Progress on the v2f model with Code_Saturne

1
Progress on the v2f model with Code_Saturne

• EDF - Manchester meeting
• 18-19th May 2009

2
The model and Code_Saturne
Accuracy

• A low-Reynolds (near-wall integration) eddy
  viscosity model derived from second moment
  closure models
• No damping functions, no wall functions, less
  empirical assumptions
• Best results on range of test cases, heat
  transfer and natural convection in particular.
• The original model is stiff (requires coupled
  solver or very small time-step)
• Degraded version available in StarCD, Fluent,
  NUMECA..
• Long collaboration Stanford, Delft, Chatou,
  Manchester (Durbin, Parneix, Hanjalic, Manceau,
  Uribe) gt several code friendly versions since
  1995.
• Present Reconsider all historical choices with
  numerical stability and known asymptotic states
  as principal objectives

Stanford 1991
TU-Delft 2004
Manchester 2004
Stanford 1996 (Fluent, STAR-CD)
Robustness
3
Starting from the model

• Uribe (2006), Laurence et al. (2004), available
  in CS (ITURB50)
• Same overall good perfomances as the original
• But lack of compliance with asymptotic behaviour
  requirement.
• No - diffusion for (does not
   feel  its B.C.)
• instead of
• Problems reported

Very low value of k
Betts Cavity
Near wall overshooting of
COLD
HOT
Very low value of k
4
Improved code friendly version
Elliptic blending

• Only a few sub-iterations needed to converge
• More robust (B.C. 0)

with

• Unlike , correct near wall behaviour
  of , hence
• No over prediction of the in the core
  region, unlike Lien and Durbin model

5
Prediction of weak turbulence
Case 1 Forced, mixed and natural convection in a
heated pipe (You et al. (2003)). Re180.

• 0.087 Forced/mixed convection
• 0.241 Relaminarisation
• 0.400 Recovery

6
Prediction of weak turbulence
Case 2 Combined natural and forced convection
(Kasagi and Nishimura (1997)) Re150, Gr9.6 105

• Upward flow in a vertical channel
• Turbulent anisotropy enhancement in the buoyancy
  aiding side

7
Improved prediction of by-pass transition

• Near wall adaptation of the equation (near
  wall terms)
• Usual modelling (also used in the )
• Launder Sharma model (1974). E term
  models the term P3 of the exact
  equation.
• Howard (2004), application to a skewed channel
• Latest version of the E term in the
  equation more robust

could be added as well.
8
Results on the T3A flat plate
9
Improvement for High/Low RE
Variables like U, or YdUdY are in fact weakly
Reynolds dependant
But near wall extra terms are generally Re-
dependant
10
Improvement for High/Low Re
11
Improvement for High/Low Re
2008 version near wall tem in the Ep. equation
2008 version near wall tem in the K equation
12
Improvement for High/Low Re
channel flow
13
Application on RAE2822 aerofoil

• Collaboration with Jeremy Benton (AIRBUS).
• Prelimiary tests on a turbulent boundary layer (
  up to 5368) Cf overprediction reported
  with the 2008 and the
  , problem cured with the latest version.

• Two cases case 9 (attached) and case 10
  (separated).
• model tested with a non-linear
  stress-strain relationship (Pettersson-Reif,
  2006) and in an Algebraic Structure based model
  (Kassinos)

• Numerical stability reported to be better than
  the SST model and the

14
Results, Cp, case 9 (attached)
COURTESY OF AIRBUS
15
Results, Cf, upper surface, case 9
COURTESY OF AIRBUS
16
Results, Cp, case 10 (separated)
COURTESY OF AIRBUS
17
Results, Cf, upper surface, case 10
COURTESY OF AIRBUS
18
Problem on the diffuser case

• No separation predicted with the
  model, ill behaviour suspected in hom. Shear
  turbulence. Re-tuning of constants needed.

19
THANK YOU FOR YOUR ATTENTION