Downforce Numbers

[magicweed]

I was wondering what some of the teams were running for actual effective downforce on their cars. I've been playing around with Javafoil for the last day, along with an excel sheet I made up that calculates Effective Downforce, Effective Drag, and Effective G's for any FSAE car given its weight and speed. Just wondering how the numbers translate, as I just finished designing a foil with a Coefficient of Lift around -3.3 that creates 500 Newtons of downforce @ 20mph.

[Marshall Grice]

java foil is cool. you can make just about any downforce you want with it. the question is at 20mph are you wings stalled?

if you get something around -6 with relatively sane aoa's then you might have a decent profile in real life, but no gurantees.

*also team smartass

[magicweed]

All the angle of attacks have been very reasonable, right around 12-14 degrees, and seem to be getting less and less as i change the variables. Has anyone done any testing on a similar simulation situation to real life?

[Agent4573]

that number seems a bit high to me.... unless your running a god dam huge wing.

[magicweed]

Rear Wing Specs:
Chord = .6096m
Span = 1.27m
Coeffiecient of Lift = 3.5
Coeffiecient of Drag = .03216
velocity = 13.5 m/s

Front Wing specs:
Chord = .5199m
Span = .8382m (sum of both sides of frame)
Same stats as above

Single element wing (because I havent gotten the hang of multi-element wings in simulations yet)

Total Downforce @ 13 m/s = 474 Newtons (~20mph)
Total Downforce @ 20 m/s = 1039 Newtons (~45mph)

I want to know how acurate these simulations can be. If I find the time I'll build a small scale model for the little windtunnel we have, but I'd like to know if Javafoil is relatively accurate, or if I should be using a more advanced CFD program.

[Scott Wordley]

Is that a multielement wing?

That force number does seem high for that speed and lift coefficient.

By my calcs you would need the wing to be about 3.5m^2.

Given the rule limitations that would make the wing about 2.5m in Chord for a 1.4m span. Thats an aspect ratio of about 0.5 which is not good.

For a wing like that it would also need to be very high above the driver, or you lose a lot of downforce.

Im presenting a couple of SAE papers in detroit next year on FSAE aero which will give you an idea of what is possible for these cars.

Keep in mind that you will lose a lot of performance going from javafoil to 2D CFD to 3D CFD to experimental.

[Marshall Grice]

we've run our wings in java foil but never with the rest of the car. we also have never measured downforce of the wings with them not on the car. so it's a little difficult to compare java foil to actual car in our case. although the java foil numbers were significantly different then what we measured and running multi element wings in java foil was even worse. I'd give numbers but our aero guy would probably kill me. i'll let him respond if he's interested.

[Jersey Tom]

112lbf of downforce is a damn lot for 20mph. Just intuitively I find it questionable.

[Z]

Stan,

Can a real wing in freestream have Cl = 3.5? Sure, but it would have to be a multi-element wing.

Can a real (inverted) single-element wing in ground effect have Cl = 3.5? Possibly, though again it is easier to do it with multi-elements.

I am not familiar with Javafoil, but your figures look right (ie. Lift = half-rho-V-squared x Cl x Area - who needs a computer!).

The three important points are that a racecar wing typically has low aspect ratio (bad), can fly in ground effect (good), and has other nearby bodies messing up the airflow (usually bad). The first two cancel each other out to some extent, and the third can be minimised with attention to detail.

I guess that Javafoil does not model the "other" bodies (wheels, etc.), but does it account for aspect ratio and the ground?

Z

PS. Here's (http://fsae.com/eve/forums/a/tpc/f/125607348/m/46710996521/r/69510777521#69510777521) where I posted on a similar topic.

[Bryan Hagenauer]

If I remember correctly, I was getting Cls of 5.5-6 range for our front 3 element wing in Javafoil- infinite span, ground effect. That isn't even close to reflecting reality.

When you start messing with multis, keep in mind that javafoil has weak code when it comes to wing size. If you increase the chord of the wing, it actually increases the Cl. So for example, if you wanted a total chord of 15", then scale each element down as the correct % of 15" to maintain an overall chord the same as the default in the program. Otherwise you get exaggerated results.

[magicweed]

It does account for the ground and aspect ratio. I just turned on the ground option, as I was doing clear air simulations. The Lift Co went to around -5.575, when the aspect ratio is at 0.

Normal? Abnormal? I'm new to Aero as I'm an EE that should've been a mechanical engineer. I'd rather design differentials, intakes and exhausts than program microcontrollers, but it took a race car to teach me that.

[Z]

Stan,

Another thing to consider is that a racecar wing doesn't have to look anything like an aeroplane wing to work well. Aeroplane wings are designed to meet two goals that are of little importance on a racecar. Namely;

1. An aeroplane wing section must be structurally very efficient. That typically means it must be "fat" (ie. deep section for bending strength, and large section area for torsional stiffness). A structurally inefficient wing will either be very heavy (bad), or will fall off (worse http://fsae.com/groupee_common/emoticons/icon_frown.gif)!

Racecar wings are short and can have multiple supports, so they can be quite thin. Thin wings are actually better aerodynamically than thick wings - try it http://fsae.com/groupee_common/emoticons/icon_smile.gif.

2. An aeroplane wing must generate an approximately constant lift force (= the weight of the plane) at largely varying speeds. So it has to operate effectively at largely varying angles of attack (to give the necessary range of Cl's). So it has a rounded nose.

If an aerofoil only has to work at one angle of attack, like mandated on most racecars, then it can have a razor sharp nose and still work well.

So an effective racecar wing can just be a curved sheet of aluminium or carbon fibre with bevelled edges. Or better yet a multi-element series of nose-to-tail curved arcs (the curvature at the noses is important). Again, try it!

Z

PS. Yes, it is normal for wings to increase their Cl when in ground effect. But an inverted wing, when too close to the ground, is likely to stall.

[Chris Boyden]

a razor sharp nose

Careful with the minimum radius of curvature rule regarding this as it applies to FSAE.

[Z]

Chris,

Yes, I was speaking generally. Just making the point that efficient aerofoils don't have to be "teardrop" shaped.

For a formula that is supposed to be relatively "free", the FSAE minimum radius for aero parts is crazy! Especially the 6mm thick trailing edges! Other body parts (eg. muffler guards, etc.) apparently can be "razor" sharp. Likewise the "no wing behind the rear wheels" rule.

But despite these crippling aero rules I still think the first FSAE team to do downforce right will clean up http://fsae.com/groupee_common/emoticons/icon_wink.gif.

Z

[Dave M]

But despite these crippling aero rules I still think the first FSAE team to do downforce right will clean up

There have been a few schools that have done aero right. UTA, UMR (had an article in racecar engineering), and Monash pop in first, we are doing well this year due to having some good aero guys on the team. Look at UTA at nationals, they rule all unless another team brings wings.

[Marimvibe]

A Cl of 3.5 is incredibly high. It would be possible for a multi-element wing (talking Monaco F1 number of elements.) It's definately higher than any airplane would be configured for full flaps-down. You definately wouldn't be getting a L/D of 10. A F1 rear wing gets maybe 4. Also, with an AR of 2, you'll lose a lot of that to upwash.

I'd suggest just going to CFD. Java foil is a simple panel method code, and you really can't get any idea of how the tires, body, frame, etc is going to effect the wing. Once you get the geometry created (my current fun little project,) the CFD itself isn't that bad.

[Bryan Hagenauer]

Agreed. I've actually accomplished more just trying out different wing/endplate designs in Floworks than I did playing with profiles in javafoil.

If you arn't able to use a more powerful 2D program than javafoil, then just pick something that matches all the theory presented by Katz and McBeath in their books, and spend some time figuring out what the best span, chord, ground height, endplate shape etc. Don't try to tweak that last 5% out of a program thats at least off by 25% anyway.

In addition, the best freestream profile in the world isn't going to do squat with a tire just inches behind/below it.

[magicweed]

Great response to this. I guess it's time to go bug the department for a good CFD program. Any recomendations? I heard Floworks in there. It looks like it's a program compatible with Solidworks models. Any programs compatible with ProE Wildfire 2.0?

[RobSchäel]

I´ve heard fluent is better than CFX if you are using a Proe Geometry.
I think GAMBIT's journaling capability allowed a parametric approach to model building to be used. In this way, frame sizes, frame lengths, number of cooling holes, etc., are variables and easily changed.
Also , CFD fits into the concurrent engineering strategy right back at the conceptual design stage and provides us with a powerful tool to improve our cars.

[Bryan Hagenauer]

You can get 2 licenses of Fluent/gambit for about $100 a year for the FSAE program by emailing university-at-fluent.com

Floworks is nice if you are using SolidWorks already, but the simplicity of the program is both its greatest asset and largest downfall.

[mtg]

Originally posted by Z:
For a formula that is supposed to be relatively "free", the FSAE minimum radius for aero parts is crazy! Especially the 6mm thick trailing edges! Other body parts (eg. muffler guards, etc.) apparently can be "razor" sharp. Likewise the "no wing behind the rear wheels" rule.

But despite these crippling aero rules I still think the first FSAE team to do downforce right will clean up http://fsae.com/groupee_common/emoticons/icon_wink.gif.

Z

Yes, some of the aero rules are retarded. They seem to make them up at will also, after going through tech with wings for a few years.

One rule I always thought was hilarious is why they don't want people running horizontal decks on the front wing- it could injure a worker if they got hit by it. So, I guess its better to get hit by the metal control arms.

The best aero you could get isn't going to "clean up" at the Silverdome. It helps, but its not like non-aero cars will be uncompetitive. You still need good drivers, well sorted car, etc etc.
If you have the aero chassis mounted, you have to run a much stiffer car, hurting grip over the "unsmooth" parking lot. If you run the aero mounted to the uprights, you have to raise the Cg to allow for the extra ~1" of clearance needed for a thick undertray floating around under the chassis.

Now, a smooth parking lot on the other hand, the stiff car will work better (than at the Silverdome).....Time will tell what the new surface is like.

It sounds like there should be more aero cars, instead of just UMR, UTA, CP Pomona, and a couple others that come and go.

[Marimvibe]

I'm using Fluent at the moment. I also intend on checking out TetrUSS at some point. It's a good code that you can get for free from NASA (as long as you're a US citizen,) but only runs on Unix and Mac and I've heard it's difficult to learn.

Be wary of CFD. Some geometries it works well for, others not so much.

[Scott Wordley]

mtg,

what proportion of your total downforce do you think is coming from the undertray?

If its less than say 40% which im guessing it is, then there is no reason not to mount your wings unsprung. Its quite easy, and very beneficial for the front wing and your wheel rates.

Have you guys been in the tunnel much, I think I remember hearing that you tested at Ford? Have you measured your downforce on track at all?

[mtg]

Originally posted by Scott Wordley:
If its less than say 40% which im guessing it is, then there is no reason not to mount your wings unsprung. Its quite easy, and very beneficial for the front wing and your wheel rates.

Have you guys been in the tunnel much, I think I remember hearing that you tested at Ford? Have you measured your downforce on track at all?

You'd have to ask Deputy Downforce about the specifics, but I doubt he'll say much.

I'm graduated and gone, so I'm not sure exactly how deep they're diving now, but the old aero packages had wind tunnel and track testing.

There could be an advantage to mounting a wing unsprung, but it adds a complication of opposite wheel travel stiffness into the mix.

[andyman61]

Like mtg said, no comment on actual downforce numbers...sorry!

As far as testing goes, yes we've had the privledge to use Ford's wind tunnel in Detriot for the past two years. We've gotten a lot of good info and flow visualization from it. This year, we've also done a series of coast down tests to see what kind of numbers we're actually seeing on the car and what kind of effects rotating tires and other stuff we can't simulate easily in the wind tunnel have.

As for unsprung aero, I'm yet to be conviced that's its benefits completely outweigh all the problems it causes (or can cause if not done correctly). The inital gains look very promising, but the devil's in the details.

-Deputy Downforce

aka

[German]

hello my name is German, I am of the University Simon Bolivar.
In these moments am I realizing a project on ailerons multi elements, and I would like to know if someone has information of them, on, to that profiles are most used?, in that angles? which are his ropes? if someone has this information (although it is an example that serves), it would be quite useful to me... my email is German_fsae@yahoo.es

[Bryan Hagenauer]

I think you will find that what you are asking for is stuff that no one is willing to share. Sorry.

As a point in the right direction though, try checking out the Selig profiles, specifically the 1221 and 1223. They are a good starting point for high lift applications.

[RiNaZ]

My lab has floworks, Star CD and Fluent. Does anybody know what the differences are. I know different program will give different output depending on what your needs are, but just checking if any of you guys had try all 3 and if you have any preference. Thanks

[Scott Wordley]

German,

I'll be presenting 2 papers on Aero for FSAE at next years SAE conference in detroit. They will have a lot of info to get you started on designing a wing package for FSAE.