WINGS

[Luke Phersson]

Thanks for the kind words guys.

MCoach,
There really isn't anything different between a 'diffuser' and a wing - all the principles remain the same. Airfoils also have a 'diffusing' section which is designed to recover pressure as quickly as possible without stalling. An undertray (which has a diffusing section) on a car just differs in that it doesn't have an aerodynamic top section, and it also has to deal with a lot more turbulent/messy air than a typical wing.

Moop,
The ground and the speed of the on coming air isn't really the whole story. Picture it like this, the car is traveling at 100km/h with no ambient wind, so using the car as the reference frame the road is moving under the car at 100 km/h, same as the on coming air. What happens underneath the front wing? The air is accelerated above 100 km/h. We now have a velocity gradient between the road and the air, this develops a kind of boundary layer.
Generally this symmetry/streamline assumption is used in wind tunnels without a rolling road to simulate ground effect. We tried this a few years ago..

Essentially you just build a mirror image of the front wing/undertray, then connect one of them to a force balance (or pressure tap it like we did^). The middle streamline between the two bodies is supposed to represent the ground - it does this far better than a non moving floor, but is still not completely accurate, due to the reasons above.

Haha yeah, aerodynamics is full of 'chicken or the egg' type problems. Did the curved streamlines create the pressure gradient, or did the pressure gradient curve the streamlines... One of the more ridiculous explanations of lift involves packets of air traveling along the top and bottom of the wing mysteriously having to meet up at the same time at the trailing edge...

There is nothing wrong with just sticking a big sheet of aluminum under your car and manually bending up an inlet or diffuser.. Then just drive different radius circles and record times. Alternatively you can fit soft springs and shock pots to measure displacement vs. velocity (beware varying ride heights though). Maybe play around with different diffuser angles? Agricultural but it will tell you if it works or not. If you can show that this was the best use of your time/resources design judges should not mark you down for that.

We actually do the different radius circles to figure out our 'effective' CL.A and CD.A, which takes into account load sensitivity and extra tyre drag due to aero.

Luke.

[Z]

Luke

"Start rant..
You’d be surprised at how few engineers actually understand how wings work – the actual mechanism for lift..."

I'm more surprised at how few Fluid Dynamicists get it! In fact, I have so much to say on this subject that I'll have to leave it for another time.

Meanwhile, thanks for the above objective information (err,... any chance of some kPa numbers for those pretty colours? ).

When I have more time I will post some numbers that I think are possible, and some simple underbodies that might do it.
~~~o0o~~~

Moop,

Don't give up on underbodies yet!

Also, as Luke said (and IMO), driving in circles is the best way to test these things. Do CFD to get a rough idea of what might work. Then plywood, ally sheet, Maxbond, duct tape, wool-tufts, circles... Then more cutting, bending, PU spray-foam, "gotta get these wool-tufts pointing in the right direction", and more circles...

Z

[Hi-oct]

Now i currently working on mulielement wings. I observe that the downforce is max and drag min when the overlap is 0, which counters the theory I learnt. Also, as gap increases downforce increases continuously. Is this the right observation or am I doing anything wrong in my CFD?

[JulianH]

I'm quiet sure that your CFD is wrong.

Increasing gap over a certain point wil definitely decrease your downforce.

Overlap = 0 is also not the optimal point when it comes to downforce in our observations. Additionally "max DF and min drag" at one single point sound quiet counter intuitive.

Could you upload a picture of your setting? That might help.

[Z]

Most of the Northern hemisphere comps are now over, so teams will be celebrating, or commiserating, and hopefully thinking hard about how to do better next year. With the exception of those very few teams who won their comps, the best way to do better is to score a lot more points.

So, how do you score a shedload more points?

Well, AERO, of course!!! (Which is why I dug up this old thread... )

So in the next post I give some "big-picture" numbers showing how DOMINANT aero can be in FSAE. In later posts (next week?) I will give some hints of how to achieve these numbers.

Z

(PS. It seems this new Forum doesn't like long posts, hence I have to break this one up...)

[Z]

WHY AERO?
===========
Consider the following two cars.

Car-1.5G.
This car corners at ~1.5G = ~15m/s^2 (lateral Gs, flat track, as many good FSAE cars can do these days).
So this might be a non-aero car with the good racing tyres currently available, which have average Mu = ~1.5 (ie. including TLS, etc.).

Car-3.0G .
This car corners at ~3.0G = 30m/s^2.
So this might also be a non-aero car that happens to have magical tyres with Mu = ~3, but good luck finding those!

So, instead, assume this is an aero car with similar (though slightly larger) tyres to the 1.5G car above, but with "drive on the ceiling" downforce at the given corner speed (ie. DownForce = Weight of car).

Note that 3G cornering is quite tame these days, with many aero racecars getting up into the 4 and 5Gs. These 4+G forces are hard work for the drivers, but are by no means impossible.

Now consider the points difference between these two cars in the Dynamic events (as calculated by the formulas in one of the 2013 Rules floating around the web...).

(Note that current good aero FSAE cars are somewhere between the above two cars, although closer to 1.5G than 3.0G. The following is intended to show how dramatic a difference high cornering speeds (or Gs) can make.)
~o0o~

ACCELERATION.
No significant difference here, because this event is won mostly at "launch" where the speed is too low for any big aero effect.
~o0o~

SKID-PAD.
This is the easiest event to put numbers to, with,
Car-1.5G having laptime = 4.8 seconds, with speed = ~11 m/s (40 kph, 25 mph), and
Car-3.0G having laptime = 3.4 seconds, with speed = ~16m/s (57 kph, 35 mph).

With the 3.0G car as "Tmin", and thus scoring the maximum 47.5 "performance" points, the "maximum" time Tmax = 1.25 x Tmin = 4.25 seconds. Thus any car slower than this, such as the 1.5G car (which would be considered very fast these days), scores NO performance points at all! In fact, to score any points a car would have to corner at greater than 1.9G.
~o0o~

AUTOCROSS.
Here the scoring formula has Tmax = 1.45 x Tmin. This means that any car taking 45% more time to complete the event than the fastest car (Tmin) scores no "performance" points out of the 142.5 points available.

The 3G car corners 41% faster (ie. x sqrt(2)) than the 1.5G car. So, ASSUMING (see below!) that these speeds are similarly translated to overall average speed, any cars cornering at 1.5G or less score maybe 10 points at most.
~o0o~

ENDURANCE.
Again the scoring formula has a scaling factor of Tmax = 1.45 x Tmin, but this time with 250 "performance" points available. So, using same assumption as above, this time any cars cornering at 1.5G or less score maybe 20 points at most.

The assumption above, namely that cornering speed advantage is also translated to the rest of the track, is a big one. But the essence of FSAE-style "autocross" racing is that it is mostly about corners and the transitions between them. Many of the recent FSAE events (from video I have seen) seem to be an almost endless series of corners with very few, and rather short, straights between them. Also, coming out of a corner faster than the other cars, and being able to enter the next corner faster, makes it much easier to develop a higher average speed along any of the straights between the corners.

So, if the 41% speed advantage is even close to true, then, so far, the 3G car has a massive advantage of over 400 points (!) over the 1.5G cars. Cars capable of 1.5G cornering are quite fast by today's standards, although current aero cars are gradually passing this mark.

But what about the oft-mentioned "big disadvantage" of aero? .....
~o0o~

FUEL ECONOMY/EFFICIENCY.
Firstly, note that in olden days racing (pre WWII), "aero" was about "streamlining" the car to reduce drag, so as to increase top speed in a straight line (the top speeds were the same in the 1930s as now (high 300s kph)). This streamlining is still useful in FSAE today, but would be used mainly to minimise fuel usage (wastage!), rather than increasing top speeds.

However, in FSAE, as in most modern racing, "aero" is mostly about adding downforce to increase corner speeds. Fortunately, this can also lower Fuel usage for the simple reason that the car can win comfortably without braking hard before corners (throwing away valuable kinetic energy), nor accelerating hard out of the corners (burning valuable fuel to regain the kinetic energy just thrown away).

[Nerd Note]
Force-vector(Dot-product)Velocity-vector = Power.
If F and V are parallel (= acceleration), then Power is high and positive (= burn fuel to make kinetic energy).
If F and V are anti-parallel (= braking), then Power is high and negative (= heat brakes to dump kinetic energy).
If F and V are almost perpendicular (= cornering at low slip-angle), then Power is almost zero (= good!).)[/End Nerd Note]

So, energetically speaking, not only are high cornering Gs almost free, but they can, in fact, save fuel!

And, very importantly, A HIGH DOWNFORCE CAR CAN ALSO BE A LOW DRAG CAR! (It is true! See next week... )
~~~~~o0o~~~~~

CONCLUSION.
So, IMO, the goal of "aero" in FSAE is to build an almost "constant speed car". This car should be capable of very high lateral Gs, but require negligible longitudinal Gs. Imagine a car like the "Wild Mouse" fairground rides. These rides have ~constant downward slope, so constant forward speed, but also a seemingly never-ending series of corners at 3+Gs lateral. For your "Initial Design Concept Meetings", I suggest you take your data-loggers to the fairgound.

The potential points advantage of this sort of car against the current "state-of-the-art" non-aero cars is far in excess of any other means of improving performance. I leave it to you students to do more detailed point and laptime simulations. But no amount of increase in power (limited by restrictor), reduction in weight (not a huge factor TLS-wise), magical new tyre compounds (good luck with that...), Nth degree refinement of suspension (again, ultimately limited by tyre Mu), nanosecond paddle-shift gear-change (), or even the best driver in the world (once again, limited to the ~1.5G that the tyres can bring...) will ever come close to the huge gains possible from aero.

Also worth noting here, a team that does the absolutely best that is possible in the Static events of Cost, Presentation, and Design, might only get a handful more points than the next best teams. That is, the Static events are scored essentially independently. But in the Dynamic events (as currently scored) an exceptionally fast team pushes all the other teams backwards. The ~400 point gain in the above examples comes from taking points AWAY from the other teams.

Before getting to the aero details, it is worth noting that these 3G numbers will require a LIGHTWEIGHT and STRONG car. The car SHOULD be lightweight because the increase in competition points won is directly proportional to the ratio of aero-downforce to car weight. Less weight means more points from a given amount of downforce. The car MUST be strong because the forces acting on the car increase in this same ratio (ie. the structural loads become much greater than those of a non-aero car!).

Hence, my many suggestions over the years to keep the rest of the car as simple as possible. So, think of a HEAVY-DUTY "brown go-kart, with aero-undertray".
~~~~~o0o~~~~~

AERO DETAILS.
Finally, what is needed from the aero to achieve the above?

Assuming a 150 kg car (not too hard, with KISS), with 50 kg jockey and the currently available Mu = 1.5+ tyres, the DF = W equation suggests the car needs ~200 kg of downforce (ie. ~2 kN), at the given corner speed. As shown above this corner speed is ~16 m/s on the Skid-Pad, and this is also close to the current average AutoX/Enduro speed. This speed drops to about 10 m/s around any hairpins, as they are specified in the Rules. But most circuits these days seem to have very few real hairpins, despite discussion on other threads asking for more of them.

So, if your car can develop at least 2 kN of downforce at 16m/s, then you should have a massive jump on the non-aero opposition. Well, at least until they catch up! And some teams are (slowly) getting there ...

In conventional aero terminology (Force = Half-Rho-V-Squared.CL.A) the above numbers imply a CL.A = ~13 m^2.
(3G cornering around a hairpin (V = ~ 10 m/s) implies CL.A = ~33 m^2.)

A fairly compact undertray can have Area = ~2 m^2.
Utilising most of the plan area allowed in the Rules (say, about 1.4 m wide, by 3 m long, less cut-outs for wheels) gives Area = ~3 m^2.

So, for 3G cornering at average Enduro speeds the total car lift-coefficient, based on plan area, needs to be ~CL = 4 to 7.

Are these CL numbers at all achievable?

Looking through aeronautical textbooks would suggest NOT LIKELY!!!!!
(Err, ... except that Handley-Page got CL = 4+ back in 1910, and that wasn't even in ground-effect... )

So, should I go and talk to the fairies at the bottom of the garden???

As always, comments and criticisms welcome.

More next week...

Z

(PS. As a teaser, putting the above figures in a more sensible way, and depending on plan area, the average pressure drop under the car for 2 kN downforce needs to be about -700 Pa to -1,000 Pa (since most aero forces come from "suction", not pressure above ambient, and 1 Pascal = 1 N/sq.m). This is less than 1% of atmospheric pressure, which is 10 tons/sq.m (= 100 kPa).)

[JulianH]

I think you are correct in one area: An Aero-Car achieving this magical numbers would destroy everything out on track.

But, after that, it gets difficult:

Note that 3G cornering is quite tame these days, with many aero racecars getting up into the 4 and 5Gs. These 4+G forces are hard work for the drivers, but are by no means impossible.


(Note that current good aero FSAE cars are somewhere between the above two cars, although closer to 1.5G than 3.0G. The following is intended to show how dramatic a difference high cornering speeds (or Gs) can make.)

I think it is not possible to achieve 4G's in FSAE. The speeds are just too low so that Aero isn't even more important. F1 cars (with lower cL*A compared to massive FSAE-Aero) achieve these figures at 200kph+ when Aero is like 2-3-4 times their weight

Good FSAE Aero cars reach over 2.5G's on a FSAE track (I heard in Lincoln it can be even more. But in Germany we logged about that). So we are not too far away. "Problematic" for your calculations is, that good Non-Aero cars pull close or even over 2G's... So the "Delta" is not that large.

SKID-PAD.
This is the easiest event to put numbers to, with,
Car-1.5G having laptime = 4.8 seconds, with speed = ~11 m/s (40 kph, 25 mph), and
Car-3.0G having laptime = 3.4 seconds, with speed = ~16m/s (57 kph, 35 mph).

With the 3.0G car as "Tmin", and thus scoring the maximum 47.5 "performance" points, the "maximum" time Tmax = 1.25 x Tmin = 4.25 seconds. Thus any car slower than this, such as the 1.5G car (which would be considered very fast these days), scores NO performance points at all! In fact, to score any points a car would have to corner at greater than 1.9G.
~o0o~

The problem here is simple. It is not possible to achieve 3.0G in Skidpad. It just won't work. It probably even comes down to the driver. As said before our car, that is able to destroy all Non-Aero cars we ever built on track is only 2 tenths faster than our last Non-Aero Car from 2011 in Skidpad. Maybe the kinematics were better back then but still...
Even further. The current FSAE "World Record" in Skidpad is somewhere at 4.6-4.7 seconds (Monash at FSAE-A 2012 or Rennteam Stuttgart at Silverstone 2013). Even the car with probably the best Aero out there cannot pull more than 1.5G in Skidpad. Because the speeds are just too low.

AUTOCROSS.
Here the scoring formula has Tmax = 1.45 x Tmin. This means that any car taking 45% more time to complete the event than the fastest car (Tmin) scores no "performance" points out of the 142.5 points available.

The 3G car corners 41% faster (ie. x sqrt(2)) than the 1.5G car. So, ASSUMING (see below!) that these speeds are similarly translated to overall average speed, any cars cornering at 1.5G or less score maybe 10 points at most.
~o0o~

You can't drive 41% faster than a Non-Aero car that is good. It simply doesn't work! In Germany 2012, the fastest non-Aero car in AutoCross was Delft with a 76.370s. 41% faster would come to a 45.05s. The fastest Aero car did a 75.933s. That means not 57kph average speed but 97.6kph. Sure the Delft car was exceptional but that is always an issue .
And the German track is "Aero friendly". It is just ridiculous to come up with such numbers and then say "ah you are so fast the other don't get any points then...".

So, if the 41% speed advantage is even close to true, then, so far, the 3G car has a massive advantage of over 400 points (!) over the 1.5G cars. Cars capable of 1.5G cornering are quite fast by today's standards, although current aero cars are gradually passing this mark.

Yes, and a 400hp car with 120kg is also nice...

But what about the oft-mentioned "big disadvantage" of aero? .....
~o0o~

FUEL ECONOMY/EFFICIENCY.
Firstly, note that in olden days racing (pre WWII), "aero" was about "streamlining" the car to reduce drag, so as to increase top speed in a straight line (the top speeds were the same in the 1930s as now (high 300s kph)). This streamlining is still useful in FSAE today, but would be used mainly to minimise fuel usage (wastage!), rather than increasing top speeds.

However, in FSAE, as in most modern racing, "aero" is mostly about adding downforce to increase corner speeds. Fortunately, this can also lower Fuel usage for the simple reason that the car can win comfortably without braking hard before corners (throwing away valuable kinetic energy), nor accelerating hard out of the corners (burning valuable fuel to regain the kinetic energy just thrown away).

So, energetically speaking, not only are high cornering Gs almost free, but they can, in fact, save fuel!

That's correct. With the new formulas, you just drive the hell out of your car. GFR 2013 in Germany won Endurance and Efficiency. That works.

And, very importantly, A HIGH DOWNFORCE CAR CAN ALSO BE A LOW DRAG CAR! (It is true! See next week... )

That's were I don't think you are correct. You presented some "ideas" in the past. And I tried them all and they all don't work. If you want high Downforce you have to pay the "Drag price". But as mentioned, it's not a big issue. Even with a possible "3G car everywhere" you just use a DRS (you need it anyway because you would break the whole thing at speeds close to 100kph).

CONCLUSION.
So, IMO, the goal of "aero" in FSAE is to build an almost "constant speed car". This car should be capable of very high lateral Gs, but require negligible longitudinal Gs.

That is true. GFRc 2013 is the best example.

So, if your car can develop at least 2 kN of downforce at 16m/s, then you should have a massive jump on the non-aero opposition. Well, at least until they catch up! And some teams are (slowly) getting there ...
In conventional aero terminology (Force = Half-Rho-V-Squared.CL.A) the above numbers imply a CL.A = ~13 m^2.

You can't pull 2kN at 16m/s. It won't work. Period.
cL*A of 13 is just not possible.

So, for 3G cornering at average Enduro speeds the total car lift-coefficient, based on plan area, needs to be ~CL = 4 to 7.

CL = 4 is possible. 5 maybe, 6 nah.



Summing Up:
Yes, Aero works. Yes, Aero gives you an advantage.
No, you can't get 400 dynamic points compared to a good Non-Aero car.
No, you can't pull 3G in a slow corner.
No, you can't achieve your magical downforce numbers.

I'm quite sure that we all are doing "bad" Aero designs. I mean the most teams are doing it for 2 years. Look at all the "first 2 years chassis". Of course in next years the cars get better. But even then, you will not achieve such numbers.
Talking about "ah everybody does it wrong" does not help. As said, we are far from "optimum" but more when it comes to efficient designs not max. downforce.

Other thing:
I heard they want to penalize Aero again because you "need it" to be fast. So probably with the 2015 Rule Change it even gets more difficult.


Cheers,
Julian

[DannytheRadomski]

In terms of building a high lateral downforce car with low longitudinal downforce, would active aero be best? Something like the MP4-12C or the Veyron? It would have to calculate angles for the wings based on speed, braking, throttle position, steering position, etc., so it might be a little much in terms of programming. Then again, the payoff could outweigh the drawbacks if it were done well.

[mech5496]

Danny, check out what Sooner racing team (and a few others) have done...

http://www.youtube.com/watch?v=7wozSqFXitY

[JulianH]

In terms of building a high lateral downforce car with low longitudinal downforce, would active aero be best? Something like the MP4-12C or the Veyron? It would have to calculate angles for the wings based on speed, braking, throttle position, steering position, etc., so it might be a little much in terms of programming. Then again, the payoff could outweigh the drawbacks if it were done well.

Active Aero is a possibility. Karlsruhe just does it with driver input.
I think the most "important cases" to activate DRS can even be programmed quite simple. Speed/Throttle > X and Steering < Y or something like that. Monash showed that the "closing" of the DRS can be done very quickly so just the braking pedal or the linear acceleratometer could be used.

Don't know how complicated that would be in a combustion car, but if you programm Torque Vectoring in an electric car, it's a bit more difficult, at least they tell me that all the time