Solidworks CFD simulations

[freebeard]

Are these 2D or 3D simulation runs?

[Vekke]

AST II Cd 0,186 cut to length with 5 degree diffusor like EQXX


EQXX Cd 0,159


EQXX dip straightened Cd 0,174


If you study the pictures closely be warned they are not level in the pictures. So you need to level them in picture program. They are level in the simulations and models...

At the moment seems the EQXX beats the templates in all aspects. Next will study the EQXX shape by lowering the rear edge 10mm steps and by maintaining similar dip. First 10mm seems to go much lower in drag. If the lowering would not start to increase the drag in some point then i start to guess the program accuracy and function.

[freebeard]

If you study the pictures closely be warned they are not level in the pictures. So you need to level them in picture program. They are level in the simulations and models...

I can't see it. Where should I look? The streamlines at the top seem to enter and exit on the same level.

[Vekke]

i have tried to put those manually level so yhey are close...

simulations are run in 3d, but so far with minimal shapes to get hang on this. still not sure if ground level works like it should as some streamlines go throung the calculation box which is set to ground in these



this one is level -10mm from EQXX gives Cd 0,152

How the level changes are modelled:

[freebeard]

I'm trying to answer my own question, what is the underlying architecture of their algorithm. I'm convinced the field of low-end CFD needs adoption of OpenVFD over Cartesian cells.

First find: https://indexcfd.com/cfd-solidworks/

Solidworks The Use
We use the models and drawing files and import all of them in any sort of CFD software as well. What this implies is that when it comes to be able to thermal analyses in addition to FEA, you want to be certain that your convective effects are expected and can end up being accurately described by the above information. Predictable effects may be defined because conditions where you have got to pushed convection and sincere convective coefficient values. ​Particleworks for SOLIDWORKS and SOLIDWORKS Flow Simulation could go with each other to expand design phase CFD simulations. which are difficult to resolve by classical main grid based CFD software like SOLIDWORKS Circulation Simulation.

They acknowledge the issue. Next is www.solidworks.com/sw/docs/Flow_Basis_of_CAD_Embedded_CFD_Whitepaper.pdf

The idea is underpinned by the choice of meshing technology in SOLIDWORKS Flow Simulation
and the impact that choosing a Cartesian-based mesh has on the way the geometry is handled,
in particular solid-fluid and solid-solid interfaces, the wall treatment used to capture boundary
layer evolution, and calculation of skin friction and heat fluxes

That's a 20 page PDF, I guess I have some homework.

The thing about OpenVFD is that it is a sparse, shallow b-tree that can be traversed rapidly with each cell carrying arbitrary attributes like temp, pressure, electrostatic charges -- whatever. Open Source. It's used in the movie industry for smoke and flames.

[Vekke]

licence for this in Finland is 23000€ range so not free at all.
3 month rental 2300€ range.

[Phase]

Any updates of rear tire deflectors?

Wish you could find a model of a Hyundai Ioniq and run it in some simulations

[freebeard]

I'm slogging through the Numerical Basis of CAD-Embedded CFD White Paper

Mesh generation is started by dividing the rectangular computational domain into a set of
rectangular cells (cuboids) formed by intersection of planes parallel to the axes of coordinate
system. The mesh can be refined (by splitting each cuboid into 8 geometrically-similar cuboids)
using various adaptation criteria that can be defined for each solid body (curvature, narrow
channels, small features, etc.) and automatically according to gradients in the solution.

It sounds like they're talking about voxels and dynamic subdivision.

https://develop3d.com › develop3d-blog › new-maxwell-render-plug-ins-announced-for-solidworks-rhino-and-others
New Maxwell Render plug-ins announced for SolidWorks, Rhino ... - DEVELOP3D
The SolidWorks release features updates to Maxwell Render 3.1.1 binaries and support for OpenVDB (i.e. Volume File) to Volumetric extension. Fixes include resolving the colour param that failed to work properly in the Transparent material type, and viewport materials that were not updated on some types of changes.

A free plug-in for Solidworks that incorporates OpenVDB, but at the rendering stage? Does it help with iterative calculations?

[aerohead]

https://ecomodder.com/forum/showthre...tml#post116392a) The images may not have been of sufficient clarity.
b) The diffuser is actually 2.8-degrees upsweep, beginning at 252mm,before the rear wheels.
c) The lower profile indicates for the bottom of the wheel fairing package, below the actual body. I tried to shade them, to provide some visual contrast to set them off from the main body. Sorry!
d) The belly pan at the nose is level, at 252mm ground clearance, however the front wheel fairing bottoms are 'LOW' and follow the front overhang at the SAE 16-degrees 'approach angle, sloping up from the face of the front tires.
Wheel fairings between the axles follow the SAE 10-degree breakover angles. According to the SAE, none of the bottoms of the wheel fairings would experience a ground strike. The lower extremities of the fairings would be level with the bottom of the wheel rims, respecting California State Law. Only the tire sidewalls would project below that level.
I apologize for the lack of clarity. Sunlight managed to degrade the ink.
The von Mises reflection of the rear upper body was a concession to room for taillights and a license plate.

The CFD is very interesting. Thanks so much!

[Vekke]

EQXX diffuser in extended mode this text?:
The diffuser is actually 2.8-degrees upsweep, beginning at 252mm,before the rear wheels.

Have done today more simulations got the ground level working which raise the numbers:
Here is the latests takeoffs. More in the facebook where is more pics:

Added today the EQXX front to model Cd 0,194 with 82mm ground clearance. Here angle of attack 0,5 degrees and diffusor at 5 degree


EQXX full siluette -10mm lowered at the rear. plus same treats as the original EQXX shape. Cd is still better than eqxx basic shape with Cd 0,188 vs without front Cd 0,177. Difference 5,8%


EQXX full siluette -10mm ground clearance 130mm (was 82mm on previous simulations) Cd 0,189. So increasing the ground clearance does not impact the drag a lot in the base body. When you add rotating wheels the lowering has almost linear drop in drag.


Ok here is my first guess how the underside should be done. Now I know from drag studies that straigth diffuser is not best for low drag so I have just adapted the roof curvature in the diffuser. Cd with this 0,184 which is pretty nice drop from 0,188. EQXX original diffusor angle is not 5 degree because in range mode it extends and lowers the angle to around 2,5 degrees. I do MK2 version to see if the upper dip style helps also underside and ofcourse the EQXX style with and without the extension for comparison.


The dip in diffusor seem to work also underside Cd 0,182. Good for two counts again just a guess on the dimensions. My guess with tweaking the shapes on curved diffusor you would get 0,01 lower drag than the 5 degree straigth diffusor.