Fine-mesh multiphysics of LWRs: two-phase flow challenges and opportunities SKC Symposium October 11-12, 2016, Hindås Klas Jareteg Chalmers University of Technology October 12, 2016
Fine-mesh multiphysics of LWRs: two-phaseflow challenges and opportunities
SKC Symposium October 11-12, 2016, Hindås
Klas Jareteg
Chalmers University of Technology
October 12, 2016
Project overviewProject overviewTwo-phase flow in LWRs
Motivation for CFD methodsTwo-fluid solversSubcooled flow
Future outlook
2
FIRE - FIne-mesh deterministic REactor modeling
Development of a fine meshcomputational tool for nuclear fuelbundles:
• integrated approach for solvingneutronics and thermal-hydraulics
• single and two-phase flow models basedon first principles
• high-resolution coupling on fine meshesusing HPC
• fuel bundle size calculations, ultimatelycoupled to coarse mesh solvers
Figure: Temperature distributionin moderator and fuel.
3
FIRE - Developments during 2011-2012
Figure: Radial temperatureprofile, diffusion neutronics
• September 2012 - June 2013: Masterthesis: ”Development of an integrateddeterministic neutronic/thermalhydraulicmodel using a CFD solver”
• Initiated work on coupledthermal-hydraulics and neutronics
• Developing basics of a cross-sectionmethodology for fine-mesh simulations
• Coupling incompressible CFD solver withdiffusion solver in OpenFOAM-1.6-ext
4
FIRE - Developments during 2012-2013• Starting as a PhD student September 2013• Implemented and studied coupled formulations for diffusion solver• Starting implementation and formulation of transport solver for neutronics• Initiating the development of a two-phase solver based on a population
balance equation (PBE)• Generalizing the geometry and the cross-section algorithms together with
meshing techniques
Figure: Void fraction, bubble mean diameter and axial velocity5
FIRE - Developments during 2013-2014
• Consolidation of the discreteordinates solver, benchmarkingagainst Monte Carlo simulations
• Continued development ofsubcooled boiling and two-phaseflow
• Extended system size simulations
Figure: Moderator temperature
6
FIRE - Developments during 2014-2015
Figure: Emergence of void fractionmeso-scale structures
• Development of transientmultiphysics for single-phase flowwith diffusion and SN neutronics
• Investigations of two-phase flowinstabilities for adiabatic bubblyflow
• Two–fluid/PBE approach forheated bubbly flow withcondensation (DQMOM/MUSIG)
7
FIRE - Developments during 2015-2016
• Extension of two–fluid/PBEapproach for heated bubbly flowwith condensation
• Investigation of phase instabilitiesin two-fluid solvers for 3D cases
• Development and implementationof a verification methodology fortransient neutronic codes Figure: Mean bubble diameter and void
fraction distribution on a heated channel 1
1K. Jareteg, S. Sasic, P. Vinai, C. Demaziere (2016). “A numerical framework for bubbletransport in a subcooled fluid flow”. Submitted to Journal of Computational Physics
8
Two-phase flow inLWRs Project overview
Two-phase flow in LWRsMotivation for CFD methodsTwo-fluid solversSubcooled flow
Future outlook
9
Motivationi. The curse of simplifications
Figure: Flow regimes (Todreas andKazimi 1993)
Flow regimes in a BWR:• Increasing void fraction• Change in characteristic scales• Different heat transfer mechanisms
Tempting not to work from first principles:• Limited mechanistic models, black box
approach• Enough degrees of freedom to close the
equations by experiments• Verify and validate for all existing scenarios
⇒ Existing typical 1D codes
10
Motivationi. The curse of simplifications
Figure: Flow regimes (Todreas andKazimi 1993)
Issues:
• What are the other scenarios?• Cost of unresolved physics:
• Safety margins• Influence on other physics
⇒ Need for fully dimensional, resolvedsimultions, CFD
10
Motivationii. The need for different models
One Method to rule them all...?• Lagrangian particle tracking?• Immersed boundary method?• Volume of fluid?• Two-fluid model in 3D?
No!• Constraints due to computational
cost• Constraints inherent to the
methods
⇒ Need for combinations or newmethods
Figure: Bubble jet simulated withtwo–fluid model
11
Two–fluid model: The Beauty or the Beast?
Advantages:• Affordable cost for larger systems• Computationally scalable• Resolving (in 3D) phase distributions
Issues:• No phase boundaries resolved• Separation of scales?• Stability issues?
Figure: High-level equation format
Figure: Domain decomposition
12
Two-fluid model: Formulation and Implementation
• Mass conservation equation:
∂αiρi
∂t +∇ · (αiρiUi) = 0
• Momentum conservation equation:
∂αiρiUi
∂t +∇ · (αiρiUiUi) = −∇ ·(αi(¯̄τi + ¯̄τ t
i ))−∇(αiP) + αiρig + Mi,
• PISO algorithm, shared pressure
• Interfacial forces including:• drag (Schiller-Naumann)• virtual mass (Cvm = 0.5)
13
Two–fluid model: Stablitiy issues?
• Theoretical literature restricted to 1D models due to:• complexity in mathematical formulation• relevance for the nuclear system codes in 1D
• Previous simulations performed in 2D systems2
• limiting the computational burden• instabilities shown (in multiple software)
• In this stage we perform 3D simulations3:• allowing instablities to grow in third direction.
2K. Jareteg, H. Ström, S. Sasic, C. Demazière (2015). “Numerical investigation of instabilitiesin the two-fluid model for CFD simulations of LWRs”. M&C 2015, Nashville, Tennessee.
3K. Jareteg, H. Ström, S. Sasic, C. Demazière (2016). “On the dynamics of instabilities intwo-fluid models for bubbly flows”. Submitted to Chemical Engineering Science.
14
Computational setupFormulation and parameter variations
Case geometry• Domain size: 10 cm × 40 cm × 10 cm• Mesh size: 32× 64× 32
Boundary conditions• Vapor velocity Ug = (0, 0, 0) m/s,
• Fully periodic• Liquid velocity Ul = (0, 0, 0) m/s
• Fully periodic• Pressure p:
• Axially: jump condition• Horizontally: periodic
15
ResultsVoid fraction, time-resolved uniformity index and velocity4
0.048
0.050
0.052
0.055α
[-]
10−4
10−3
10−2
10−1
100
Φ[-]
0 50 100 150 200Time [s]
0.025
0.050
0.075
0.100
Velo
city
[m/s
]
Ug
Ul
(a) With virtual mass
0.040
0.080
0.120
α[-]
10−5
10−4
10−3
10−2
10−1
100
101
Φ[-]
0 50 100 150 200Time [s]
0.08
0.16
0.24
0.32
Velo
city
[m/s
]
Ug
Ul
(b) Without virtual mass
3K. Jareteg, H. Ström, S. Sasic, C. Demazière (2016). “On the dynamics of instabilities intwo-fluid models for bubbly flows”. Submitted to Chemical Engineering Science
16
ResultsHorizontal effects5
Horizontaldiscontinuity duringgrowth of Φ
Cell-wise discontinuityin void fraction
Meso-scaleheterogeneities largerthan cell size
Horizontaldiscontinuity duringgrowth of Φ
Cell-wise discontinuityin void fraction
Meso-scaleheterogeneities largerthan cell size
Horizontaldiscontinuity duringgrowth of Φ
Cell-wise discontinuityin void fraction
Meso-scaleheterogeneities largerthan cell size
Horizontaldiscontinuity duringgrowth of Φ
Cell-wise discontinuityin void fraction
Meso-scaleheterogeneities largerthan cell size
4K. Jareteg, H. Ström, S. Sasic, C. Demazière (2016). “On the dynamics of instabilities intwo-fluid models for bubbly flows”. Submitted to Chemical Engineering Science
17
ResultsMagnitude of virtual mass force?
-1.50.01.5
F vm,x/F
d,x
[-]
-2.00.02.04.0
F vm,y/F
d,y
[-]
0 50 100 150 200Time [s]
-3.0-1.50.01.5
F vm,z/F
d,z
[-]
Figure: Magnitude of virtual mass force relative to the drag force, presented for eachCartesian component.
18
Extension of two–fluid model: Subcooled flow
Problem: Two–fluid methodology treating vapor and liquid phases asinterpenetrating continua - separate bubbles not resolved
Remedy: PBM• Retrieve bubble size distribution• Size-dependent formulations of
particle forces and condensation
Bubble size
Nu
mb
er
den
sity
Population balance equation:
∂n(x, r, t)∂t +∇x ·
(∂x(x, r, t)
∂t n(x, r, t))
+∇r · (U(x, r, t)n(x, r, t)) = S(x, r, t)
19
Efficient tracking of bubble size distribution
Figure: Mean bubble diameter and void fraction distribution on aheated channel 6
6K. Jareteg, S. Sasic, P. Vinai, C. Demaziere (2016). “A numerical framework for bubbletransport in a subcooled fluid flow”. Submitted to Journal of Computational Physics
20
Summary of work on the two–fluid model
Numerical stability investigations of two–fluid model:• Lack of hyperbolicity also for 3D cases• Virtual mass has a stabilizing effect• Results from two–fluid model trustworthy?
Two–fluid model with PBM:• Retrieve bubble size information from Eulerian approach• Compared PBMs for stability, performance and accuracy
21
Future outlookProject overviewTwo-phase flow in LWRs
Motivation for CFD methodsTwo-fluid solversSubcooled flow
Future outlook
22
Future outlook
Project outlook:• Continued development of multiphase flow methodologies• Coupled, transient neutronics/thermal-hydraulics with multiphase flow
Fine-mesh multiphysics outlook:• Large scale simulations with highly resolved physics growing• Computer science growing in importance (data structures, parallelism,
heterogeneous architecture)
23
Publications and conferences
Publications2016 K. Jareteg, H. Ström, S. Sasic, C. Demazière (2016). “On the dynamics of instabilities in
two-fluid models for bubbly flows”. Submitted to Chemical Engineering Science
K. Jareteg, S. Sasic, P. Vinai, C. Demaziere (2016). “A numerical framework for bubbletransport in a subcooled fluid flow”. Submitted to Journal of Computational Physics
2015 K. Jareteg, S. Sasic, P. Vinai, C. Demazière (2015). “Development of a Coupled Two-Fluid/DQMOM Methodology for Heated Bubbly Flows”. CMFF2015, Budapest, Hungary
K. Jareteg, R. Andersson, C. Demazière (2015). “Development and test of a transient fine-meshLWR multiphysics solver in a CFD framework”. M&C 2015, Nashville, Tennessee
K. Jareteg, H. Ström, S. Sasic, C. Demazière (2015). “Numerical investigation of instabilitiesin the two-fluid model for CFD simulations of LWRs”. M&C 2015, Nashville, Tennessee
E. Pettersen, C. Demazière, K. Jareteg, E. Schönfeldt T, B. Lauritzen (2015). “Developmentof a Monte-Carlo based method for calculating the effect of stationary fluctuations”. M&C 2015,Nashville, Tennessee
H. Ström, S. Sasic, K. Jareteg, C. Demazière (2015). “Behaviour and Stability of the Two-FluidModel for Fine-Scale Simulations of Bubbly Flow in Nuclear Reactors”. International Journal ofChemical Reactor Engineering 13.4, pp. 449–459
K. Jareteg, P. Vinai, S. Sasic, C. Demazière (2015). “Coupled fine-mesh neutronics andthermal-hydraulics - modeling and implementation for PWR fuel assemblies”. Annals of NuclearEnergy, Special Issue: ”Multi-Physics Modeling of LWR Static and Transient Behavior”
A. Jareteg, K. Jareteg, S. Sasic (2015). “Formulation of stresses in dry granular flows”.CMFF2015, Budapest, Hungary
24
2014 K. Jareteg, P. Vinai, C. Demazière (2014). “Fine-mesh deterministic modeling of PWR fuelassemblies: Proof-of-principle of coupled neutronic/thermal–hydraulic calculations”. Annals ofNuclear Energy 68.0, pp. 247 –256
K. Jareteg, P. Vinai, S. Sasic, C. Demazière (2014). “Influence of an SN solver in a fine-meshneutronics/thermal-hydraulics framework”. PHYSOR 2014, September 28 - October 3, Japan
C. Demazière, K. Jareteg (2014). “Developing a course in nuclear reactor modelling and goingfrom campus-based to web-based teaching”. PHYSOR 2014, September 28 - October 3, Japan
H. Ström, S. Sasic, K. Jareteg, C. Demazière (2014). “On the validity of the Two-fluid modelfor simulations of bubbly flow in nuclear reactors”. 13th International Conference on MultiphaseFlow in Industrial Plants (MFIP)
2013 K. Jareteg, P. Vinai, C. Demazière (2013). “Investigation of the possibility to use a fine-meshsolver for resolving coupled neutronics and thermal-hydraulics”. M&C 2013, Sun Valley, Idaho
25
Workshops and conferences
2016 International Conference on Multiphase Flow (ICMF), May 2015 (Presentation title:Numerical simulations of subcooled boiling using a two-fluid/DQMOM methodol-ogy)
2015 Gothenburg Region OpenFOAM User Group Meeting, November 2015 (Presenta-tion titles: ”Increasing performance and cutting computational time - examplesand thoughts on acceleration at different code levels” and ”A Two-Fluid/DQMOMMethodology For Condensation In Bubbly Flow”)
SKC Symposium, Uppsala, October 2015 (Presentation title: ”Transient fine-meshsimulations in LWR fuel assemblies: phenomena and methodologies”)
CMFF 2015, Budapest, Hungary, September 2015 (Presentation title: ”A Two-Fluid/DQMOM Methodology For Condensation In Bubbly Flow”)
SIAMUF Spring meeting, Lund, May 2015 (Presentation title: ”Using an opensource computational framework in research”)
RPNC 2015, Chalmers, Gothenburg, April 2015 (Presentation title: ”Fine-meshmultiphysics modeling within FIRE - Challenges and opportunities with CFDmethodologies in reactor physics calculations”)
26
2014 Gothenburg Region OpenFOAM User Group Meeting, November 2014 (Presenta-tion title: ”Multiphysics simulations of nuclear reactors and more”)
Participation to SIAMUF Meeting, Gothenburg, November 2014
Mid-term presentation of project, Chalmers, November 2014 (Presentation title:”Towards fine-mesh multiphysics simulations of LWR fuel assemblies”)
SKC Symposium 2014, KTH, Stockholm, October 2014 (Presentation title: ”To-wards fine-mesh multiphysics simulations of BWR fuel assemblies”)
PHYSOR 2014, Kyoto, Japan, September 2014 (Presentation title: ”Influence ofan SN solver in a fine-mesh neutronics/thermal-hydraulics framework
Gathering MSc/PhD students - industry, Chalmers, June 2014 (Presentation title:”Multiphysics simulatios of nuclear reactors - Modeling and implementation forfine-mesh simulations”
9th OpenFOAM Workshop, Zagreb Croatia, June 2014 (Presentation and abstracttitles: pUCoupledFoam - an open source coupled incompressible pressure-velocitysolver based on foam-extend and Multiphysics simulations of Light Water Reactorsusing OpenFOAM®
NAFEMS Nordic, Göteborg, May 2014 (Presentation and abstract title: Multi-physics simulations of Light Water Reactors using a CFD approach)
Westinghous Electric Sweden, Västerås, April 2014 (Presentation title: ”FIRE –Project presentation at Westinghouse Sweden, Västerås”
Participation to ”Konferens om lärande och undervisning”, Chalmers, January 2014
27
2013 SIAMUF Fall meeting, ABB, Västerås, November 2013 (Presentation title: ”Mul-tiphase flow simulations for multiphysics simulations for BWRs within FIRE”
Gothenburg Region OpenFOAM User Group Meeting, November 2013 (Presenta-tion title: ”Coupled calculations in OpenFOAM - Multiphysics handling, structuresand solvers”)
SKC Symposium 2013, Gimo, Östhammar, October 2013 (Presentation title: ”FIRE- Status presentation”)
RPNC 2013, Kjeller, Norway (Presentation title: Deterministic fine-mesh coupledneutronics and thermal-hydraulics within FIRE”)
Participation to SIAMUF Spring meeting, Ekerö, Stockholm, June 2013
M&C 2013, Sun Valley, Idaho, May 2013 (Presentation title: ”Investigation of thepossibility to use a fine-mesh solver for resolving coupled neutronics and thermal-hydraulics”
2012 Participation to SIAMUF Fall meeting, Varberg, October 2012
SKC Symposium 2012, Chalmers, Göteborg, October 2012 (Presentation title: De-velopment of an integrated deterministic neutron/thermal-hydraulic model using aCFD solver
SIAMUF Spring meeting, Sundsvall, May 2012 (Presentation title: Development ofan integrated deterministic neutronic/thermal-hydraulic model using a CFD solver
28
Thank you! Questions?
29
References:N. Todreas, M. Kazimi (1993). Nuclear systems I: Thermal hydraulicfundamentals. Taylor & Francis, Levittown, USA
30