experience in neutronic/thermal-hydraulic coupling in...
TRANSCRIPT
Experience in Neutronic/Thermal-hydraulic
Coupling in Ciemat
Miriam Vazquez (Ciemat)
Francisco Martín-Fuertes (Ciemat)
Aleksandar Ivanov (INR-KIT)
2nd Serpent International Users Group Meeting,
Madrid 2012
Outline
1. Introduction
2. Coupling scheme
3. BWR pin benchmark
4. Other applications
5. Conclusions
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 2
Introduction
• This work is framed in nuclear reactor modern analyses:
• High accuracy requirements
• Trend towards multi physics simulations
• Coupled n transport (MC)/ thermal-hydraulics developed worldwide:
• TH feedback
• Exact geometry description
• Energy point-wise cross sections, angular dependency
• MC (+TH) codes are usually the reference for deterministic codes
• MCNP coupled with subchannel codes at pin-by-pin level, but none for
large geometries:
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 3
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 4
Neutronic TH Applied to Who XS temperature dependence
MCNP STAFAS (subchannel)
HPLWR FA (Waata, 2005) Pseudo material
MCNP STAR-CD (CFD) 3x3 PWR fuel pins
(Seker et al., 2007)
5K temperature interval
MCNP COBRA-EN 4 PWR FA (Capellan et al., 2009)
Pseudo materials
MCNP COBRA-TF PWR FA (Sanchez and Al-Harmy, 2009)
Psuedo materials
MCNP ATHAS (subchannel)
CANDU-SCWR FA
(Shan et al., 2010)
Pseudo materials
MCNP5 SUBCHANFLOW LWR pin and FA
(Ivanov et al., 2011,2012)
Pseudo materials
MCNP THERMO(subchannel)
1/8 PWR core
(Kotlyar et al.,2011)
Polynomial expansion
MCNP and TRIPOLI4
SUBCHANFLOW FLICA4
3x3 BWR pins
(Hoogenboom et al., 2010)
Pseudo materials
MCNPX & COBRA
• MCNPX is a stochastic multi-particle Monte Carlo transport code.
• Detailed geometry description, point-wise cross sections.
• Neutron flux estimations for critical and also subcritical systems, as it
includes spallation models.
• keff in critical systems with the KCODE transport mode.
• COBRA-IV is a deterministic code:
• Fluid conservation equations: flow & enthalpy distribution in rod
bundles or cores.
• Heat transfer equation is used to calculate fuel temperature.
• Steady state and transients calculations.
• Different coolants: water, liquid metals or gas.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 5
COBRA-IV Update for Fast Reactors
Applications
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 6
7
MCNPX/ COBRA Coupling
The coupling scheme is based on the iterative use of both codes, steady state
condition:
• MCNPX computes cell power distribution with a flux tally (F4) multiplied
by fission cross section.
• COBRA calculates fuel, cladding and coolant temperature and coolant
density in each cell.
Data are exchanged by means of external file sharing.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 7
Computational Method
• Driver program written in C++
• I/O interface COBRA
• I/O interface MCNP
Requirements:
• Input with the name of the files and cross section libraries.
• MCNP and COBRA input decks with the same discretization.
• Fuel, cladding and coolant of each pin/FA and axial level is described with a cell and have a different material.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 8
Approach for Temperature Effects
(i) Thermal effects in MCNP:
Density → Cell density → ∑(coolant,…)
Temperature → Material cross section → σ(fuel)
• * MCNP adjusts thermal scattering cross section with the TMP card.
• * Doppler broadening of absorption and fission cross section: the library
should be compiled at the suitable temperature.
•
• (ii) Thermomechanical effects (fast reactors):
Cell dimension → Surfaces.
We are working on the dependence of reactor geometry with temperature
to obtain the real reactivity in hot steady state and to calculate expansion
reactivity coefficient (important to license fast reactors under unprotected
transients)
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 9
Cross Section Handling
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012
Alternatives to update the cross section libraries:
Generate explicit cross section libraries with NJOY → Time consuming and memory problems.
On-The-Fly Doppler broadening → Is not yet included in the MCNP release.
Approach using pseudo materials (our choice) → Accuracy depends on the interpolation interval.
The new material composition is a weighted combination of nuclide X cross
section at lower T1 and higher T2 temperature.
In fast reactors, this approach provides deviations 10 times lower with an
interpolation interval of 200 K than with a library 100 K different.
• JEFF 3.1.1 cross section libraries compiled at CIEMAT each 100 K
21
12
12 1 w=w
TT
TT=w
10
Convergence
• Since MCNP adopts a stochastic method, each computed power has an
uncertainty.
• The convergence criterion must be more robust than the uncertainty to
guarantee the convergence
• It is calculated in the first step as the maximum relative uncertainty
among all the MCNP tally results.
• After the first step the temperature convergence is checked in each cell.
• An internal exercise showed that the temperature uncertainty due to the
power uncertainty is 5 times larger in average at pin level and 2 times larger
at core level.
ε<T
TT
i
ii
1
1
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 11
BWR pin benchmark problem
Benchmark data was obtained from:
A.Ivanov, V.Sanchez, J.E. Hoogenboom, “Single BWR pin
coupled Monte Carlo -Thermal Hydraulic analysis”, PHYSOR
2012, Knoxville, April 15-20, 2012
Comparison of results between MCNP-SUBCHANFLOW and
MCNP-COBRAIV coupled codes.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 12
Geometrical model
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 13
Single fuel with reflective radial surfaces.
Diferences in the coupling scheme
MCNP-
SUBCHANFLOW MCNP-COBRA
Initial conditions Converged TH solution from
KENO/SUBCHANFLOW Averaged TH values
Convergence criterion ε=max|ΔTi|<0.15% Δk<σ Adjust N to satisfy this criterion
max|ΔTi|<ε ε=max σqi=0.3%
Temperature dependence of fuel cross section
Using pseudo-materials 50K interval
Using pseudo-materials 100K interval
Themal scattering data Interpolation between data files
No interpolation. Use the closest library available.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 14
Model comparison
MCNP-
SUBCHANFLOW MCNP-COBRA
Number of iterations 14 13
Histories per cycle/ Active cycles
100000/350 500000/350
Fuel temperature Using pseudo-materials 50K interval
Using pseudo-materials 100K interval
Coolant and cladding temperature
Constant Closest available XS library
Two phase correlation Chexal-Lellouche drift flux model
Armand model
Sub-cooled boiling Bowring correlation No sub-cooled boiling model
Fuel conductivity λ=f(T) λ=a+b*T+C*T2 Fits to f(T)
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 15
Power and k∞ during iterations
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 16
Fuel temperature
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 18
2% higher peak temperature
Relative Power
12 % greater peak power predicted by MCNP-COBRA due to
the higher density in the upper part.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 19
Analysis of SFR FA and core
• The coupled code has been used for fast reactors applications with a
detailed geometry at pin level in the ESFR FA and at channel level in the
ESFR and MYRRHA cores.
• In the full core MCNP model, FAs have been grouped in rings of similar
power to complete the calculation in reasonable amount of time (3-4 h
using 128 processors) and not exceeding our machine memory limitations
(2GB).
• Coupling calculations with the working horse ESFR:
ESFR fuel assembly with 271 pins, 30 axial levels. It converges in 3 iterations
Δkeff=182 pcm. Max ΔPower=0.6%
ESFR core with 11 rings, 10 axial levels. It converges in 3 iterations Δkeff=182
pcm. Max Δpower=3%.
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 20
Power density evolution in the central pin of
the ESFR FA.
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 21
Power density in the lower part
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 22
Power density in the upper part
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 23
CR movement with TH feedback in the
ESFR core to achieve criticality
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 24
CR insertion length (cm)
Keff (σ pcm)
0 1.01353 (3)
37.673 0.99844 (3)
34.9133 0.99977 (3)
34.5172 0.99998 (3)
Movement of the 24
Control and Shutdown
rods (CSD)
Linear power per FA in the upper axial level (10th)
Conclusions
• A computational tool has been developed for N-MC/ TH coupled analysis
• Static conditions, 3D temp map effects on keff.
• Flexible geometry modeling. Can be used for full core & subchannel
analysis.
• It has been compared with others MC/TH coupled codes. The discrepancies
are because of the TH model.
• Quick and robust behavior for fast reactors. A method to accelerate
convergence is needed for LWR applications.
• .Memory limitations in full core applications with burned fuel due to the use
of pseudo-materials
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 25
Acknowledgments
- Haileyesus Tsige-Tamirat and Luca Ammirabilie (JRC-IET)
- The co-authors of the BWR pin benchmark problem: V. Sanchez
(KIT) and Prof. Hoogenboom (Delf University of Technology)
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 26
2nd Serpent International Users Group Meeting,
Madrid, September 20, 2012 27
Thank you for your attention!!
References
• [1] C.L. Waata, Coupled Neutronics/Thermal-hydraulica Analysis of a High-Performance Light-
Water Reactor Fuel Assembly, FZKA 7233, Forschungszentrum Karlsruhe, 2005.
• [2] V. Seker, J.W. Thomas, T.J. Downar, others, Reactor Physics Simulations with Coupled Monte
Carlo Calculation and Computational Fluid Dynamics, in: Proceedings of the International
Conference on Emerging Nuclear Energy Systems (ICENES-2007), Istambul, Turkey, 2007: pp.
15–19.
• [3] N. Capellan, J. Wilson, S. David, O. Meplan, J. Brizi, 3D coupling of Monte Carlo neutronics
and thermal-hydraulic calculations as a simulation tool for innovative reactor concepts, in:
Proceedings of Global 2009, Paris, France, 2009.
• [4] V. Sanchez, A. Al-Hamry, Development of a Coupling Scheme Between MCNP and COBRA-
TF for the Prediction of the Pin Power of a PWR Fuel Assembly, in: Proceedings of International
Conference on Mathmatics, Computational Methods & Reactor Physics (M&C 2009), Saratoga
Springs, New York, 2009.
• [5] J. Shan, W. Chen, B.W. Rhee, L.K.H. Leung, Coupled neutronics/thermal–hydraulics analysis
of CANDU–SCWR fuel channel, Annals of Nuclear Energy. 37 (2010) 58–65.
• [6] J.E. Hoogenboom, A. Ivanov, V. Sanchez, C. Diop, A flexible coupling scheme for Monte Carlo
and thermal-hydraulics codes, in: International Conference on Mathematics and Computational
Methods Applied to Nuclear Science and Engineering (M&C 2011), Rio de Janeiro, Brazil, 2011: p.
22.
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 28
References
• [7] A. Ivanov, V. Sanchez, U. Imke, Develpment of a Coupling Scheme between MCNP5 and
Subchanflow for the Pin- and Fuel Assembly-Wise Simulation of LWR and Innovative Reactors, in:
Proceedings of International Conference on Mathmatics, Computational Methods Applied to
Nuclear Science and Engineering (M&C 2011), American Nuclear Society, Rio de Janeiro, Brazil,
2011.
• [8] A. Ivanov, V. Sanchez, U. Imke, Optimization of a Coupling Scheme between MCNP5 and
Subchanflow for high Fidelity Modeling of LWR Reactors, in: Proceedings of International
Conference on M, American Nuclear Society, Knoxville, Tennessee, USA, 2012.
• [9] D. Kotlyar, Y. Shaposhnik, E. Fridman, E. Shwageraus, Coupled neutronic thermo-hydraulic
analysis of full PWR core with Monte-Carlo based BGCore system, Nuclear Engineering and
Design. 241 (2011) 3777–3786.
• [10] A. Ivanov, V. Sanchez, J.E. Hoogenboom, Single Pin BWR Benchmark Problem for Coupled
Monte Carlo- Thermal Hydraulics Analysis, in: Proceedings of PHYSOR 2012, Knoxville,
Tennessee, USA, 2012.
2nd Serpent International Users Group
Meeting, Madrid, September 20, 2012 29