Aerodynamics

[celter]

First of all, I am not a physicist, and no expert on aerodynamics, so bear over with me if this is a stupid question: Is the nose-cone design on the Model S the best solution to reduce the drag coefficient? The nose cone has a smooth surface. While this may appear to be optimal with regards to the drag coefficient, I am not completely convinced.

Take the golfball example. I found this explanation online:

“The surface of a golfball is made of small dimples. These dimples acts as “turbulators” and induce turbulence in the layer of air next to the ball. In some situations, a turbulent boundary layer reduces drag.
There are two types of flow around an object: laminar and turbulent. Laminar flow has less drag, but it is also prone to a phenomenon called "separation." Once separation of a laminar boundary layer occurs, drag rises dramatically because of eddies that form in the gap. Turbulent flow has more drag initially but also better adhesion, and therefore is less prone to separation. Therefore, if the shape of an object is such that separation occurs easily, it is better to turbulate the boundary layer (at the slight cost of increased drag) in order to increase adhesion and reduce eddies (which means a significant reduction in drag). Dimples on golf balls turbulate the boundary layer.”

We know that a dimple-less golf ball will only travel half the distance of a normal golf ball. Could the same aerodynamic principle be used by designing the Model S nose cone with dimples? Other car-manufactorers have never been able to do this because the air intake at this place makes this impossible. But the Model S could. ( And it might look pretty cool )

I am probably completely wrong, but explain me why.

 

[ElSupreme]

First of all, I am not a physicist, and no expert on aerodynamics, so bear over with me if this is a stupid question: Is the nose-cone design on the Model S the best solution to reduce the drag coefficient? The nose cone has a smooth surface. While this may appear to be optimal with regards to the drag coefficient, I am not completely convinced.

Take the golfball example. I found this explanation online:

“The surface of a golfball is made of small dimples. These dimples acts as “turbulators” and induce turbulence in the layer of air next to the ball. In some situations, a turbulent boundary layer reduces drag.
There are two types of flow around an object: laminar and turbulent. Laminar flow has less drag, but it is also prone to a phenomenon called "separation." Once separation of a laminar boundary layer occurs, drag rises dramatically because of eddies that form in the gap. Turbulent flow has more drag initially but also better adhesion, and therefore is less prone to separation. Therefore, if the shape of an object is such that separation occurs easily, it is better to turbulate the boundary layer (at the slight cost of increased drag) in order to increase adhesion and reduce eddies (which means a significant reduction in drag). Dimples on golf balls turbulate the boundary layer.”

We know that a dimple-less golf ball will only travel half the distance of a normal golf ball. Could the same aerodynamic principle be used by designing the Model S nose cone with dimples? Other car-manufactorers have never been able to do this because the air intake at this place makes this impossible. But the Model S could. ( And it might look pretty cool )

I am probably completely wrong, but explain me why.

The dimples of the golf ball reduce the size of the low pressure wake (that essentially pulls the ball backwards) by creating turbulent air. It also slightly increases the skin friction but the benefits of a smaller wake are much greater. It is most useful when an object starts to taper down in size. Dimples are great for an object that do not have a stable orientation (ball spinning all sorts of ways) but for fixed objects like cars you can introduce subtle features that do the same thing.

But putting dimples on the front edge of a body is in general not the correct place.

 

[NigelM]

 

[Johan]

Mythbusters: Dimpled Car MiniMyth : Video : Discovery Channel

Update: Nigel, you beat me to it!

 

[celter]

Well, that “Mythbusters” video wasn’t exactly like my idea of a nose cone with small dimples J

Wouldn’t dimples on the nose cone create a thin layer of turbulent air, like an air cushion, in front of the car. And wouldn’t the friction between the air and this air cushion be less than between the air and the car?

I have also heard that racing boats have less drag in water if the hull isn’t polished. Is that a fact?

 

[ElSupreme]

Well, that “Mythbusters” video wasn’t exactly like my idea of a nose cone with small dimples J

Wouldn’t dimples on the nose cone create a thin layer of turbulent air, like an air cushion, in front of the car. And wouldn’t the friction between the air and this air cushion be less than between the air and the car?

I have also heard that racing boats have less drag in water if the hull isn’t polished. Is that a fact?

So fluids, and thus aerodynamics are hard. A lot of things are counter intuitive. And when you start dealing with compressible fluids (air, not water) it gets even harder. And that stupid Mythbusters episode makes things worse.

Smooth fluid flow is called laminar flow. The fluid near your surface is moving very slowly and as you move away from the surface the fluid starts moving more rapidly. You end up with very little skin friction in this setup, but you create a wake. Turbulent flow is when you have high speed fluid (and some random directionality also) moving right next to your surface. You have higher skin friction but you create less wake. Laminar flow also is very thick (sometimes inches) in the air that is not moving at speed, and effectively increases your surface's frontal area.

A lot of the time it is better to reduce your wake with increased skin friction as it is a major product of drag. It really depends on the fluid (air, water, maple syrup, etc.), the shape of your object, the size of your object, and the surrounding fluid movements (crosswind, headwind, etc.). The dimples on a golf ball greatly reduce the wake behind the ball, but increase skin friction all around the ball. In a perfect world you would have laminar flow in the front of the ball and turbulent around the belt (widest point) and then transition back to laminar flow behind the ball. This can be accomplished with a 'belt', or channel at the widest point (really slightly before the widest point). But the problem with golf balls is that they spin all around so you put dimples everywhere.

In general you don't gain a lot with thin layers of turbulent flow versus the slightly wider layers of laminar flow when dealing with air. It generally isn't dense enough to matter. Water on the other hand is a very different ball game. You also lose advantage as your surface gets longer. As the increased skin friction is constant over the whole length, your thick layer does'n grow as length is increased.

And there is a reason why both Jamie and Adam were completely baffled by the dimples on the car experiment. Because it shouldn't have had better gas mileage. If you did that same test on a Model S I bet you would get complete opposite results. That Taurus just wasn't optimized for aerodynamics from the start. A small rib of clay in a few key spots probably would have done even better. All that experiment showed was that further testing was needed to determine how to improve aerodynamics. Putting dimples everywhere isn't the right thing to do. Not to mention a couple of degrees of temperature could produce similar results. (I know they are more scientifically rigorous than they put on the show but there are so many variables driving a car outside)

And as for your racing boats I bet you will see most of their hulls polished and smooth. I bet you get some rough areas near where the surface tapers or changes directions, but most of the surfaces should still be smooth (or at least really really close and coated in a 'slippery' material). In water you use a rough surface to reduce the size of your boundary layer at the cost of higher skin friciton. Those super speedos used in the 2008 Olympics, which are now banned, use this principal (along with some super slippery materials that reduce the actual skin friciton). But the longer your surface the less advantage you get. There is a reason why airplanes and boats have been super smooth for a long time, it works. Now with greater material science we can get super small features, and slippery surfaces that allow this give take to move more towards the 'rough' skin area.

 

[richkae]

(I know they are more scientifically rigorous than they put on the show but there are so many variables driving a car outside)

I sure hope so. I love the show, but some of their "science" is painful to watch sometimes.
The show they did about driving cars with the windows down vs. air conditioning was terrible.

 

[Alpha]

I was reading about the Sparrow EV recently and they actually put dimples on the sides- they claimed that it did give better aerodynamics... Dimples on the very front seems like it would act more as a dam to capture air (and thus increase drag.) The sides seems more reasonable if there is anything to it -- That's just my intuiton though and could be way off...

Here are some good pics of the dimples on the Corbin Sparrow...
Endless-sphere.com View topic - CORBIN Sparrow Electric Vehicle

 

[celter]

@elsupreme: thank you for this lesson in aerodynamics. You wrote: "In general you don't gain a lot with thin layers of turbulent flow versus the slightly wider layers of laminar flow when dealing with air. It generally isn't dense enough to matter." That means that you gain something, but it wouldn´t be noticeable on coefficient of drag on a car? Another question: How much laminar flow is there really around a car? I know there´s virtually no laminar flow around a boat hull.

 

[Alpha]

Speaking of aerodynamics, If you haven't seen this yet, it's pretty wild!

"The pizza pan method"

Aero wheels - Page 39

I do hope Tesla brings back the aero wheels as an option on the Model S in some capacity (and that they maybe look a little more cool than the pizza pan ones :tongue

 

[ElSupreme]

@elsupreme: thank you for this lesson in aerodynamics. You wrote: "In general you don't gain a lot with thin layers of turbulent flow versus the slightly wider layers of laminar flow when dealing with air. It generally isn't dense enough to matter." That means that you gain something, but it wouldn´t be noticeable on coefficient of drag on a car? Another question: How much laminar flow is there really around a car? I know there´s virtually no laminar flow around a boat hull.

A sailing boat (sorry I always think sailboat when someone says boat, or submarine) hull is almost 100% laminar flow around the fully submerged part (unless there are aqua dynamic features to prevent it). A power boat probably not, I don't know.

When I say in general I mean most of the time you lose efficiency when inducing turbulent flow for the purposes of reducing boundary layer size (not for keeping adhesion on a tapering body). The skin friction typically is greater than the slightly larger effective frontal area. But with 'slippery' surfaces or short bodies it isn't always true (fluids are HARD to deal with).

In reality there are 2 ways to reduce 'wake' that I was talking about. If you push your hand through water two things happen.

The water on top and bottom of your hand move, such that you move more effective surface area of water than the size of you hand. Reducing this extra area is typically not worth the extra skin drag that comes with inducing turbulent flow. Really only recently has material science (size of tiny features and 'stickness' and better computer models has this even been possible). Think of this as wake off to the side of of a boat.

The second thing that happens is there is a low pressure wake behind your hand (you will see the water level drop here displaying lower pressure). This is where dimples on a golf ball come into play. Turbulent air sticks to the surface better than laminar flow. Around the back of the ball (tapering body) turbulent air follows the ball before separating. This effectively reduces the size of the low pressure area behind the ball. This type of flow GREATLY reduces drag. Think of this as wake behind the boat.

These two forces are slightly different. The first is due to the mass of fluid you have to push out of the way. The wider the boundary layer the more fluid you have to move. The second is due to low pressure behind an object and high pressure in front of the object. This produces a direct force on the object and is normally the greatest part of 'drag'. The last part of drag which sometimes is greater than the first, sometimes less than the first, is due to actual skin friction on the object. And decreasing this typically leads to increasing the first.

But for fixed shapes small features rather than pseudo-random dimples works better. But requires actual testing. Dimples work if you know the general area where you want to induce turbulent flow to keep 'adhesion' but are too poor/lazy/rushed to do real testing.

Sorry this stuff is sort of hard to cut down to a single post.

 

[Alpha]

...
Sorry this stuff is sort of hard to cut down to a single post.

Wow! You really know a lot about this stuff. What kind of engineer are you anyway? Sounds like maybe directly in that field?
So it sounds like the bottom line is that dimples can really make a difference, they just have to be applied very intelligently and in the right places...
Why don't we see dimples on a lot more cars, or on every car? Is it the ugliness factor?

 

[ElSupreme]

Wow! You really know a lot about this stuff. What kind of engineer are you anyway? Sounds like maybe directly in that field?
So it sounds like the bottom line is that dimples can really make a difference, they just have to be applied very intelligently and in the right places...
Why don't we see dimples on a lot more cars, or on every car? Is it the ugliness factor?

I am a degreed mechanical engineer and deal with pipe flow (the second easiest type of fluids, the easiest is flow over an infinite flat plate) all the time. And never really had to deal with compressible flow. And when doing real calculations you don't do the actual fluids problem. You use a simplified model which is still hard to deal with. I was (sort of still am) an avid cyclist and so looked at a lot of practical aerodynamic problems.

Fluids and Power Electronics were two of my favorite classes in college. Those and my physics classes were great.

As to why we don't see dimples there are really 3 reasons. First, is people wont buy cars with weird dimples on them. Second, is that non-dimple features can provide the same effects. Say the gap between body panels. Little lips and valleys can do that too. And third, this level of aerodynamics is hard/time consuming/expensive/has a small overall effect that car manufacturers haven't spent much time on it. The 'look' of a car will sell way more units than a single percent fuel economy improvement. And that even on cars a small crosswind will completely change the aerodynamics of the car that often times your 'improvement' isn't really all that effective.

 

[Alpha]

Yeah it does look quite ugly, especially if taken to extremes, like they did on Mythbusters - so if you can get the same effect without the dimples that makes sense. On the other hand, some cars are already extremely ugly, (the leaf and prius come to mind) and people who buy those might be sold on the dimples idea too if they thought they would get more mpg for it... Heck, maybe covering the car in feathers might make a difference - seems to work for birds pretty well OK, now that's a complete joke. By the way I love your "Uh-oh." avatar - makes me think of Jurassic park - is that a velociraptor? (recently saw that old movie again on Blu-ray... so I have that on the brain.)

 

[celter]

@elsupreme: I am always impressed when someone has extended knowledge of something I don´t know so much about. Thank you !

 

[Zextraterrestrial]

I will agree, Transport Phenomena, Fluids and Electronics were my favorite classes (besides economic botany - awesome class!)

So based on that I feel a little 'retarded' after trying to mount my airspeed pitot tube to the front of my car. with the pressure wave, I seem to get really low #'s and haven't been able to get it in a good location. any ideas? I am trying to keep it from looking too crappy with it. I started off with it a little closer to the center but higher. Now I have mounted it to a CF tube to try and extend it out further but still only saw 26-30 mph when I hit 60 today.

should I still be seeing the boundary layer here? probably, huh?
I had it mounted to a truck roofrack and was seeing 140 mph(going ~80), which I expected due to the location being just above he windshield, so I think it is working ok

 

[JohnQ]

Where's your static source? Placement of the static source is just as important as the pitot tube.

 

[Zextraterrestrial]

Where's your static source? Placement of the static source is just as important as the pitot tube.

The pitot has a static on the side ~ 1" back from the tip. It was meant for RC application.

I probably need to be almost 1' out from the nose?

The altimeter works pretty good

 

[ElSupreme]

I have a feeling you are getting some ground effect placing the pitot tube so close to the ground. High pressure is going to build up near the ground as there is very limited space (espically when the suspension lowers) for it to go.

A quick glance at my car I am thinking the side mirrors are probably the cleanest place to put them. But I think the big ridges on the hood are there to 'shield' the side mirrors from wind, much like the weird LEAF headlamp housings.

The only other place that looks clean to me is under the frunk lid, or the lip above the headlamps. Neither place will really hide the pitot tube well though. Also are you sure you have enough extension in front of your vehicle? I remember having an AE friend in college that commented that pitot tubes were really about twice as long as they look, or you would imagine on airplanes. And they are much less flat on front (less high pressure area) and don't have any ground effect to speak of. Going further out would only help regardless of location.

The other option is to go off the side of your car, but potential for damage is probably pretty high.

At the hood I don't think there is very much side displacement of air at all. It was raining lightly on my way to work this morning, and I had a fresh RainX application, so I wasn't using the wipers. The water trails were only about 5deg from vertical. So I don't think left to right placement would be a big issue.

EDIT: I am thinking you need to be ~8-12" off the front of the car if you are doing the hood. And longer the closer to the ground you get.

 

[JohnQ]

That could be part of your problem. I presume you have the sensor unit in the passenger compartment. While it won't read quite right if you just vent the static port to the interior it might be worth trying to isolate the problem. Meaning, is your pitot source sufficiently out of the boundary layer or is your static source reading too high an ambient pressure. It's a little bit of a black art (and mostly trial and error) to get the static location right on a full size aircraft.