Software Aerodynamic Simulation software

[sitha241]

Any sugestions on any free air flow software that allows inporting 3D objects and testing air resistance on difirent conditions?

 

[Urwumpe]

Any sugestions on any free air flow software that allows inporting 3D objects and testing air resistance on difirent conditions?

OpenFOAM is the only really good that I know so far. You can combine it with Paraview.

 

[sitha241]

Airflow

cant find win 7 32 bit instaler if openfoam,and I am looking for a 3 d model of an old car,so I can test its aerodynamics cause I am preping it for race on airstrip car is renault 4 but I canot find a free 3 model and I am not skilled at making one :/ anyone help?

 

[Urwumpe]

cant find win 7 32 bit instaler if openfoam,and I am looking for a 3 d model of an old car,so I can test its aerodynamics cause I am preping it for race on airstrip car is renault 4 but I canot find a free 3 model and I am not skilled at making one :/ anyone help?

First of all, you better don't use Windows 7 for OpenFOAM (but there are binaries for an older version of OpenFOAM around, if you just use Google properly), but a Linux box. And then not one, better use a few good Linux computers with lots of RAM and good CPU - and a good HDD for storing the results, which can be pretty large.


I don't know what you expect there, so let me tell you a bit of my job, which is writing software for the automated analysis of CFD simulation results. The resulting data of a simulation run, that focuses mostly on the inside of the engine compartment, is a few hundred Gigabytes large, and takes a few hours on the 9000 core cluster of a small local car company here. (Still listed among the TOP 500)

you can use simpler geometry for less accurate results, but faster calculations.

Another aspect there: There are special people around, like some coworkers of me, who do nothing but creating and optimizing the geometries for such simulation runs. I could point you to the management of the engineering services department of my company, but I doubt you can even afford a friendly price for the services.

So, in brief: Unless you learn to do it yourself and get good in it, you will have problems finding help there, because those who can do the job exist, have a well running market and will be expensive to hire for you.

 

[sitha241]

airflow

Ok I got the point beside that I am studying physics so I will try to run some caculations on my own..I just thinked that It was'n to be so hard to find simple softweare and sice there is a vast amount of data at output because I have computer that runs on coal I give up on that..Il try to do it myself tnx anyway

 

[Urwumpe]

Ok I got the point beside that I am studying physics so I will try to run some caculations on my own..I just thinked that It was'n to be so hard to find simple softweare and sice there is a vast amount of data at output because I have computer that runs on coal I give up on that..Il try to do it myself tnx anyway

Partial differential equations are only simple for a single Voxel. :lol:

Try getting an old OpenFOAM version for windows into your hands and play with it a bit, the tutorials should run on a moderate PC already, if you give it time.

 

[boogabooga]

cant find win 7 32 bit instaler if openfoam,and I am looking for a 3 d model of an old car,so I can test its aerodynamics cause I am preping it for race on airstrip car is renault 4 but I canot find a free 3 model and I am not skilled at making one :/ anyone help?

O.K. Time out.

We are talking about computational fluid dynamics here, right?
Keep in mind that CFD is done in "negative space" so to speak. You build a grid around the geometry that you want to model, then compute the external flowfield. Don't expect a 3D computer model like what you see in Orbiter to be useful as it is, unless you want to model the flow through a hollow car...

But, you could use the model as a template to construct a computational grid around it. However, CFD grids are built differently than 3D models, and constructing this grid will probably require special software.

Anyway, CFD takes some experience and knowledge to get anything useful out of it. This is a topic taught at the graduate level or upper-class undergraduate level in universities. There are issues associated with numerical stability, turbulence modeling, etc. Don't expect to take a 3D model and something you download off of the internet and do drag optimization study in one day. Also, it helps if you actually understand fluid mechanics first! If you understand the mathematics of partial differential equations, how to discretize them, and how to maintain numerical stability while solving their discretized equations, well that is even better.

 

[RisingFury]

Ok I got the point beside that I am studying physics so I will try to run some caculations on my own..

Your optimism makes me think you're in the early stages of studying Physics. As you "grow up" (A professor's joke at my university), you'll figure out that you'll solve every megalithic scenario by the time you're done you graduate :lol:

 

[Arrowstar]

But, you could use the model as a template to construct a computational grid around it. However, CFD grids are built differently than 3D models, and constructing this grid will probably require special software.

Oh, it will most certainly will. Building a mesh for a useful CFD analysis is the most complicated part of the job in my experience. The mesh defines the results, the stability of the calculations, everything. I've taken upper level coursework in CFD and I still struggle to put together all but simple meshes.

If you want my advice, avoid CFD at all costs until you can learn it from an experienced professional (such as a professor) correctly. Stick to the basic 1-D equations of fluid mechanics for now and save yourself the hours of headache that will be your (likely failed) attempt at using CFD.

 

[n72.75]

There's a reason ANSYS Fluent isn't free.

 

[boogabooga]

There's a reason ANSYS Fluent isn't free.

Or cheap.

 

[sitha241]

Your optimism makes me think you're in the early stages of studying Physics. As you "grow up" (A professor's joke at my university), you'll figure out that you'll solve every megalithic scenario by the time you're done you graduate :lol:


you are correct,I am second year on college..I'l try to work something out in 1 D..tnx everyone for advices..

 

[Juanelm]

If you want to model low-speed (subsonic) aerodynamics of an aircraft, there are grid-free CFD codes that are way simpler to use than traditional grid-based CFD. For example look at XFLR5 open source code:

http://www.xflr5.com/xflr5.htm

Or the AVL code,

http://web.mit.edu/drela/Public/web/avl/

However, if you want to model supersonic or hypersonic flow (like re-entry aerodynamics), you need grid-based, hard to use properly, expensive, extremely time-intensive CFD codes.

 

[T.Neo]

This entire thread is the motivation for me to stick a disclaimer on all my addons that says "If the aerodynamics of this addon are unrealistic, it is because I haven't got a clue what the aerodynamics actually are"... :uhh:

 

[Urwumpe]

This entire thread is the motivation for me to stick a disclaimer on all my addons that says "If the aerodynamics of this addon are unrealistic, it is because I haven't got a clue what the aerodynamics actually are"... :uhh:

I write "Pretuned by German Engineering" on any add-on, that flies too good to be true. :lol:

 

[n72.75]

It might be fun to use OpenFOAM to model turbulent stall characteristics of the Deltagilder. They could be implemented with a look-up table like SSU.

Next time I have a few weeks free ...

 

[Thunder Chicken]

I don't know what you expect there, so let me tell you a bit of my job, which is writing software for the automated analysis of CFD simulation results. The resulting data of a simulation run, that focuses mostly on the inside of the engine compartment, is a few hundred Gigabytes large, and takes a few hours on the 9000 core cluster of a small local car company here. (Still listed among the TOP 500)

you can use simpler geometry for less accurate results, but faster calculations.

Another aspect there: There are special people around, like some coworkers of me, who do nothing but creating and optimizing the geometries for such simulation runs. I could point you to the management of the engineering services department of my company, but I doubt you can even afford a friendly price for the services.

So, in brief: Unless you learn to do it yourself and get good in it, you will have problems finding help there, because those who can do the job exist, have a well running market and will be expensive to hire for you.

Ditto what Urwumpe said. I have a PhD in mechanical engineering with a concentration in thermal-fluids and I worked for many years doing CFD modeling and simulation. It is expensive in resources. It is not currently something that someone can simply download, run, and be guaranteed a reasonable answer. There are engineers at every step of the process, from the solid modeling, to the meshing, to the selection of appropriate mathematical models and numerical routines, to the evaluation of the validity of the final results. You need a good understanding of fluid dynamics, thermodynamics, and numerical methods.

 

[Jerryphan]

It will be cheaper than just k