Transcript edf_0509.ppt

Laminar Kinetic Energy
Models in Code_Saturne
Presented by Clare Turner
Combining the strengths of UMIST and
The Victoria University of Manchester
Presentation Outline
• Introduction to the problem
• Models for implementation
• The first test case
• Progress of the 1st model
• Progress of the 2nd model
• Summary of actions
Combining the strengths of UMIST and
The Victoria University of Manchester
Industrial Application
• Original application: rear wing of a Formula 1 car
• Can be applied to any transition case
• Examples:
 Pressure drop in fuel assemblies
 Efficiency of combustion process
 Turbomachinery
 Wind turbine blades
Combining the strengths of UMIST and
The Victoria University of Manchester
Model Equations – 2005 model
Walters-Leylek 2005
Combining the strengths of UMIST and
The Victoria University of Manchester
Model Equations – 2008 model
Walters-Cokljat 2008
Combining the strengths of UMIST and
The Victoria University of Manchester
T3 Flat Plate Tests
Combining the strengths of UMIST and
The Victoria University of Manchester
Boundary conditions
Inlet conditions determined by FSTI curve in STAR-CD
Inlet conditions:
k = 0.0536 m2/s2 ; ε = 1.35 m2/s3 ; U = 5.08 m/s
Combining the strengths of UMIST and
The Victoria University of Manchester
The 2005 Model: Initial Results
Combining the strengths of UMIST and
The Victoria University of Manchester
Corrections in the Code (1)
• Incorrect wall boundary conditions
Combining the strengths of UMIST and
The Victoria University of Manchester
Corrections in the Code (2)
• Inlet value of ω incorrect; ω ≡ ε/kT
• Error in laminar kinetic energy transport equation
• The lack of diffusion caused convergence problems
Combining the strengths of UMIST and
The Victoria University of Manchester
Correction of the 2005 Model (1)
Combining the strengths of UMIST and
The Victoria University of Manchester
Corrected Version of the 2005 Model
Combining the strengths of UMIST and
The Victoria University of Manchester
Further Work Required
• This test case has zero pressure gradient so fΔP
term has been ignored
Combining the strengths of UMIST and
The Victoria University of Manchester
Further Work Required
• The next test case has a pressure gradient,
therefore the change in Ω must be calculated
• Currently have Ω stored as a common – similar to
how S2KW is stored from visklw.F routine
• Is there a simple way of determining whether the
gradient is positive at each cell?
• Also concerns about adapting to parallel
computations
Combining the strengths of UMIST and
The Victoria University of Manchester
Differences Between Models
2005 Model
Combining the strengths of UMIST and
The Victoria University of Manchester
2008 Model
The 2008 model: Test
Combining the strengths of UMIST and
The Victoria University of Manchester
Areas For Error
• From experience, the low profile reflects
incorrect ω at the wall
• Looking at routines which affect this
• Intending to look at budgets, ω values are
smaller than 2005 model
Combining the strengths of UMIST and
The Victoria University of Manchester
Summary
• Including two kL-kT-ω models (ITURB=61)
• For the 2004 version require inclusion of fΔP
term – require help!
• Errors in 2008 model, need to contact
authors but also may be problems in routines
– have a list of edited routines
Combining the strengths of UMIST and
The Victoria University of Manchester
Any Questions?
Combining the strengths of UMIST and
The Victoria University of Manchester