Underbody aero

[wytbishop]

Removed...becuase it was wrong and misleading.

[youdirtyfox]

I was going to say you'd want to be racecar low for it to make any difference.....

My chassis frame rails are about 2" from the deck

[wytbishop]

After reading a bit more I have to refine my previous statements...

The thing about the common daily driven automobile is that typically the bottom is flat and the top is convex. What this means is that as air flows over the car it naturally accelerates...as it would over a wing...and the pressure goes down accordint to the Bernoulli Principle. As it flows under the car there is essentially no aerodynamic effect. The net result is that the typical car actually experiences positive lift because the pressure above the car is lower than the pressure below the car.

A symmetrical object like the one in the top of the figure experiences that airflow acceleration both above and below so in a free airstream the aero effects cancel each other and it experiences no lift or downforce. However as you move that symmetrical body closer to the ground the air flowing underneath accelerates more than above, due to the restriction.

Whereas with the flat bottom shape, the closer it gets to the ground the more the air is blocked by the sharp edge until essentially all flow is over the top and the only affect is lift. So moving a flat bottom shape very close to the ground eliminates any sort of aero effect under the car entirely. What the bottom figure shows is that you cannot generate downforce with a flat bottom. The best you can do is to generate the least lift by having the ride height at about h/L=0.05 or greater. Any lower and you block the flow under the car and accentuate the lifting affect of the air flowing over it.

I totally misread that figure in my first post.

[wytbishop]

So...having said all that...and I've been thinking about this stuff for a few weeks now, the best solution would be to reshape the bottom of the car as much as possible to cause the desired airflow acceleration.

In my mind this would involve sealing the area from the firewall back to the rear edge of the fuel tank and creating a downward sloping underpan from the lip of the front bumper to the firewall and then a rear pan with largish venturis sloping upward from the rear of the tank to the rear bumper at about 7-10º (which is optimal according to what I've read). The idea is to try to make the underside of the car slightly convex. The cool bit is that the venturis if done properly will create little vortices under the car that increase downforce. You'll never create more downforce than that upward force (lift) which is created by the shape of the top of the car, but you can probably do a lot better at neutralizing it and making the car more stable.

The problems are, a) remodeling the front bumper, rear bumper and possibly reshaping the floor of the cargo area to accomodate the necessary sloping pans and b) you'd have to do a lot of trial and error to find something that actually worked and felt right while driving.

[Ryan]

Igor, I did eat my beret. No problem.

An idea from diffusers to to channel what air does go under the car into one stream, accelerating it by reducing the area it flows through.

So an hourglass shaped channel under the car. Also a nozzle at the front.

[wytbishop]

Or you could just make something that looks cool and not care about the aerodynamics...cuz really...it's not going to get a lot better no matter what you do.

[Ryan]

I was gonna say that, but I thought peoples feelings would get hurt.

[Josh]

By putting a smooth plate on the underside of the car you are reducing the dirty airflow that it creates, but with the way the car is designed already there will be little improvement over stock. This application is purely for looks on a street car.

Now if you were tracking your car and your lap times were consistent, then I could see, perhaps, yielding some benefit reducing the dirty airflow reducing the drag. That could increase your distance and fuel economy, and potentially a little bit quicker in theory. and I say "could" loosely.

its for aesthetics

[fowljesse]

Hmm. I'd like to get a Geo Metro, a blowtorch, and a softball. I could heat up a spot, and dent it in with the ball, over, and over. I remember a Motorcycle helmet a friend had that was dimpled like that, to reduce neck fatigue. I have wondered about this for a long time. Thanks Nd4SpdSe!

[Ryan]

I actually had a discussion about this today. The golf ball dimpling stimulates a turbulent boundary layer which results in less skin friction drag, which is only part of the drag story, the other half being pressure. Skin friction is more important at low speeds, pressure is more important at high speed. The transition from "high" to "low is such that basically all we are ever concerned about in normal, human sized life is considered "high" speed.

[wytbishop]

I am starting to learn how to use Autodesk CFD Simulation. I'm hoping in the coming months I'll be able to figure out how to run a wind tunnel simulation.

Seriously...half of being an Engineer these days (which I am not...just so we're clear) is learning to use all the damn software. Since I started my design project I've picked up, Algor Simulator, Alias Automotive and Inventor Fusion...in addition to AutoCAD Mechanical and standard Inventor I was already familiar with.

[Ryan]

Wyt, when it comes time I can give you some insight in how to run an accurate wind tunnel sim. There are a few very important things that are a b---- to get working properly, such as a moving ground plane.

[fowljesse]

I am very interested in this subject, and really hope you get a simulation going. I want to build a 100mpg vehicle, probably starting with a Metro.

[wytbishop]

Wyt, when it comes time I can give you some insight in how to run an accurate wind tunnel sim. There are a few very important things that are a b---- to get working properly, such as a moving ground plane.

The thing with the computational software like FEA and CFD simulators is that you have to really understand the physics to create an accurate simulation. I have to relearn all the FEA stuff and I've never really been taught the aero stuff in that much detail at all. It's going to take a fair bit of reading to get to where I need to be to do it properly.

I will probably take you up on that Ryan. Thanks.

[Ryan]

FEA and CFD are not related really... I mean, they're both software that solve mind bogglingly large matricies with mind bogglingly large and complicated equations...

For every single element in a CFD analysis, the computer solves at least these equations:



It does that over and over through time, and then the whole thing over and over as an iteration to determine convergence or lack thereof. Its not unusual for CFD sims to have CPU times measured in days, even for some pretty insane computers.


but you don't need to understand both.