025 50 75 100 125 150 175
REL AT IV E CU64 A CT IVA TIO N
0.0
0.3
0.6
0.9
1.2
1.5
FIT TE D CU RV E
A CT IVA TION DATA
Jo BES SEL
FUN CTIO N
REX
A XIA L F LU X S HA PE
D IS TA N CE A BO VE R EA CT O R FL OO R (c m)
050 100 150 200 250
R EL A TIV E CU 6 4 A CT IV A TI ON
0.0
0.3
0.6
0.9
1.2
1.5
F ITTED CU R V E
A CT IV A TI ON D A TA
CO SIN E
FU N C T IO N
HEX
Buckling = (2.405/REX)2+ (/HEX)2
UNRESTRICTED / ILLIMITÉ 55