Khamsin Virtual Racecar Challenge 2015

[CAEdevice]

A question about the engine inlets: is there any inner template?

[cdsavage]

A question about the engine inlets: is there any inner template?

No inner template for the engine inlet, this was omitted for simplicity. There has been an extra point added to K4.3 stating that there must be a plausible volume for engine inlet air to reach the engine, like at the end of K4.2.

[MadMatt]

Hi Julien,

I was using Paraview 4.1, so after unistalling it, I installed Paraview 3.98.
In that case the simulation stoppes after a few seconds: "pvbatch can not be found"

Where is ParaView 3.98 installed? In my case it is installed in (and it is found automatically):

Code: Select all

C:\Program Files (x86)\ParaView 3.98.0

I still have the "pvbatch can not be found" error that I mentioned previously, even by installing Paraview 3.98 in the x64 folder (I have the latest Paraview version in the x64 folder). I would like to try that CFD package to see how different the results are with the software I am using so I can calibrate things a bit.

[julien.decharentenay]

@MadMatt Can you let me know which directory ParaView is installed and check that there is an executable called pvbatch.exe in the bin directory? I am slightly surprised as the software should only check whether the file exists...

Calendar: Chris explained the logic behind the choice of track type. We think that the low downforce tracks are less interesting as they focus mostly on reducing the number of downforce inducing devices to reduce drag as much as possible. This is a case where less is probably more. In comparison, the high and mid downforce tracks are more challenging design-wise.

There are a number of reasons behind the two non-scoring rounds:
- Trying to foster opportunity for innovations - as these tracks are non-scoring, it is an opportunity to try to push the envelope further;
- Giving an opportunity to benchmark the design prior to the scoring rounds;
- July/August period is a holiday period in the northern hemisphere. It would be unfair to have a scoring round during this period;
- Based on last year season, the top positions remain very similar from round to round. More scoring rounds does seem to result in a more competitive environment.

At the end of the day, I would be happy either way.

Mesh question: There has been some discussion on computational resources, costs and trying to even out the playing field. It is inevitable that there will be some difference between people opprtunities (software, time, resources, experience). At this stage, the mesh has not yet been finalised and I wanted to put the following question back to you as there are two options:

Option 1: Keep the mesh parameters fairly similar to last year. When the symmetry is used, the mesh size is around the 1.2 to 1.6 millions cells depending on the exact meshing parameters. Overall, a simulation may be undertaken in around 2 hours on a dual-core computer (or at least, this is the turn-around time I get from Amazon AWS m3.large at a cost of around US$0.14 per simulation).

Option 2: Increase the mesh resolution slightly to have an overall mesh size of around 2.5 to 3 millions cells to have a simulation turn-around time of 10-12 hours on Amazon AWS (m3.large instance) at a cost of around ~US$1.

[julien.decharentenay]

I obtained a different problem during running: https://dl.dropboxusercontent.com/u/522 ... rgence.txt

The STL is the same of the previous test (the one completed without refinement).

I am glad that the divergence trapping does work. I am working on a very similar issue at the moment - which is preventing me from releasing the next version. I hope to be able to get past it as it has been bugging me for a few days now (and for some reason the meshing is not working either)... Can you confirm whether you are using inlet and exhaust - at this stage I would recommend not using these?

Two things you might be able to check:

(1) dimensions of the STL, which should show up in the log file. These should be in meters.
(2) ensure that the STL mesh does not have holes. I use netfabb Basic to check the mesh and fix issues. It does help in quite a few cases.

[CAEdevice]

I obtained a different problem during running: https://dl.dropboxusercontent.com/u/522 ... rgence.txt

The STL is the same of the previous test (the one completed without refinement).

I am glad that the divergence trapping does work. I am working on a very similar issue at the moment - which is preventing me from releasing the next version. I hope to be able to get past it as it has been bugging me for a few days now (and for some reason the meshing is not working either)... Can you confirm whether you are using inlet and exhaust - at this stage I would recommend not using these?

Two things you might be able to check:

(1) dimensions of the STL, which should show up in the log file. These should be in meters.
(2) ensure that the STL mesh does not have holes. I use netfabb Basic to check the mesh and fix issues. It does help in quite a few cases.

Hi, I confirm that I'm not using exhaust and that units are meters.
About the mesh: I checked it before starting the simulation and it has no holes (with SpaceClaim that has some STL dedicated tools)

EDIT:
New run (with a 2015 draft)
https://dl.dropboxusercontent.com/u/522 ... D_test.log
https://dl.dropboxusercontent.com/u/52296215/test.txt

I still have convergence problem.

[CAEdevice]

I'm running a new simulation with a different geometry file (new structure tree, I have the idea that something goes wrong with the names (given to bodies/surfaces or to assembly components?)

I have to questions about the PPT document.
1) The distance between the ground and the reference surface is 40mm?
2) Is there a distance along the y direction (ex. from the origin to the FW axis)?

The PPT images about CAD orientation are confused about the Y direction (+ or - ?) and the part names in the image are not the same or the previous text explanation.

[MadMatt]

A question about the engine inlets: is there any inner template?

No inner template for the engine inlet, this was omitted for simplicity. There has been an extra point added to K4.3 stating that there must be a plausible volume for engine inlet air to reach the engine, like at the end of K4.2.

Can the plausible path for this engine inlet go through the cockpit template? What I am seeing here is that the engine inlet on top of the cockpit can be located quite far up front (making it sit over the cockpit template). Therefore is this accepted?

@MadMatt Can you let me know which directory ParaView is installed and check that there is an executable called pvbatch.exe in the bin directory? I am slightly surprised as the software should only check whether the file exists...

4.3 is installed in Program Files x64 (aka Program Files), and 3.98.1 is installed in Program Files x86, and both have the pvbatch.exe in the bin directory.

[cdsavage]

A question about the engine inlets: is there any inner template?

No inner template for the engine inlet, this was omitted for simplicity. There has been an extra point added to K4.3 stating that there must be a plausible volume for engine inlet air to reach the engine, like at the end of K4.2.

Can the plausible path for this engine inlet go through the cockpit template? What I am seeing here is that the engine inlet on top of the cockpit can be located quite far up front (making it sit over the cockpit template). Therefore is this accepted?

No, the 'plausible path' must be inside the bodywork but excluding the space taken up by the cockpit template.

EDIT: I think it would make things less ambiguous if we moved the forward limit on the engine inlet rearwards by a small amount, this way it's clear that if the engine inlet is on top of the cockpit, it must sit above the cockpit template, not in front of it. I intend to change the 855mm dimension to 875mm or so in the final release.

[cdsavage]

I'm running a new simulation with a different geometry file (new structure tree, I have the idea that something goes wrong with the names (given to bodies/surfaces or to assembly components?)

I have to questions about the PPT document.
1) The distance between the ground and the reference surface is 40mm?
2) Is there a distance along the y direction (ex. from the origin to the FW axis)?

The PPT images about CAD orientation are confused about the Y direction (+ or - ?) and the part names in the image are not the same or the previous text explanation.

Originally One-Click CFD required the ground plane at z = 0 but I believe this is being changed.

Other than that, if built to the guide files, the car should be in the correct position/scale. Front axle at y = 0, with the rear of the car being +y. After the change to One-Click CFD, the reference plane should be at z = 0, with the ground plane, from memory, at z = -45 but I will check with Julien.

If you are needing to make other changes to position / scale before using One-Click CFD, let me know so I can alter the guide files.

[CAEdevice]

Thanks for the explanation. I have the same doubt about 40mm or 45mm distance.
I run the simulation, but I'm still having a divergence issue.

[RicME85]

Think I am not going to bother with this year.
Have sat down and started working on a car about 10 times in the last month and so far have a blank file.
Just not getting my head around things and not enjoying it.
I thought LMP style regs would be interesting buy its just not working for me.

[cdsavage]

Think I am not going to bother with this year.
Have sat down and started working on a car about 10 times in the last month and so far have a blank file.
Just not getting my head around things and not enjoying it.
I thought LMP style regs would be interesting buy its just not working for me.

Sorry to hear that. Have you had a look at the starter parts? Are there any other pre-built parts in particular that we could supply to make it easier to get started? Any rules that you're finding particularly annoying/difficult?

[RicME85]

Yeah I've looked at the starter parts, plonked the wheels and suspension into the file once and started looking at the floor but again just stopped.

I use a lot of morphine (and have done for a lot of years) so find concentration and comprehension difficult at times.
I have a big interest in F1 which I think helped me get my head around the regs previously but with LMP I have watched probably a combined total of 3hrs in the years I have watched F1 (since 92).

I dunno, something might suddenly click one day and I will produce a car but I cannot say for certain.

[CAEdevice]

I had difficulties with LMP at first because I have been more interested in F1 too (since 1985!), but after reading everything I could find about C group, Imsa, LMP I could appreciate closed wheels.

When I began to model my 2015 car I didn't know what to model first: the starting part provided by Chris and some pictures I found on the web helped very much.

I can send you a basic release of a LMP car I'm using to compare my cfd solver to openfoam: it is not optimized and it is quite conventional, but it can help you to concentrate about details and development.

Wich CAD are you using? I have native SWX format but I can export to STEP or STL or SktechUp.