WG3: Extension of CFD Codes to TwoPhase Flow Safety Problems

1
WG3 Extension of CFD Codes to Two-Phase Flow
Safety Problems

• D. Bestion (CEA, France),
• H. Anglart (RIT, Sweden),
• B.L. Smith, M. Andreani, (PSI,Switzerland)
• J. Mahaffy (PSU, USA),
• E. Komen (NRG, Netherland),
• P. Mühlbauer (NRI, Czech Rep),
• M Scheuerer (GRS, Germany)
• T. Morii, F. Kasahara (JNES, Japan)
• With additional input from E. Laurien, T.
  Watanabe, A. Dehbi

2
Objectives and Content of Work

• The objective is to give orientations for the
  future development and assessment of two-phase
  CFD tools to be used in Nuclear Reactor Safety
  (NRS) problems
• A classification of NRS problems requiring 2
  phase CFD use
• A classification of different modelling
  approaches
• The specification and analysis of needs in terms
  of physical assessment,
• The specification and analysis of needs in terms
  of numerical assessment ( including the selection
  of a matrix of numerical benchmarks)

3
NRS problem which require two-phase CFDwas
replaced by NRS problems where extension of CFD
capabilities to two-phase flow may bring real
benefit

• For each NRS problem
• Description of the issue
• Why extension of CFD to two-phase flow may bring
  real benefit
• Degree of maturity of present two-phase CFD tools
  to treat the problem
• High maturity was applied to the case in which
  sufficient information was available, all related
  phenomena were identified well, and models were
  developed for each phenomenon, improvements may
  be welcome for some of them.
• Medium maturity was applied when a publicised
  background exists, most basic phenomena are
  supposed to be well identified and some models
  exist which require improvements and validation.
• Low maturity was applied to the case in which
  no trusted information was available on the
  validity of existing models.
• References

4
NRS problems where extension of CFD capabilities
to two-phase flow may bring real benefit
5
Classification of Two-phase CFD approaches

• Classification with respect to the averaging of
  basic equations
• Space averaging
• 3D model for porous medium
• 3D model for open medium
• Filtering turbulent scales and two-phase
  intermittency scales
• All turbulent scales are filtered (RANS models)
  Reynolds Averaged Navier Stokes
• Only some scales are filtered (two-phase LES)
  Large Eddy Simulation
• All turbulent scales are predicted (DNS) Direct
  Numerical Simulation
• Phase averaging or field averaging
• Homogeneous for a two-phase mixture
• Two-fluid model
• Multi-field models
• Additional transport equations, which are used
• Transport or turbulent quantities k-?, Rij-?,
• Transport of interfacial area or particle number
  density,
• Use of Interface Tracking/Capturing Technique
• Classification   with respect to
• Eulerian-Eulerian method
• Eulerian-Lagrangian method

6
Classification of 3D CFD models
Single -Phase
Two -Phase
1 phase for porous body
Two-fluid for porous body
1 phase RANS models
Two-fluid for open medium
C CF MD F D ? ?
Two-fluid LES LIS ITM??
1 phase LES models
LES ITM
1 phase DNS models
Two-phase pseudo-DNS with ITM
7
Porous or open medium approaches Some issues
require investigations at both scales or coupling
between scales or use of smaller scales for
developing models for more macroscopic scales
8
Identification of gaps in present models for open
medium

• Specification of averaging /filtering of basic
  eq.
• Identification of the local flow configuration
• Free surface location
• Bubbly flows
• Drag and turbulernt dispersion forces
• multi-scale turbulence
• poly-dispersion
• Wall functions
• Droplet flows
• Droplet dispersion force deposition
• Separate phase flows and free surface flows
• interactions between waves, turbulence and
  interfacial transfers, droplet entrainment
• Droplet entrainment from the wave crests
  Breaking of waves with entrainment of bubbles

9
Ex PTS in two-phase conditions
COSI ECCS injection tests
Important phenomena have to be simulated At
system scale At local scale (RANS) Up to
microscopic scale (DNS?)
10
Identification of gaps in the technology of
coupling CFD with other numerical simulation tools

• CFD must be considered as part of a multi-scale
  and multi-disciplinary approach to reactor
  simulation including
• Coupling CFD with 0D or 1D models
• Coupling CFD in open medium with CFD in porous
  medium
• Coupling with neutron kinetics, structure
  mechanics, fuel thermo-mechanics
• Coupling between two-scales requires further RD
  work on
• Creating an interface for the data exchange
  between the modules
• Solving space discretization problems (e.g.
  staggered or non-staggered grids).
• Solving time discretization of the coupling
  (explicit, implicit coupling)
• Coupling between different physical models
• Defining inlet conditions for the finer module
  based on the variables calculated by the more
  macroscopic module
• Preserving mass balance and energy balance at the
  coupling plane

11
Specification and analysis of needs in terms of
physical assessment (i.e. Validation)

• The physical validation of two-phase CFD models
  will require a two-level matrix
• Global validation for each industrial application
  may use already existing data used for system
  codes SET and IET with limited instrumentation
• Local validation on basic flow experiments, with
  advanced instrumentation in new SET able to
  measure local flow parameters and to validate all
  the local transfers.
• Each item of the list of NRS issues should be
  further analysed to precisely define the needs in
  terms of physical validation.
• CSNI organized a Workshop on Instrumentation In
  Santa-Barbara, (1996). It is recommended to
  organize a new workshop on instrumentation to
  update the State of the art and with special
  emphasis on new local techniques required for
  two-phase CFD tool validation

12
Specification and analysis of needs in terms of
numerical assessment

• Requirements for Numerical schemes
• Robustness
• Adaptation to complex geometry
• Accuracy, reducing numerical diffusion
• Efficiency
• Mass and energy conservation
• Preservation of wave propagation processes
• Well-posedness of the mathematical problem
• Selection of a matrix of numerical benchmarks
• A first list is established
• It is recommended to organize a first series of
  benchmarking exercises for existing two-phase CFD
  tools based on a LINX test, a water sloshing
  test, and a rising Taylor bubble

13
(No Transcript)
14
Concluding remarks

• A rather large variety of approaches
• The degree of maturity ranges from very low
  to acceptable
• Further RD is still required for all of them
• It is recommended to focus on a limited number of
  high priority issues.
• Two recommendations
• a new workshop on instrumentation
• a first series of benchmarking exercises

15
Proposal for a WG3-step 2

• WG3-step 2 should include
• Update of the list of NRS issues for which
  two-phase CFD may bring real benefit and update
  of the evaluation of degree of maturity of
  two-phase CFD tools
• Selection of a limited number of NRS issues
  having a high priority and for which two-phase
  CFD has a reasonable chance to be successful in a
  reasonable period of time.
• Review of the existing database for validation
  of two-phase CFD application to the selected NRS
  problems. Identification of needs for additional
  experimental validation
• Identification of a matrix of numerical
  benchmarks of special interest for the selected
  NRS problems
• Establish the foundation of Best Practice
  Guidelines for two-phase CFD application to the
  selected NRS problems