[electronbee]

On the Eco-Modder Forums there are people doing mock air flow animations with Flow Simulator by S.I.Chernyshenko.

Basically, you make a profile of your car in black and white, upload it, and wait for the video. It's not scientific as he states since this project was designed for speed and simplicity over pure data driven CFD.

But, despite that, many people on the Eco-Modder Forums still use it. Any ways, I did four of them:

,

,

,

.

Nifty, ain't it?

 

[AztecRol]

That is pretty cool. Thanks for sharing.

 

[Wrenchmeister]

Cool.

Is the windshield angle right?

 

[electronbee]

The windshield angle should be right. As I took a profile pic of an Element that I found off of the internet and then converted it to a black and white picture. However, in that process or in the downsizing process to 256x128 the angle may have been approximated.

But, I did not alter any angles or dimensions. The only things that I added were the subtraction of the wheels and the addition of the roof rack and spoilers.

I think I'll try one with that spoiler from the Honda Mugen site. I personally did not like that look but it might help with MPG.

 

[Rob Dobbs]

when I get home, im totally going to put a little box in front of a smaller box and see what airflow that gives.

 

[racerc2000]

The windshield angle should be right. As I took a profile pic of an Element that I found off of the internet and then converted it to a black and white picture. However, in that process or in the downsizing process to 256x128 the angle may have been approximated.

But, I did not alter any angles or dimensions. The only things that I added were the subtraction of the wheels and the addition of the roof rack and spoilers.

I think I'll try one with that spoiler from the Honda Mugen site. I personally did not like that look but it might help with MPG.

if your referring to that stacked spoiler. i believe it would be hard to test as air goes under and over it

 

[electronbee]

One of the pitfalls of Flow Illustrator, or rather, limitations, is that is it strictly 2D. Therefore, you have to abbreviate the picture somewhat. The only real way would be to do an actual wind tunnel test and collect data. But, that can be expensive and actually, its hard to find a place that has a wind tunnel big enough for a car. A lot of universities have small ones for models but the large ones are rather rare. Also, there are no large scale models of the Element. That 1/43 one that was available is really small to try any aero modifications.

Also, the resolution of Flow Illustrator greatly abbreviates the airflow. You can model above and below the spoiler if you were to make a BMP of just the roof line and the spoiler. That would probably work and give you an idea.

 

[Eww-an E]

Very cool, thanks for sharing.

 

[racerc2000]

apparently going by the eco modder formula. the "spoiler" helps for aero and economy the rear deflector (old style) hurts economy

judging by shape and profile the spoiler acts as an extension of the roof with a sharp cut off which is basically ideal for us thinking inside the box ^.^

I still have to finish my belly pan and do a tuft test ^.^

 

[psschmied]

The video isn't clear enough to let me see how the rack section was simulated. I would expect to see turbulence vortexes behind each crossbar.

If the simulation used the support feet profile, not the crossbars modeled as two ovoids, detached from and above the roof line, the output wouldn't be valid. (It'd be like modeling the wheels as solid cylinders.)

Another limitation of the simulator is boundary layer effects. At low resolutions, a digitized profile has ragged step irregularities that introduce surface turbulence not experienced by a continuously curved smooth surface.

Surface modeling is more important than you'd suspect. Modeling of dolphins and sharks that didn't correctly describe their skins led to calculated maximum speeds that were only 25-30% of their observed speeds.

 

[electronbee]

Unfortunately, as the simulation is only 2D you have to cut off parts and then make some assumptions. Or, you can do multiple simulations, such as an enlarged part of the roof rack feet only. I chose to eliminate the roof rack feet as their overall aerodynamic foot print is really quite small in comparison to the rest of the vehicle.

Ideally, you'd get some real CAD program that can do CFD and simulate a rolling tire to get a true visualization of what is occurring underneath the car and in the wheel wells. But, another thing to consider is that those Eco-Modders address every possible detail to increase MPG -to the point of connecting and disconnecting belt driven accessories to reduce parasitic engine loss-- and it is basically known/proven that the rear of the vehicle offers the most improvment of aero mods. Followed by a belly pan, then, you address the things that make smaller improvements like the wheels and mirrors.

I'll try the simulations again with a 512x256 image. Some people had issues with the 512x256 size so I stayed at 256x128. If 512 works I'll bump up to 1024, etc.

 

[electronbee]

So, the 512x256 did not work. This thing is kinda tricky so you have to be careful with the parameters. I dropped the time down from 40 seconds to 10 seconds to see if that will work.

 

[curt504]

FWIW got .4 MPG or so from removing the SC's stock racks and gas tank cage.

Anyone do some real world experiments? A scan or ultra gauge makes the data collection on real time MPG easy.

curt