modeling exterior aero for Formula Student Car

FSAE_student · November 4, 2011, 12:41

Hello,

I am a university student who is currently trying to run CFD simulations on STAR CCM+ V6 for my university's Formula Student team. I have been learning/using STAR CCM for last month or so and although i have learned a lot about meshing, modeling, etc, i have recently realized that i have been running simulations extremely inefficiently (basically modeling an entire wing, not just half and covering it with a box that i did not need). Anyway, I was wondering whether someone could please give me some quick tips on the following topics:

1) do i have to cover an imported CAD part with some kind of geometry and extract the external volume to model the airflow? or can i just somehow simulate the flow with just the wing and no other geometry covering it?
2) How can i make the mesh finer independently for the leading and trailing edges of the wing while keeping the mesh relatively course for the endplates/wing surfaces?

Those are the two main questions i have, but if anyone has done a similar job and has some tips for a novice, that would be much appreciated.

Thank you very much for your help.

FSAE_student

willimanili · November 5, 2011, 08:13

Hi,

1) If you like to simulate the airflow around a wing you need a meshed volume for the air around the wing. So if you just have the wing itself you cant mesh a region around it. So, your way to cover it with a volume, maybe a box, cylinder etc., to extract the external volume is the right way to get a meshable air volume.

2) My suggestion is to use volumetric control. You will find more information how to use volumetric control in the user guide.

abdul099 · November 5, 2011, 16:46

Additionally to willimanili's reply:
For your second question, you might create a part in CAD, import it and use it as volumetric control. That gives you more flexibility than volume shapes, as it could be a part which for example follows a curved trailing edge. Creating the parts in CAD is easier than creating it in the CCM+ CAD modeller.

Some general hints:
When you want to simulate the behaviour of different wings on the same body (or vice versa), it might be worth to think about splitting the model. For example, create a box with the wing and a box with the whole car without the wing-box. You can mesh them independently which gives you the option not to mesh the whole model just for a minor change at the wing but just the small box with wing.

Also think about a consistent naming convention for all parts while modelling the geometry. Race car models can get pretty big (had some F1 cars with more than 120M cells

) and automated postprocessing by a macro on a cluster is a nice feature. Also you might set up some of the commonly used settings by a macro, it could save a lot of mouse clicks.

FSAE_student · November 7, 2011, 23:50

thank you for the response,

I have tried your ideas regarding volumetric control, however I believe i am doing it incorrectly because even though i am setting values for it, it still seems to use the same size throughout. I have also tried creating a shape to sorround the leading edge and then applying volumetric control just to the shape, but that also doesnt seem to be working.

Perhaps if someone would explain to me what exactly does the volumetric control do i can try and use it correctly. It seems that when i understand what a certain control does, i can better understand how to use it.

Also, anyone have a cell count that i should be aiming for, for a three element rear formula-style wing with endplate. I am cutting the wing in half and surrounding it with a box that barely covers it.

Thank you for the help,

FSAE_student

ping · November 8, 2011, 04:28

star-ccm+ has amazingly good help and tutorials on most subjects especially on meshing - can i suggest you spend some time reading then experiment and improve your model, since the help already provided above seems very sensible - eg read this section: Generating Mesh > Tailoring the Mesh > Working with Volumetric Controls

abdul099 · November 11, 2011, 14:08

Regarding volumetric controls, do what ping has suggested: The user guide is not there to increase the download size of the package...
And it's easier for you to look to the user guide than it is for me suggesting anything without knowing what you did.

Regarding the cell count:
First of all, I hope, your domain is bigger than a box which barley covers the wing. It's fine as long as your box has some interfaces to a bigger domain, e.g. for the rest of the car.
A Formula 1 rear wing with three elements and an endplate has about 25 million cells when using a trimmed mesh. Please note, this cell count is for a half wing cut at the symmetry plane.
A whole Formula 1 car half model has about 100 Million cells plus maybe some additional 10 Million cells.
So running the model with an appropriate mesh resolution could easily be too much.

November 4, 2011, 12:41	modeling exterior aero for Formula Student Car	#1
FSAE_student New Member Ido Waksman Join Date: Nov 2011 Location: USA Posts: 3 Rep Power: 15	Hello, I am a university student who is currently trying to run CFD simulations on STAR CCM+ V6 for my university's Formula Student team. I have been learning/using STAR CCM for last month or so and although i have learned a lot about meshing, modeling, etc, i have recently realized that i have been running simulations extremely inefficiently (basically modeling an entire wing, not just half and covering it with a box that i did not need). Anyway, I was wondering whether someone could please give me some quick tips on the following topics: 1) do i have to cover an imported CAD part with some kind of geometry and extract the external volume to model the airflow? or can i just somehow simulate the flow with just the wing and no other geometry covering it? 2) How can i make the mesh finer independently for the leading and trailing edges of the wing while keeping the mesh relatively course for the endplates/wing surfaces? Those are the two main questions i have, but if anyone has done a similar job and has some tips for a novice, that would be much appreciated. Thank you very much for your help. FSAE_student

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
modeling car radiator	petrile_83	OpenFOAM Running, Solving & CFD	1	January 5, 2017 05:11
CO modeling of a Car park..	Ryan	FLUENT	2	February 4, 2010 05:46
Convergence issues for modeling of flow over a car	blin	FLUENT	2	January 11, 2010 11:58
again a 2d car modeling and calculate CD	reza79	ANSYS	1	October 19, 2009 16:13

November 5, 2011, 08:13		#2
willimanili Member Join Date: Feb 2011 Location: DE-PB Posts: 56 Rep Power: 15	Hi, 1) If you like to simulate the airflow around a wing you need a meshed volume for the air around the wing. So if you just have the wing itself you cant mesh a region around it. So, your way to cover it with a volume, maybe a box, cylinder etc., to extract the external volume is the right way to get a meshable air volume. 2) My suggestion is to use volumetric control. You will find more information how to use volumetric control in the user guide.

November 5, 2011, 16:46		#3
abdul099 Senior Member Join Date: Oct 2009 Location: Germany Posts: 636 Rep Power: 22	Additionally to willimanili's reply: For your second question, you might create a part in CAD, import it and use it as volumetric control. That gives you more flexibility than volume shapes, as it could be a part which for example follows a curved trailing edge. Creating the parts in CAD is easier than creating it in the CCM+ CAD modeller. Some general hints: When you want to simulate the behaviour of different wings on the same body (or vice versa), it might be worth to think about splitting the model. For example, create a box with the wing and a box with the whole car without the wing-box. You can mesh them independently which gives you the option not to mesh the whole model just for a minor change at the wing but just the small box with wing. Also think about a consistent naming convention for all parts while modelling the geometry. Race car models can get pretty big (had some F1 cars with more than 120M cells ) and automated postprocessing by a macro on a cluster is a nice feature. Also you might set up some of the commonly used settings by a macro, it could save a lot of mouse clicks.

November 7, 2011, 23:50		#4
FSAE_student New Member Ido Waksman Join Date: Nov 2011 Location: USA Posts: 3 Rep Power: 15	thank you for the response, I have tried your ideas regarding volumetric control, however I believe i am doing it incorrectly because even though i am setting values for it, it still seems to use the same size throughout. I have also tried creating a shape to sorround the leading edge and then applying volumetric control just to the shape, but that also doesnt seem to be working. Perhaps if someone would explain to me what exactly does the volumetric control do i can try and use it correctly. It seems that when i understand what a certain control does, i can better understand how to use it. Also, anyone have a cell count that i should be aiming for, for a three element rear formula-style wing with endplate. I am cutting the wing in half and surrounding it with a box that barely covers it. Thank you for the help, FSAE_student

November 8, 2011, 04:28		#5
ping Senior Member Ping Join Date: Mar 2009 Posts: 556 Rep Power: 20	star-ccm+ has amazingly good help and tutorials on most subjects especially on meshing - can i suggest you spend some time reading then experiment and improve your model, since the help already provided above seems very sensible - eg read this section: Generating Mesh > Tailoring the Mesh > Working with Volumetric Controls

November 11, 2011, 14:08		#6
abdul099 Senior Member Join Date: Oct 2009 Location: Germany Posts: 636 Rep Power: 22	Regarding volumetric controls, do what ping has suggested: The user guide is not there to increase the download size of the package... And it's easier for you to look to the user guide than it is for me suggesting anything without knowing what you did. Regarding the cell count: First of all, I hope, your domain is bigger than a box which barley covers the wing. It's fine as long as your box has some interfaces to a bigger domain, e.g. for the rest of the car. A Formula 1 rear wing with three elements and an endplate has about 25 million cells when using a trimmed mesh. Please note, this cell count is for a half wing cut at the symmetry plane. A whole Formula 1 car half model has about 100 Million cells plus maybe some additional 10 Million cells. So running the model with an appropriate mesh resolution could easily be too much.