CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Flow around a body (Automotive) (

Roland July 16, 2006 17:33

Flow around a body (Automotive)
Hi all

Im having difficulty finding out how to determine a couple of core force values in a car/windtunnel simulation.

The first value is downforce balance front to rear. That is, if I add a rear wing to increase downforce, I can get an overall value of lift/downforce easily, but I cannot determine the balance of forces front to rear. Is there a way to divide the chassis at its (known) center of gravity (i.e. via a plane), and then calculate vertical force for each section (front vs. rear) in Post?

Closely related, I wanted to determine the aerodynamic CG of the test vehicle, but again there does not seem to be a simple function for this.

Any insight is appreciated-


Joe July 17, 2006 05:07

Re: Flow around a body (Automotive)
Quick and dirty method: split the patches covering the vehicle verticaly at the CoG. You can then apply the relevant functions to the back-half-of-the-car etc.

Robin July 17, 2006 10:38

Re: Flow around a body (Automotive)
I assume that for the aerodynamic CG you are looking for the center of pressure (i.e. point about which there are no moments). There are no functions in Post to calculate this directly, but it is relatively straigtforward to get this. Simply calculate the components of the force and torques (about the global X,Y,Z axes) on your model, then apply simple linear statics to find the point about which the torques go to zero.

Regards, Robin

Roland July 17, 2006 13:47

Re: Flow around a body (Automotive)

Thanks very much for the thoughts. I had assumed that the center of pressure would require some external calculation.

The patching approach might be burdensome if the CG changes in the model (i.e. overall balance with aero still requires a shift of mass in some iterative process). Is there a method to define patches as 'areas of interest' in Post, vs defining these during mesh creation?

Robin, the idea of using torques seems interesting here. If I know the CG, and I know the total downforce (or lift), couldnt I use the torque function to infer the balance of downforce forward and aft of the (CG) reference plane?

Best, Roland

Robin July 18, 2006 08:10

Re: Flow around a body (Automotive)
Hi Roland,

It's simple linear statics. You could write the whole thing out in expressions to get three expressions for the center of pressure. First year engineering stuff.

Currently there is no way to select individual cells in Post, but you can create patches by generating a contour plot with appropriate contour levels, then creating a User Surface object with the "from Contour" option, which allows you to create a user surface at a particular contour level.

Regards, Robin

Roland July 18, 2006 17:22

Re: Flow around a body (Automotive)

Thanks as always for the creative approaches to these problems.

A question on torque- In a non-rotating domain, how is the torque axis location inferred (i.e. pitch axis for an auto chassis) by CFX? Can this axis be changed with a coordinate frame, etc?

Thanks again- Roland

Bak_Flow July 19, 2006 07:52

Re: Flow around a body (Automotive)
Hi Roland,

creative? Yes but somtimes I wonder where the terms come from ...??? LOL

Robin mentioned this comes from a simple first year course in Linear Statics?

The procedure of determining force and moment resultants from distributed forces is in the field traditionally known as mechanics of RIGID bodies.

Linear Statics is the field of mechanics of DEFORMABLE bodies where in general behaviour may be linear or non-linear. This is beyond what you require here.

Just a mix up of terms probably.



Robin July 19, 2006 08:42

Re: Flow around a body (Automotive)
Hi Bak_Flow,

Do you actually have a point, other than demonstrating how gosh-darn smart ye are?

True, Linear Statics is the field of deformable bodies, but you may also recall that it introduced the calculation of force and moment balances (of course, this may differ from one University to the next). My point was that the methods introduced in such a first year course can be applied here, not that you have to do the analysis of deformable bodies.

Regards, Robin

opaque July 19, 2006 09:18

Re: Statics/Linear Statics?
Hi Bak_Flow

Not to defend Robin in his use of the term Linear Statics, but checking just a few of the basic books on mechanics the chapter is usually called Statics (w/o Linear).. Sure it is explained first for RIGID bodies, but it is not limited to it. The concept of equilibrium (from Newton) does not mention anything about rigid/deformable bodies or linear/non-linear constitutive relations.

People go to different schools, have different background, speak different languages. While reading information on an international forum such as cfd-online, we must try to put it into context the information and confirm what we understood (two-way communication)



PS. I did not like the first year remark either.. That depends which school you go/went to.

Roland July 19, 2006 16:37

Re: Flow around a body (Automotive)
Im new to CFX, and I have learned a great deal about methods to leverage native functionality in non-obvious ways from Robin's (and others') posts. If I were an expert in this, I would not need to ask for guidance.

My followup question was not about CP, but on the primary question on front/rear downforce calculation ('if' it would be simpler to calculate downforce balance front/rear by using torque on the pitch axis combined with the overall lift/downforce value from the function calculator); this seems easier than defining patches and calculating forces separately. Effectively I needed/need to know- how is the torque axis in a non-rotating domain defined? I have not found this information in the literature or help files. Is the pitch axis simply defined as mid-point of the body (i.e. nose to tail on an auto chassis), or by some coordinate frame, etc?

Thanks in advance to anyone who can help clarify- Regards, Roland

Bak_Flow July 19, 2006 19:43

Re: Statics/Linear Statics?
Hello Opaque,

I like your comments.

This was exactly my point that if Roland were to look up this subject in a textbook, (presumably because one would like a richer and more fundamental understanding) they would not likely find it under the topic of linear statics.

I certainly realize that there are a broad range of terms around the world and over time for such studies in mechanics. However, the term that linear statics would not be the choice search words...I infact Googled these and found only info on linear deformable body analysis.

As for Robin's silly comments about my intent, IQ, etc...I have no point is probably clear to everyone else and that that will suffice.



PS first year of what? PhD, grade school, grad school....10 year jail term? LOL

Robin July 20, 2006 07:44

Re: Flow around a body (Automotive)
Hi Roland,

The torque function returns the torque about a specified global or local coordinate system axis. It doesn't actually have any specific relationship to rotating frames, other than it being a quantity of interest to turbomachinery design.

So, to answer you question more directly. Yes, you can specify a local coordinate system about which to calculate torque. You should be able to find further documentation on this in the on-line CFX help.

Best regards, Robin

Robin July 20, 2006 09:13

Re: Statics/Linear Statics?
Hi Bak_Flow,

I apologize for berating you. I took your comments personally and should not attack your for this.

Regards, Robin

All times are GMT -4. The time now is 00:51.