CFD - Computational Fluid Dynamics, Motorsport, Formula 1

[julien.decharentenay]

Hi,

I just joined the forum as the software I am developing, Khamsin (a Google SketchUp plugin for CFD Modelling), has been mentioned a couple of times earlier in the year. If anyone is interested in having a go at it and would like some help, feel free to contact me at info@hibouscientificsoftware.com.au. Be aware that one of the limit of the current version is that it does not allow for the modelling of the rotation of the wheels (which limits its useful for F1 whole car CFD modelling purposes).

I would love if someone would be willing to send me a geometry (or point me on a decent model in the SketchUp warehouse) so I can prepare a video tutorial.

Thanks in advance.
Julien de Charentenay

[niazaliahmed]

Hi I am a 2.5 years experience CFD Engineer. Have been doing CFD for optimizing race car aerodynamics for a few privateer race drivers. Would be delighted to help privateer racers who want some free CFD Simulations done.

[astracrazy]

anybody here able to give me a hand with autodesk invertor and autodesk simulation cfd?

I'm trying to use my sketchup model, i can import it into inventor but it won't let me set-up the fluid volume on it to then export to cfd (once the volume is set you can automatically do this). Anybody got any ideas where i'm going wrong? Please pm me

[astracrazy]

does anyone here now how you can simulate exhaust air flow in solidworks? Thanks in advance

[flyboy2160]

does anyone here now how you can simulate exhaust air flow in solidworks? Thanks in advance

you can simulate a steady state* - not pulsed - flow by picking the exhaust diameter face and inserting a boundary condition. if the exhaust face is normal to the pipe axis, just leave the default flow direction as the face normal. set the boundary condition as an INPUT velocity (or INPUT mass flow if you prefer). this seems strange, but you are introducing a flow INTO the computational domain.

similarly, you can remove a flow from the field, such as going into an internal duct, by setting the boundary condition of the face to OUTPUT. this last trick allows you to examine the detailed flow going through a radiator or something internal as a separate case from the external flow. after getting the answer for the radiator duct flow, you can input the radiator exhaust flow back into the overall model. of course, if you have a supercomputer, you could just do it all in one model......

(*there is a way to make time dependent flows, but i've never used it.)

[astracrazy]

so basically the face of the exhaust exit, i can set as a boundary condition which needs to be an input velocity? Can you set the temperature?

Thanks

[flyboy2160]

so basically the face of the exhaust exit, i can set as a boundary condition which needs to be an input velocity? Can you set the temperature?

Thanks

give me a few minutes. i'll start up one of the models in which i've done this and post back.

[flyboy2160]

so basically the face of the exhaust exit, i can set as a boundary condition which needs to be an input velocity? Can you set the temperature?

Thanks

yes, you can set the properties of the input and output streams to be different from those of the nominal domain flow.

[flyboy2160]

anybody here able to give me a hand with autodesk invertor and autodesk simulation cfd?

I'm trying to use my sketchup model, i can import it into inventor but it won't let me set-up the fluid volume on it to then export to cfd (once the volume is set you can automatically do this). Anybody got any ideas where i'm going wrong? Please pm me

i have this also, but it's a few years old. i gave up trying to use it because it demanded a solid cfd domain that had the studied shape hollowed out from it and had meshing errors in which the cfd mesh severely encroached into the "hollow." the new version with adaptive meshing may be better.

[astracrazy]

thanks for your help. Can have a little play with it when i get home

i gave up with Autodesk, Solidworks is much more user friendly (for me anyway)

[Th0R]

I just wanted to mention that neither autodesk nor solidWorks should be used for simulation exhaust gas flow. These are both construction tools, which have the limited abillity to calculate flows. But the results are far from accurate. Even with special cfd software exhaust gas flow will be hard to predict. I dont think that this topic is something anyone who wants to do cfd as a hobby should work with.
You should start simulating smaller things, like wing proflies or something like this. After that try to interpret the results of the simulations (for example compare them with results found in literature). When i try out a software i simply start to simulate the airflow over a plane. This is a situation were you can actually calculate how well your simulation program works. For more complex geometries you will need meassurments to back up and adjust your simulations. Without that you will only produce coloured images without meaning.

[slimjim8201]

I just wanted to mention that neither autodesk nor solidWorks should be used for simulation exhaust gas flow. These are both construction tools, which have the limited abillity to calculate flows. But the results are far from accurate.

This is a pretty bold statement without any supportive reasoning. Please, feel free to enlighten the forum with facts and figures. Any pertinent information you have regarding results inaccuracy would be very welcome.

[Th0R]

ok i will try to, as i dont have any cfd software at home and at work i can only use it for things which i cannot publish here i will simply line out some general things. I only have experience with the cfd-functionallity of solidworks 2010 professional edition and compared to that the actual starccm and cfx versions. I actually do not have any knowledge about autodesk cfd functionallity but i think like solidworks it will not match the capacities of a specallised cfd software. Perhaps someone else can elaborate on that.

The first and obvious reason why cfd software has to be better, is that it is still in use. If cad programs would be as good in simulating cfd as special cfd tools, nobody would buy and use cfd software, because you need to buy cad software anyway

Ok now straight forward to some facts:

What do you need when using cfd tools.

1. I want good control about the generated mesh, i want to manually improve / mesh when needed. I want different mesh types to minimize nummerical errors in boundary layers or arround shock waves. In solid works these mesh options were very limited.

2. I want to control the types of boundarys. I dont want to simply chose a type of boundary and look what happens. Maybe i have some meassurement results (for example heat distribution at an outlet.) I want to be able to map these things to the boundarys. Simply taking mean vaules will not work if you want accurate results. Here again solidworks lacks options.

3. I want to be able to automate repeative steps in my work process. The cfd tools i mentiond earlier have great macro techniques for that, solidworks hasn't.

4. Solving process. Options to choose different solvers, preconditioners, maybe even add another existing solver....

5. Results.
I need to be able check the numerical errors for each cell seperatly. If i cant do this i cant say that the heat/perssure velocity etc is correct at this point. Again solidworks lacks options here.


Last but not least i dont want to bash solidworks or autodesk, they both are great cad tools. I would also give it a try to simulate simple flows with them. But when i want to simulate exhaust flow, i wouldn't even trust the specalised cfd software.

[Jersey Tom]

I just wanted to mention that neither autodesk nor solidWorks should be used for simulation exhaust gas flow.

I actually do not have any knowledge about autodesk cfd functionallity...

[Smokes]

Thor you need solidworks flowworks prefably 2012 version( 2010 and 2011 didn't seem to stable)its an add on package and you can simulate injection of gas streams into a flow.
but i would say flow works is bias towards the eletronics industry and does not have transient solver for it that works well.
Open foam and ansys will do pretty good predictions but you have to know what you are simulating and how you mesh the grid.but like any simultion If you input random boundarys conditions you will get a random answer.