[KVRC] Khamsin Virtual Racecar Challenge 2016

[cdsavage]

Is the problem arising due to the geometry we submitted or is it a flaw in the cfd simulation?

It's hard to say.

[CAEdevice]

I have never had similar issues this year and I have run more than 25 simulations since January. All the issues I had with occfd are related to parallel computing or to the use of the external tools. I don't want to criticize the design of other teams (consider my point of view it a suggestion), but I see from the pictures that the geometry quality is sometimes very low.

My job often consists in fem/cfd geometry preparation and I think that the ideal starting cad model would be a watertight (=solid) STEP.

[Alonso Fan]

Well we have quite a few teams being affected by the issue. I know that my submission wasn't water tight. If the issue persists, perhaps we could, just this once, have a reschedule of the first race, and push the other races back by a few days. That way it will give everyone a chance to sort out their car issues and hopefully the issues will be sorted. The only problem is that all your efforts with the current simulations will be wasted, and that's not good

[cdsavage]

I have never had similar issues this year and I have run more than 25 simulations since January. All the issues I had with occfd are related to parallel computing or to the use of the external tools. I don't want to criticize the design of other teams (consider my point of view it a suggestion), but I see from the pictures that the geometry quality is sometimes very low.

My job often consists in fem/cfd geometry preparation and I think that the ideal starting cad model would be a watertight (=solid) STEP.

To some extent I agree, but it's also happened with the Mantium and TalnoRacing entries, both of which look pretty clean. The TalnoRacing entry appears to come from a solid modelling package.

[variante]

My CFD runs (with the latest version of OCCFD) did show that "very high/very low" behaviour on the cooling inlets (inlets only). On the other hand, the official results I received did not show that problem at all.

My "private" test -run on the car that I submitted- revealed a good 60Pa.m2, while the official results gave me an extremely high 110Pa.m2

[TalnoRacing]

The TalnoRacing entry appears to come from a solid modelling package.

Yes I use Solidworks 2014 to draw my car as 1 solid body.

[HP-Racing]

I have never had similar issues this year and I have run more than 25 simulations since January. All the issues I had with occfd are related to parallel computing or to the use of the external tools. I don't want to criticize the design of other teams (consider my point of view it a suggestion), but I see from the pictures that the geometry quality is sometimes very low.

My job often consists in fem/cfd geometry preparation and I think that the ideal starting cad model would be a watertight (=solid) STEP.

As far as i know, snappyHexMesh only accepts .stl or .obj files (someone correct me if i'm wrong). The KVRC team would have to convert all of the entries back to .stl if we submitted STEP files.

[CAEdevice]

I have never had similar issues this year and I have run more than 25 simulations since January. All the issues I had with occfd are related to parallel computing or to the use of the external tools. I don't want to criticize the design of other teams (consider my point of view it a suggestion), but I see from the pictures that the geometry quality is sometimes very low.

My job often consists in fem/cfd geometry preparation and I think that the ideal starting cad model would be a watertight (=solid) STEP.

As far as i know, snappyHexMesh only accepts .stl or .obj files (someone correct me if i'm wrong). The KVRC team would have to convert all of the entries back to .stl if we submitted STEP files.

You are right.
What I mean is that no operations should be done on stl files (rule check, preprocessing, geometry repair, ...). The staff (Chris) should receive a STEP and then (very last operation) it should be converted it into stl.

The STEP format has been created for data exchange, STL was only the way the workstations used to communicate the final geometry to the first stereolitography machines (today we would say "3d printing").

[HP-Racing]

I have never had similar issues this year and I have run more than 25 simulations since January. All the issues I had with occfd are related to parallel computing or to the use of the external tools. I don't want to criticize the design of other teams (consider my point of view it a suggestion), but I see from the pictures that the geometry quality is sometimes very low.

My job often consists in fem/cfd geometry preparation and I think that the ideal starting cad model would be a watertight (=solid) STEP.

As far as i know, snappyHexMesh only accepts .stl or .obj files (someone correct me if i'm wrong). The KVRC team would have to convert all of the entries back to .stl if we submitted STEP files.

You are right, but what I mean is that no operations should be done on stl files (rule check, preprocessing, geometry repair, ...). The staff (Chris) should receive a STEP and then (very last operation) it should be converted it into stl.

The STEP format has been created for data exchange, STL was only the way the workstations used to communicate the final geometry to the first sterolitography machines (today we would say "3d printing").

But if the conversion to stl fails to produce perfect files then someone would have to repair those files too (ignoring the fact that the the surfaces for cooling, exhaust, etc... still need to be extracted which is another job which needs to be done).

Don't get me wrong, we are fortunate enough that we don't need to care which format we submit, but I personally think using .stp would make even more work for Julien and Chris.

[CAEdevice]

I don't agree. If the original geometry is good, the CAD used to convert it into stl will generate a good stl.

I obtain good stl with SWX, Creo, SolidEdge and even with FreeCAD. Chris is using NX (I think), so it will not be a problem to generate a perfect stl.

Also consider that we are using quite raw (but light) STL for the reason I wrote above: Chris needs to work on them. If the stl would be generated at the end of the process, it would not be a problem to use more refined stl.

I am not sure that the problem is caused by bad geometry, but my tests with occfd show that a good cad model helps the mesher (snappy) very much.

Finally: it is quite difficult to manage/modify an stl, dedicated tools, like MeshLab (free), make great things, but to work on the original step would make much more easy to prepare the geometry.

Ps: the STEP format include the assembly structure, so there will be no need for Chris to isolate inlets/outlets.

[RicME85]

Thing is the competition was based around Sketchup to start with because of Julien's Khamsin plugin. Sketchup means that the competition is accessible to a variety of people.
Moving to STEP would mean messing around with exporting from Sketchup in something like DXF then moving to another program to export in STEP. That is a) a pain and annoying plus b) will probably cause issues with the geometry too

[CAEdevice]

Hi Ric, I don't want to insist on this point, since it is not sure that the issue is related to geometry. What is sure is that to manage/prepare/modify a step in case of problems would much easier.

The use of Sketchup is not necessary. A very good tool would be freeCAD, but lots of commercial tools (es.Solidworks, OSD/CreoElementsDirectPersonal, Solid Edge,...) have cheap or free editions for students and for non commercial applications.
It is worth to learn a real CAD, both for professional and personal purposes.

You might say that a real CAD is more difficult to learn than SketchUp, but they are both much easier to understand than aerodynamics a (and we are talking about an aero dynamic contest ).

[graham.reeds]

Is anyone moving the foot well box up or down? Or the suspension? (K2.1)

[RicME85]

When I was doing the full class I moved the footbox and suspension to their highest positions.

[AratzH]

My actual car has footbox full up and suspension full down. It's all about flow management.