Automotive Aerodynamics

[homeforsummer]

Continuing this from the discussion in the Toyobaru thread, so it doesn't go too far off topic

It actually isn't, at least not a water formed teardrop.

The shape of a teardrop is partially cause by its Viscosity, the ideal shape as speed record attempt shows is more an elongated shape.

and teardrops doesn't account ground effects, which you have to in a car.

I should clarify: It is, but with automobiles the most practical shape is a half teardrop, where the beginning of the roofline marks the highest point of the theoretical teardrop shape. The natural aerodynamics of a drop of water are influenced by surface tension - the aerodynamics of a rigid vehicle structure aren't.

What you do with everything forward of the windscreen is actually largely academic compared to what you do from the roof backwards, but obviously it helps to try and make things as slick as possible. You'll also notice some of the most aerodynamic cars - GM EV1, Mk1 Honda Insight - are also semi-teardrops from above, tapering narrower as you move toward the back of the car.

They don't taper fully obviously because then you'd end up with the hugely long vehicle I mentioned, so all manufacturers cut off the body at some point, which is the next best thing.

As for ground effects, that can be minimised by a flat floor. Or by using a front splitter and side skirts to channel as much air as possible away from under the car, but that also increases drag. If you can't minimise the air going under the car, you want to make sure it slips under with as little effort as possible. Turbulence causes lift, which is why race cars have perfectly flat floors.

Here's a diagram of the effect half-teardrop I'm describing.

(Alternatively, think the experiment they did on Mythbusters with the motorcycle)

You can apply that half-teardrop to any car, and you'll see that things like Prii, Insights, CR-Zs, old CR-Xs, the Mazda MX3 and a bunch of other cars conform to it and I suspect, looking at the roofline, the Toyota 86 too, until the base of the rear screen at least.

The only major difference is the aforementioned lateral tapering of the more economy-orientated cars.

As for land speed record cars, their elongated length is very little to do with aerodynamics, funnily enough. Obviously they have to slip through the air as efficiently as possible, but they also need to be suitably long for stability at massive speeds (or even moderately high ones, like some of the "specials" at Bonneville speed week and the like) and the ideal teardrop shape would result in a vehicle that's too short. Not to mention the fact that the huge jet engines they use wouldn't fit.

 

[niky]

Was about to reply myself. Rocket and Jet powered Land Speed Record cars have an atrocious coefficient of drag. (The Space Shuttle is even worse, though...) Others are compromised by the need to maintain the silhouette of the production model.

Where engine size and body class are not an issue, such as with human powered vehicles, you see a more classic teardrop shape... Especially as they can run very thin wheels and don't need to bother with ground effects or lift-canceling downforce.

 

[Keef]

You should probably clarify that the aspect ratio of the tear drop shape is very important to its efficiency. Long, thin teardrops are much more capable of achieving high speeds than relatively short, fat ones like that Audi. Of course, the Audi's teardrop is probably 30 feet long or more, but a fighter jet's is considerably longer.

There's also the Area Rule to pay attention to. You can make your airplane with a basic teardrop shape, but don't attach the wings at the fattest part, and definitely don't make them straight. That's a good way to ensure you won't be going anywhere very fast.

 

[Zenith]

You should probably clarify that the aspect ratio of the tear drop shape is very important to its efficiency. Long, thin teardrops are much more capable of achieving high speeds than relatively short, fat ones like that Audi. Of course, the Audi's teardrop is probably 30 feet long or more, but a fighter jet's is considerably longer.

Agreed. Length is hugely important. The ratio needed of length:vertical drop is 6:1 to prevent flow separation and a turbulent wake. Not practical for a car.

Also. Drag coefficient is not a measure of how aerodynamic the car is, as some people seem to believe. A drag coefficient is exactly that; a coefficient. It stays the same no matter the frontal area. A tiny Corvette model will have the same Cd as the full size car.

 

[homeforsummer]

Agreed. Length is hugely important. The ratio needed of length:vertical drop is 6:1 to prevent flow separation and a turbulent wake. Not practical for a car.

Which is why many cars are given a kamm tail. In the absence of the length to keep laminar flow, the next best thing is to cut the length sharply to minimise turbulence from flow separation.

Also. Drag coefficient is not a measure of how aerodynamic the car is, as some people seem to believe. A drag coefficient is exactly that; a coefficient. It stays the same no matter the frontal area. A tiny Corvette model will have the same Cd as the full size car.

Agreed. CdA (drag x area) is the more important figure, as it takes into account frontal area too. However, Cd is still a good measure of the relative aerodynamic efficiency of a particular shape. Given the same frontal area, a car of Cd 0.26 will be more aerodynamically efficient than one of 0.30.

My first post is a little misleading as it looks like I'm being ignorant of those two things, but I did actually point them both out in the Toyobaru thread! Just thought it better to continue the discussion in its own thread.

 

[Jim Prower]

Something interesting about raindrops is that they're not...teardrop shaped. They actually look like this...

and given this, they're not really all that aerodynamic. otherwise, rain would hurt much, MUCH more!

 

[Exorcet]

What you do with everything forward of the windscreen is actually largely academic compared to what you do from the roof backwards

I think it's about the same. Auto design research is less prevalent than aircraft, but a number of studies have been done on the flow structure and the effects of different parts of the car.

Wolf-Heinrich Hucho has pretty much summarized everything in his book, Aerodynamics of Road Vehicles. I need to finish it at some point.

As for ground effects, that can be minimised by a flat floor.

This increases the ground effect. If you don't want high speed air under the car that you don't want to deal with, use an airdam and create a mess of low speed vortices under the car. It's not as efficient as controlling the flow, but it's better than doing nothing and is much simpler and cheaper.

If you go the flat bottom route you might increase drag on the wheels which requires more money/design to fix. You also need a rear diffuser or that high speed air just goes off on its merry way out the back and you get a drag penalty.

Or by using a front splitter and side skirts to channel as much air as possible away from under the car, but that also increases drag. If you can't minimise the air going under the car, you want to make sure it slips under with as little effort as possible. Turbulence causes lift, which is why race cars have perfectly flat floors.

Air under the car isn't a bad thing, and it can take some thinking to decide where you want to route it. The less that goes under, the more that goes above. This increases stagnation (pressure) at high pressure areas. You also get stronger pressure recovery at the back of the car, but it's not necessarily a 1:1 trade.

The bigger concern for drag is not where the air goes, but what it does. Side skirts and whatnot are more about stopping vortex formation, because the vortex is going to rob you of potential pressure recovery behind the car.

You can apply that half-teardrop to any car, and you'll see that things like Prii, Insights, CR-Zs, old CR-Xs, the Mazda MX3 and a bunch of other cars conform to it and I suspect, looking at the roofline, the Toyota 86 too, until the base of the rear screen at least.

The teardrop shape is good and it can be considered an overall strategy for car aero, but it's not really a preferred shape, and even if the side profile was "perfect" 3D effects come into play and make the aero more complex than for a hypothetical radially symmetric raindrop.

In particular you have a few areas of noticeable vortex formation on nearly every car: cowl, a/b/c pillar, rear edges, mirrors. These can take a significant amount of R&D time as there effects on drag can be high. Surprisingly, cars are fairly tolerant of pretty squared off front ends. A boxy shape with rounded corners can produce a fairly good CD value compared to a really smooth curved nose. Sometimes, curvature is actually made more abrupt to create suction peaks at the front of the car to reduce drag.

As for land speed record cars, their elongated length is very little to do with aerodynamics, funnily enough.

Air changes going from your average car to these. They are in the compressible flow region, which is something you'll only really get to with a Veyron or SSC. Further up, you go transonic (which is where the aera rules applies, it does nothing at lower speeds) where things really start to change.

But the other thing with jetcars is, the jet is part of the car. Engines that work on air create negative drag - you can't distinguish between the natural drag of a body and the momentum imparted to the air by the engine. The jet is basically creating the teardrop shape effect just by existing (and there is a little help from the air outside the jet actually following the tear drop.

Obviously they have to slip through the air as efficiently as possible, but they also need to be suitably long for stability at massive speeds (or even moderately high ones, like some of the "specials" at Bonneville speed week and the like) and the ideal teardrop shape would result in a vehicle that's too short. Not to mention the fact that the huge jet engines they use wouldn't fit.

For the really fast ones, a tear drop is draggy because it will develop a shockwave. Other than that I don't follow, the tear drop shape should still work. It wouldn't be too short, you could just make it longer. Also as above, since the section with the jet engine is kind of taken out of the whole drag equation, it's OK if it sticks out a little.