qualification of ansys software as a framework of cfd and ... · pdf fileframework of cfd and...

© 2005 ANSYS, Inc. 1 ANSYS, Inc. Proprietary

Qualification of ANSYS Software as a

Framework of CFD and Multi-Physics

Integration for Thermohydraulic Applications

Qualification of ANSYS Software as a

Framework of CFD and Multi-Physics

Integration for Thermohydraulic Applications

Thomas FrankANSYS Germany, 83624 Otterfing

[email protected]

© 2006 ANSYS Germany

Post-FISA-2006 Workshop No. 4: „Qualification of Advanced Numerical Simulation Platforms as Support for Decision Making“, EC, Luxembourgh

16. March 2006

OverviewOverview

• ANSYS CFX & Workbench product overview

• ANSYS CFX Modeling Capabilities

– Single and Multiphase Flows

– Turbulence

– Heat and Mass Transfer / Thermohydraulics

– Code Coupling

• Solver Technology, Performance & Efficiency

• Code Maintenance, Education & Support

• Model Validation

• Applications in Nuclear Reactor Safety (NRS)

• Summary



16. March 2006

ANSYS & CFX in European NRSANSYS & CFX in European NRS

FRAMATOME ANP

FZK

OKB GidropressPodolsk, RU

Josef Stefan Institute,Ljubljana,

SLO

JRC IE



16. March 2006

CFX – Workbench IntegrationCFX – Workbench Integration

Set-up

Mesh

ANSYS FEA

Workbench



16. March 2006

ANSYS WorkbenchANSYS Workbench

• single geometry representation

• full CAD integration

• fully parameterized model

• Unified Meshing Initiative (UMI)



16. March 2006


• meshing, flow setup, CFD solver run from Workbench

• integrated post-processing



16. March 2006


• CFD & CSM geometry

• CFD & CSM meshing

• CFD��CSM 2-way coupling



16. March 2006

HeatHeat

TransferTransfer

SolidSolid

MechanicsMechanics

MagnetismMagnetismFluidFluid

MechanicsMechanics

ElectricityElectricity

ANSYS Multi-Physics

Multiphysics-Environment & FSIMultiphysics-Environment & FSI

1D Codes1D CodesNeutron Neutron

Kinetics Kinetics

(Development)(Development)AcousticsAcoustics



16. March 2006

CFX Single Phase CapabilitiesCFX Single Phase Capabilities

• Steady state & transient simulation

• Sub-, trans- and supersonic flows

• Wide variety of turbulence models (18+):

• Standard k-εεεε

• k-ωωωω, SST

• RSM (k-εεεε and k-ωωωω based)

• Algebraic RSM (EARSM)

• Scale resolving turb. models (LES, DES, SAS)

• …

• Material database, Multi-component fluid mixtures

• Chemical reactions, combustion models,…

• Conjugate heat transfer (CHT walls)

• Radiation



16. March 2006

CFX Multiphase Flow ModelingCFX Multiphase Flow Modeling

Lagrange Particle TrackingEuler-Euler Multiphase

Multiphase Flow CFD

• Continuous ‚carrier‘ phase �� Euler

• Tracking of single particles or particle groups

• Interaction with carrier phase�� 2-way coupling

• Allows for broad size distributions

• Limited to disperse flows

• ODE’s for particle properties

• Interpenetrating continua

• Full N-phase framework (N≤≤≤≤20)

• Phase indicator function

• Phase weighted averaging

• Additional unknowns �� consequence of averaging

• Empirical closure

• PDE transport eq.’s for unknowns



16. March 2006

CFX Multiphase Flow ModelingCFX Multiphase Flow Modeling

Multiphase Flow ModelsMultiphase Flow Models

Homo-geneous

VOF (Ul=Uk)

Homo-geneous

VOF (Ul=Uk)

Heterogeneous

MixturesHeterogeneous

Mixtures

Mixture

ModelMixture

Model

Particle

ModelParticle

Model

Free Surface

FlowsFree Surface

Flows

Inhomo-

geneousVOF (Ul≠≠≠≠Uk)

Inhomo-

geneousVOF (Ul≠≠≠≠Uk)

Monodisperse(length scale dP)Monodisperse(length scale dP)

Polydisperse(MUSIG)

Polydisperse(MUSIG)

CavitationCavitationFluidized Beds/

Kinetic TheoryFluidized Beds/

Kinetic Theory

ASMASM



16. March 2006

CFX Solver Models CFX Solver Models

• Turbulence models, Transition

• Mixture models

• Eulerian multiphase: full N-phase, homogeneous, inhomogeneous,Multicomponent mass transfer

• Free surface models:surface tension, wall contact

• Lagrangian multiphase

• Phase change models: cavitation, boiling, condensation, evaporation, solid/liquid

• Porous media models

• Real fluid models

• User-Fortran, CCL, User Customization

• Heat transfer models, CHT

• Radiation models: P1, DTM, Monte Carlo, multiband, multigrey, specular, anisotropic scattering, multiphase support

• Combustion models: EBU, coupled multistep, kinetic, flamelet, premixed, partially premixed

• High speed flow models

• Transient models

• 2-way FSI (ANSYS WB, MpCCI)

• Moving Mesh

• Multi-domain physics

• More …

• full feature matrix interaction

• fully parallelized (PVM, MPI)

• coupled solver

• algebraic multigrid (AMG)

• single & double precision



16. March 2006

CFD � 1d-Code CouplingCFD � 1d-Code Coupling

• interfaces for CFX with AMESim / GT-Power

• mechanical / hydraulic BCs

• interfaces for ATHLET & RELAP-5 planed

Inflow

Hydraulic BC

Mechanical BC

1d Code

Hydraulic system

mɺ

p

F

d

1d Code

Mechanical system

Outflow



16. March 2006

Example: Opening ValveExample: Opening Valve

Mechanical BC

Hydraulic BC

Opening0 [atm]



16. March 2006

Solver EfficiencySolver Efficiency

• CFX-5 coupled solver

– hydrodynamic system (U-V-W-P system of eq’s)

– coupled chemical reactions

– coupled multiphase

• Algebraic Multigrid Solver

– highly convergent

– multigrid solver used by all models

– enhanced robustness

– effort grows linear with mesh resolution

• Parallelization & High-Performance Computing

– fully and highly efficient parallelized solver (PVM, MPI)

– memory scalability and speed-up

– flexible partitioning methods

– HPC architecture support (with limitations)



16. March 2006

Algebraic Multigrid SolverAlgebraic Multigrid Solver

• Grid refinement:

– constant (linear) performance with grid refinement

-6,00

-5,00

-4,00

-3,00

-2,00

-1,00

0,00

0 10 20 30 40 50 60 70 80 90 100

Work Unit

log

(RM

S R

es

idu

al)

25K Nodes

100K Nodes

400K Nodes

10

• Sensitivity to grid aspect ratios:– aspect ratios: 3, 30, 300

– AMG shows better performance then geometric multigrid

-6,00

-5,00

-4,00

-3,00

-2,00

-1,00

0,00

0 20 40 60 80 100 120

Work Unit

log

(

RM

S R

esid

ual)

3 GMG

30 GMG

300 GMG

3 AMG

30 AMG

300 AMG

10



16. March 2006

Solver Efficiency: ParallelizationSolver Efficiency: Parallelization

Parallel Scaling Using Different Interconnects

20.0

25.0

30.0

35.0

40.0

45.0

50.0

55.0

60.0

65.0

70.0

20 30 40 50 60 70

Number of CPUs

Pa

ralle

l S

pe

ed

up

SMP

100 Base-T

Gigabit

Hyperfabric2

ideal linear

• Test results courtesy of Dr. M. Ehrig, HPTC

• PA 8800 superdome

• 80 million nodes

• Low data transfer

algorithm



16. March 2006

Code Maintenance, Support,

Education & Training

Code Maintenance, Support,

Education & Training

• Fully maintained commercial CFD code�� extensive user documentation and tutorials,

benchmark test cases, examples,…

• Full user support & training@ ANSYS & In-house

• Engaged in BPG for single and multiphase flows�� ECORA, ASTRESA proposal�� ERCOFTAC, QNET-CFD, …

• Education and training:�� 4th Joint FZR & ANSYS CFX Short Course

& Workshop on MPF in NRS, 26.-29. June ’06�� Regular CFX Seminars & Updates



16. March 2006

Qualification of ANSYS

Software


Software

• Software engineering process according to strict guidelines derived in the project SEMPA (SEMPA – Software Engineering Methods for

Parallel Applications)

• maintained Quality Assurance database

• QA defects are tracked until they are corrected by a developer and the solution is approved

• nightly execution of more then 350 regression tests for basic physical models on all ~10 supported hardware platforms �� more then ~3500 simulation runs each day

• results of regression tests are compared to previously evaluated results

• failures / differences in regression tests are marked in a web interface and will be tracked by responsible developers immediately



16. March 2006


Software (cont.)


Software (cont.)

• Internal database of automated validation test suites

– stored with all mesh, input and output data

– references are reliable experimental or analytical data

– all validation material available to customers for own testing purposes on request

• Selected validation tests are repeated on new software versions

• 3 stages of Pre-Release testing:

– Beta testing program with strong user participation

• Functional and acceptance testing of new software versions prior to release by internal testing group

• Participation in national/EU projects focused on validation of CFD models and codes



16. March 2006

Model ValidationModel Validation

• large effort of ANSYS CFX in model validation

• participation in EU / national research projects:

- Project on fluid mixing and flow distribution phenomena in the PWR primary circuit

- Evaluation of Computational Fluid Dynamic Methods for Reactor Safety analysis

- Advanced 3d two-phase flow simulation tool for application to reactor safety

- German CFD Network on Nuclear Reactor Safety Research

… and much more, e.g. QNET-CFD, PRECCINSTA, EXPRO, ALESSIA, FLOMANIA, DESider, Cavitation



16. March 2006

Bubbly Flow Model Validation FZR MT-Loop and TOPFLOW Database

Bubbly Flow Model Validation FZR MT-Loop and TOPFLOW Database

TOPFLOW

MT-Loop



16. March 2006

Validation: Bubbly FlowsTurbulent Dispersion Force

Validation: Bubbly FlowsTurbulent Dispersion Force

0,00

1,00

2,00

3,00

0 5 10 15 20 25

Radius, mm

No

rm. A

ir V

olu

me

Fra

cti

on

k-eps + RPI TD (0.5)

k-eps + FAD TD

SST + RPI TD (0.5)

SST + FAD TD

Data

Model improvement



16. March 2006

Validation: Polydisperse FlowsInhomogeneous MUSIG Model

Validation: Polydisperse FlowsInhomogeneous MUSIG Model

0.0

5.0

10.0

15.0

0.0 25.0 50.0 75.0 100.0

x [mm]

Air

vo

lum

e f

rac

tio

n [

-]

Exp. FZR-074, level C

Exp. FZR-074, level O

Exp. FZR-074, level R

CFX, Inlet level (z=0.0m)

CFX, level O

CFX, level R



16. March 2006

SlugsbetweenValves

Scheme of the ROCOM test facility at FZ Rossendorf

Boron Mixing Experiments at

ROCOM Test Facility, FZR

Boron Mixing Experiments at

ROCOM Test Facility, FZR



16. March 2006

50 s

14 s

40 s

68 s

Transport of Slugs -

Streamlines

Transport of Slugs -

Streamlines

Courtesy of Th. Höhne, FZ Rossendorf



16. March 2006

Transport of SlugsTransport of Slugs

Courtesy ofTh. Höhne, FZ Rossendorf



16. March 2006

Transient Slug MixingTransient Slug Mixing




16. March 2006

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

0 10 20 30 40 50 60 70 80 90 100

time / s

the

ta / -

Experiment

CFX

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

0 45 90 135 180 225 270 315 360

azimuthal position / °

theta

/ -

Experiment

CFX

t=40.0 s

CL 1 CL 4

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 45 90 135 180 225 270 315 360

Azimuthal Position /°

Mix

ing

Sca

lar

/ -

Time=40 sec

CFX

CL-1 CL-4 CL-3 CL-2

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 45 90 135 180 225 270 315 360

Azimuthal Position /°

Mix

ing

Sca

lar

/ -

Time=50 sec

CFX

CL-1 CL-4 CL-3 CL-2

ReactorInlet

Upper Downcomer

Local Θx,y(t) (42.5°) Θϕ at 40 s

Θϕ at 40 s Θϕ at 50 s

Quantitative Results of

Slug Mixing

Quantitative Results of

Slug Mixing




16. March 2006

ECC Injection after SBLOCAECC Injection after SBLOCA

• UPTF demonstration case

• full-scale simulation of the primary system of the four

loop 1300 MWeSiemens/KWU PWR

• Fluid-Fluid mixing in Cold Leg and Downcomer during ECC injection after SBLOCA

• Further UPTF cases

investigated by GRS

Courtesy of Sander Willemsen



16. March 2006

VVER RDB SimulationVVER RDB Simulation

• Analysis of a VVER-1000 RDB (NPP Kozloduy, BG)OECD Coolant Transient Benchmark

• Temperature transient in a single Cold Leg

• Structures > 5mm resolved by the mesh

By Bieder / CEA

General Grid

Interface

By CFX 5Courtesy of M. Böttcher, FZ Karlsruhe

9.8 Mill. 0.8 Mill. 2.6 Mill.



16. March 2006

Initial and Boundary Conditions

281-286 MW Core Power Output (Assumption: linear increase up to t=1800s)

4834268.64

4682268.63

4718268.72

4737268.61

Mass flow

rate [kg/s]

Cold leg temp

[°C]

Loop No ColdLeg Temperatures

540

542

544

546

548

550

552

554

556

558

0 500 1000 1500 2000

time [s]

tem

pe

ratu

re [

K]

ColdLeg 1

ColdLeg 2

ColdLeg 3

ColdLeg 4

VVER SimulationVVER Simulation

Courtesy of M. Böttcher, FZ Karlsruhe



16. March 2006

Temperature Distribution in

VVER during Temp. Transient

Temperature Distribution in

VVER during Temp. Transient

Temp[°C]

mixing coefficient

Cold Leg 3

Core assembly outlet temperature

initial state:

final state:




16. March 2006

Temperature Distribution

in VVER RDB

Temperature Distribution

in VVER RDB




16. March 2006

Temperature at Hot Legs 1-4HotLeg1

544

546

548

550

552

554

556

0 500 1000 1500 2000

Time [s]

Te

mp

era

ture

[K

]

experiment

CFX 2nd order

CFX 1st order

HotLeg 2

544

544,5

545

545,5

546

546,5

547

547,5

548

0 500 1000 1500 2000

Time [s]

Te

mp

era

ture

[K

]

experiment

CFX 2nd order

CFX 1st order

HotLeg 3

544,4

544,6

544,8

545

545,2

545,4

545,6

545,8

546

546,2

546,4

0 500 1000 1500 2000

Time [s]

Te

mp

era

ture

[K

]

experim ent

CFX 2nd order

CFX 1s t order

HotLeg 4

544

544,5

545

545,5

546

546,5

547

547,5

548

548,5

549

549,5

0 500 1000 1500 2000

Time [s]

Tte

mp

era

ture

[K

]

experim ent

CFX 2nd order

CFX 1s t order

Accuracy of

Measurements

+ 2K !

VVER Simulation – Hot Leg

Temperature Transients

VVER Simulation – Hot Leg

Temperature Transients




16. March 2006

• ThAI test facility 60m3

• four stages:

– Helium injection

– vertical steam injection

– horizontal steam injection

– flow dissipation

• 7700s duration of the experiment

• CFX 5.7 simulation

• 105664 grid cells, symmetry assumption

Validation: ISP 47 (Step 2)Validation: ISP 47 (Step 2)

Courtesy of M. Heitsch, GRS



16. March 2006


• Velocity distribution at T=1000s, 3000s, 5500s




16. March 2006

• good agreement of CFX simulation with measurements for 1st

stage (Helium injection)

P1: Pressure in upper plenum

1.0

1.1

1.2

1.3

1.4

1.5

1.6

0 1000 2000 3000 4000

Time (s)

Pre

ssu

re (

bar)

EXPBT

CFXGRS1

DEFJNE1

GASFZK1

GOTAEC1

GOTNAI1

phases

P1: Pressure in upper plenum

1.0

1.1

1.2

1.3

1.4

1.5

1.6

0 1000 2000 3000 4000

Time (s)

Pre

ssu

re (

bar)

EXPBT

ASTGRS1

ASTIRS1

COCGRS1

COCLEE1

COCLEI1

CONIJS1

FUMUPI1

KUPIPP1

MELSTU1

TONIRS1

phases





16. March 2006

H1: Helium concentration in upper plenum

0.0

0.1

0.2

0.3

0.4

0 1000 2000 3000 4000

Time (s)

Heli

um

Vo

lum

e F

racti

on

(-)

EXPBT

CFXGRS1

DEFJNE1

GASFZK1

GOTAEC1

GOTNAI1

phases

H1: Helium concentration in upper plenum

0.0

0.1

0.2

0.3

0.4

0 1000 2000 3000 4000

Time (s)

Heli

um

Vo

lum

e F

racti

on

(-)

EXPBT

ASTGRS1

ASTIRS1

COCGRS1

COCLEE1

COCLEI1

CONIJS1

FUMUPI1

KUPIPP1

MELSTU1

TONIRS1

phases


• good agreement of CFX simulation with measurements for 1st stage (Helium injection)




16. March 2006

Containment Analysis VVER 440-213 (PAKS)


• LBLOCA with H2

release

• 20 passive autocatalytic recombiners (PAR)

• bulk and wall condensation

• pressure peak suppression system

• 25 days on 8 processors (AMD)




16. March 2006



Effect of H2 passive autocatalytic recombiners:




16. March 2006

SummarySummary

• ANSYS CFX – qualified software integration platform

• Multi-physics simulation environment:– CFD, Thermohydraulics

– Fluid-Structure-Interaction (2-way FSI)

– CFD – 1d-Code coupling

– Radiation, Combustion, CFD – EMAG coupling

• Vast manifold of physical models

• Robustness, Scalability & Performance for large applications

• Maintained CFD code, Support & Education

• Large continuous effort in model dev. & validation

• Customer oriented development



16. March 2006

Thank You!

qualification of ansys software as a framework of cfd and ... · pdf fileframework of cfd and...

Documents