help in external aerodynamics(which tubulence model is best)

sat · March 21, 2011, 21:15

hello sir
i am sathish(student) doing project in external aerodynamics
1) can u tell me which turbulence model is best for solving external aerodynamics
2)whether it is necessary to run k-e model before running any turbulence model.if yes why?

cfd_newbie · March 21, 2011, 21:42

Quote:

Originally Posted by sat

hello sir
i am sathish(student) doing project in external aerodynamics
1) can u tell me which turbulence model is best for solving external aerodynamics
2)whether it is necessary to run k-e model before running any turbulence model.if yes why?

Hi Sathish,
1)It totally depend on the kind of physics involved (there is no fixed answer for this), the mach number etc. Please provide more details about the kind of problem you are trying to solve.

2) No it is not necessary to run k-e before using any other turbulence model.

sat · March 21, 2011, 21:53

thank you sir
actually i am doing project in xperimental and numerical verification of a commercial car body(i ve taken honda accord)scale 1:14 for this i ve to find cd value.so i asked help in choosing turbulence model

cfd_newbie · March 22, 2011, 00:13

Quote:

Originally Posted by sat

thank you sir
actually i am doing project in xperimental and numerical verification of a commercial car body(i ve taken honda accord)scale 1:14 for this i ve to find cd value.so i asked help in choosing turbulence model

hi Satish,
Take a look at http://www.cfd-online.com/Wiki/Ahmed_body
which is a good case for validation of vehicle aerodynamics. Read some papers related to it. What mach number are you targeting ?
My guess is SST turbulence model may suit your requirements.

sat · March 22, 2011, 04:22

its not mach no. i am targetting cd value now its coming around 0.22 but i have to bring it to 0.3 so that it will be near to actual

Martin Hegedus · March 22, 2011, 11:06

Not sure if I understand this correctly.

Do you have experimental results for a 1:14 scale car? It sounds like this is the case. I assume Cd=0.3 comes from this experiment. Can you show us a picture or two of your experimental setup?

Do you have experimental results for a full scale car?

Are you running CFD for the scaled and full scale car? I assume Cd=0.22 is for the scaled down car. I also assume you are matching the Reynolds number for the scaled down car. Can you show us a picture of the contours of V/Vinf for your CFD?

Is your ground plane moving (or not) for both the experiment and numerical results? I'm assuming that the car is placed on the ground. In this case the ground plane for the experiment MUST be moving at the same speed as the on coming wind. Otherwise, there will be a boundary layer from the floor. If there is a B.L. from the floor, you will need to model that in your CFD run. Not easy, unless you know the profile.

What is your Reynolds number based on the length of the car?

I'm not very knowledgeable in car aerodynamics, but I assume that a majority of your drag is coming from base drag. (This is something you can check with your CFD run) Base drag includes your back window, if separation occurs over it. At these lower Reynolds numbers, Cd could be sensitive to Reynolds number. In the experiment did you trip the boundary layer at the leading edge of the car? If the car was smooth, and the flow was laminar up to your back window, then the base drag from your experiment would be higher than the base drag for a fully turbulent CFD run.

sat · March 23, 2011, 02:06

i dont ve full car reading...
i ve only for scaled down model
it is considered as ground for both experimental and cfd

presently i ve taken only for the drag and lift values from the experiment.. pressure i need to take after making arrangemens

Martin Hegedus · March 23, 2011, 02:49

Quote:

Originally Posted by sat

it is considered as ground for both experimental and cfd

I assume this means the ground is moving for both?

What is your Reynolds number and did you trip the boundary layer in the experiment?

morteza08 · March 23, 2011, 11:12

Quote:

Originally Posted by sat

hello sir
i am sathish(student) doing project in external aerodynamics
1) can u tell me which turbulence model is best for solving external aerodynamics
2)whether it is necessary to run k-e model before running any turbulence model.if yes why?

kw-sst .. for aiorfoil modelling
I got the best results

Martin Hegedus · March 23, 2011, 12:11

OK, I measured my car, it is approximately 15 ft. It is not a Honda Accord. Assuming I'm traveling at 60 m/hr (sorry, I'm in US, so English units). That comes out to a Reynolds number of about 8.4e6. That's turbulent. However, if I'm at 1/14 scale traveling at 60 m/hr then my Re number would be 6.0e5. That is in the laminar/transition region. If the B.L. did not get tripped then a fully turbulent turbulence model will under predict the drag from the backside separated region.

sat · March 26, 2011, 14:55

it is fixed for both
reynolds i am getting 2.3xe5

Martin Hegedus · March 26, 2011, 15:00

Quote:

Originally Posted by sat

it is fixed for both
reynolds i am getting 2.3xe5

And? How did you fix it?

sat · March 26, 2011, 21:57

i built one platform which is like road and i fitted that to the test cross section

March 21, 2011, 21:15	help in external aerodynamics(which tubulence model is best)	#1
sat New Member sathish kumar Join Date: Feb 2011 Location: coimbaatore/hosur Posts: 7 Rep Power: 15	hello sir i am sathish(student) doing project in external aerodynamics 1) can u tell me which turbulence model is best for solving external aerodynamics 2)whether it is necessary to run k-e model before running any turbulence model.if yes why?

March 22, 2011, 04:22	thanku sir	#5
sat New Member sathish kumar Join Date: Feb 2011 Location: coimbaatore/hosur Posts: 7 Rep Power: 15	its not mach no. i am targetting cd value now its coming around 0.22 but i have to bring it to 0.3 so that it will be near to actual

March 26, 2011, 14:55	to martin hegedus	#11
sat New Member sathish kumar Join Date: Feb 2011 Location: coimbaatore/hosur Posts: 7 Rep Power: 15	it is fixed for both reynolds i am getting 2.3xe5

March 26, 2011, 21:57	to martin	#13
sat New Member sathish kumar Join Date: Feb 2011 Location: coimbaatore/hosur Posts: 7 Rep Power: 15	i built one platform which is like road and i fitted that to the test cross section

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
gamma-ReTheta turbulence model for predicting transitional flows	FelixL	OpenFOAM Programming & Development	123	August 30, 2022 11:50
Superlinear speedup in OpenFOAM 13	msrinath80	OpenFOAM Running, Solving & CFD	18	March 3, 2015 05:36
simulation results for k-w model and SST model	Li	CFX	7	June 29, 2007 04:19
Multiphase flow. Dispersed and free surface model	Luis	CFX	8	May 29, 2007 18:13
Kato-Launder model	sam	Main CFD Forum	13	September 21, 2006 10:15

March 21, 2011, 21:53		#3
sat New Member sathish kumar Join Date: Feb 2011 Location: coimbaatore/hosur Posts: 7 Rep Power: 15	thank you sir actually i am doing project in xperimental and numerical verification of a commercial car body(i ve taken honda accord)scale 1:14 for this i ve to find cd value.so i asked help in choosing turbulence model

March 22, 2011, 11:06		#6
Martin Hegedus Senior Member Martin Hegedus Join Date: Feb 2011 Posts: 500 Rep Power: 19	Not sure if I understand this correctly. Do you have experimental results for a 1:14 scale car? It sounds like this is the case. I assume Cd=0.3 comes from this experiment. Can you show us a picture or two of your experimental setup? Do you have experimental results for a full scale car? Are you running CFD for the scaled and full scale car? I assume Cd=0.22 is for the scaled down car. I also assume you are matching the Reynolds number for the scaled down car. Can you show us a picture of the contours of V/Vinf for your CFD? Is your ground plane moving (or not) for both the experiment and numerical results? I'm assuming that the car is placed on the ground. In this case the ground plane for the experiment MUST be moving at the same speed as the on coming wind. Otherwise, there will be a boundary layer from the floor. If there is a B.L. from the floor, you will need to model that in your CFD run. Not easy, unless you know the profile. What is your Reynolds number based on the length of the car? I'm not very knowledgeable in car aerodynamics, but I assume that a majority of your drag is coming from base drag. (This is something you can check with your CFD run) Base drag includes your back window, if separation occurs over it. At these lower Reynolds numbers, Cd could be sensitive to Reynolds number. In the experiment did you trip the boundary layer at the leading edge of the car? If the car was smooth, and the flow was laminar up to your back window, then the base drag from your experiment would be higher than the base drag for a fully turbulent CFD run.

March 23, 2011, 02:06		#7
sat New Member sathish kumar Join Date: Feb 2011 Location: coimbaatore/hosur Posts: 7 Rep Power: 15	i dont ve full car reading... i ve only for scaled down model it is considered as ground for both experimental and cfd presently i ve taken only for the drag and lift values from the experiment.. pressure i need to take after making arrangemens

March 23, 2011, 12:11		#10
Martin Hegedus Senior Member Martin Hegedus Join Date: Feb 2011 Posts: 500 Rep Power: 19	OK, I measured my car, it is approximately 15 ft. It is not a Honda Accord. Assuming I'm traveling at 60 m/hr (sorry, I'm in US, so English units). That comes out to a Reynolds number of about 8.4e6. That's turbulent. However, if I'm at 1/14 scale traveling at 60 m/hr then my Re number would be 6.0e5. That is in the laminar/transition region. If the B.L. did not get tripped then a fully turbulent turbulence model will under predict the drag from the backside separated region.