A question on vehicle aerodynamics

[xxChrisxx]

Its never going to win a fuel endurance race though is it...

0.3 may be fairly good for a road going car but doesn't cut the mustard for ultra low drag racing.

 

[famousken]

Its never going to win a fuel endurance race though is it...

0.3 may be fairly good for a road going car but doesn't cut the mustard for ultra low drag racing.

But it makes up for it in cool points ;)

 

[xxChrisxx]

Very true.

 

[Cyrus]

koenigsegg ccx isn't teardrop shaped and it has a drag coefficent of 0.3

That doesn't mean anything...it's the wetted area that matters if you want to compare two cars. I could define the reference area a mile wide for a mini van and have a drag coefficient of 0.000000000000000000000000007.

 

[famousken]

That doesn't mean anything...it's the wetted area that matters if you want to compare two cars. I could define the reference area a mile wide for a mini van and have a drag coefficient of 0.000000000000000000000000007.

The wetted area is just the frontal area of the car isn't it? So to figure drag you would use
Cd X A^2?

Here is a link to a car that VW is working on, most definatly teardrop shaped
http://www.ultimatecarpage.com/car/1316/Volkswagen-1-Litre.html

 

[xxChrisxx]

That doesn't mean anything...it's the wetted area that matters if you want to compare two cars. I could define the reference area a mile wide for a mini van and have a drag coefficient of 0.000000000000000000000000007.

Why the frig would you do that? You COULD define the area as that but you'd be utterly stupid to as it tells you nothing useful. Thats like saying I'm going to work out the Reynods number for a flow over a wing and define the characterisitc length as the diameter of the earth. Also say you've calculated your Cd empirically fby finding the drag force and then calcualting Cd.

And of course it means something Cd is how you compare drag between the shape of cars to take out the annoying variables like area and speed, its nondimensional for a bloody reason.

He was making a statement that the drag coeffient for a CCX is 0.3... which it is. Just like the Cd for a Lotus 7 is 0.7. And just like the Cd for an F1 car varied according to how much downforce its set to. They are all calcualted in a standard way. So there is bugger all wrong with his statement.

Can you just take a step back and stop yourself from making stupid posts simply for something to do. I wouldn't mind you being argumentative if you contibuted something remotely useful to furthering the discussion.
More to the point, what on Earth was the point of your post, how was it relevent?

EDIT: And don't bother waving your hands and trying to justify yourself by saying CdA is more important than Cd, because it isnt. If you are going for that mentality you may as well just throw everything in and talk about Fd. He was referring to the shape only.

 

[Cyrus]

Why the frig would you do that? You COULD define the area as that but you'd be utterly stupid to as it tells you nothing useful. Thats like saying I'm going to work out the Reynods number for a flow over a wing and define the characterisitc length as the diameter of the earth. Also say you've calculated your Cd empirically fby finding the drag force and then calcualting Cd.

And of course it means something Cd is how you compare drag between the shape of cars to take out the annoying variables like area and speed, its nondimensional for a bloody reason.

Please go back and re-read my post. The point is that if the cars do not have the same frontal area it is meaningless to directly compare the drag coefficient.

He was making a statement that the drag coeffient for a CCX is 0.3... which it is. Just like the Cd for a Lotus 7 is 0.7. And just like the Cd for an F1 car varied according to how much downforce its set to. They are all calcualted in a standard way. So there is bugger all wrong with his statement.

Provided they have the same reference area, then yes. Is there a standard refernce area used in automotive aerodynamics?

Can you just take a step back and stop yourself from making stupid posts simply for something to do. I wouldn't mind you being argumentative if you contibuted something remotely useful to furthering the discussion.
More to the point, what on Earth was the point of your post, how was it relevent?

EDIT: And don't bother waving your hands and trying to justify yourself by saying CdA is more important than Cd, because it isnt. If you are going for that mentality you may as well just throw everything in and talk about Fd. He was referring to the shape only.

I have reported this portion of your post.

 

[Cyrus]

PS, do you notice how the tabulated values here are wetted parameters?

http://en.wikipedia.org/wiki/Automobile_drag_coefficient

I.e., $$C_D A$$.

While designers pay attention to the overall shape of the automobile, they also bear in mind that reducing the frontal area of the shape helps reduce the drag. The combination of drag coefficient and area is CdA (or CxA), a multiplication of the Cd value by the area.

In aerodynamics, the product of some reference area (such as cross-sectional area, total surface area, or similar) and the drag coefficient is called drag area. In 2003, Car and Driver adapted this metric and adopted it as a more intuitive way to compare the aerodynamic efficiency of various automobiles. Average full-size passenger cars have a drag area of roughly 8.5 ft² (.79 m²). Reported drag area ranges from the 1999 Honda Insight at 5.1 ft² (.47 m²) to the 2003 Hummer H2 at 26.3 ft² (2.44 m²). The drag area of a bicycle is also in the range of 6.5-7.5 ft².[16]

Also, don't put words in my mouth:

EDIT: And don't bother waving your hands and trying to justify yourself by saying CdA is more important than Cd, because it isnt. If you are going for that mentality you may as well just throw everything in and talk about Fd. He was referring to the shape only.

 

[Cyrus]

Please read page 307, notce anything special about the y-axis?

http://books.google.com/books?id=nM...sult&resnum=4#v=onepage&q=wetted area&f=false

This is why you can't add the contributions of drag in coefficient form from each part of the car unless they all have the same reference area. Each car is going to have a different reference area, which is typically the frontal area of the car. Hence, why you can't simply compare one Cd value to another and have it be meaningful. This is why saying things like 'car X has a Cd of such and such" is as meaningless as boasting about your car having 900HP, and failing to mention it weighs as much as a big rig.

 

[Cyrus]

And of course it means something Cd is how you compare drag between the shape of cars to take out the annoying variables like area and speed, its nondimensional for a bloody reason.

No, this is where your problem lies. Cd is not defined the way it is to "take out the annoying variables like area and speed". Cd is a nondimensional parameter for a bloody reason you failed to mention, similitude. People don't put full size cars inside wind tunnels because its very, very expensive. Instead, they put scale models. In order to put a scale model, you must have dynamic similitude, which occurs through the buckingham-Pi theorem of nondimensionalization. You can then SCALE UP the value of Cd from the model to the full size production car by scaling the REFERENCE AREA.

This does not mean Cd is a valid metric for comparison of cars. The wetted area is a better metric.

EDIT: And don't bother waving your hands and trying to justify yourself by saying CdA is more important than Cd, because it isnt. If you are going for that mentality you may as well just throw everything in and talk about Fd. He was referring to the shape only.

My paragraph above explains this in detail.

 

[xxChrisxx]

Hmm ok I've decided I am going to have to be careful trying to explain the reasoning on with this one as you do have me worried I am wrong. It's been a while since I've had to do any aero work.

As we arent comparing cars directly only the shapes. And although you need the same reference area to directly compare shapes. Cd gives a better indication of how good the shape is when you know nothing about the car than CdA. CdA is a better metric if you want to compare cars themselves and not just the geometry.

You come from an aerospace background don't you?. As cars are bluff bodies comparison methods differ slightly, wetted area doesn't really mean much if the flow has separated and the drag is coming from the pressure. We always use frontal area for a car. (unless you are using wetted area to mean frontal area)

The reason for this is that CdA can give misleading ideas when talking about bluff bodies rather than spteamlined. A low Cd car with a high CSA can give a higher CdA when compared to a high Cd car with a low CSA.

For example:

Ok we have a small family hatchback shaped car with a drag coefficient of 0.3 with a frontal area of 0.5m^2
We have much larger modern GT car with a drag coefficent of 0.25. With a frontal area of 0.7m^2.

The CdA for both is:
Hatch: 0.15 m^2
Sports: 0.175 m^2

Now if we knew nothing about the Cd of these cars, but wanted to pick one of the shapes we could have to conclude the hatchback is better on data availalbe when the GT car produces more drag simply becuase it is bigger doesn't lead to a good comparison as if the Gt car were the same size it would give a lower CdA. Basically the shape of the GT car goes through the air better than the hatchback.

As you stated Cd is there for simlitude, which allows scale models to be used. Imagine two cars of different CSA are just scales models of the car that you want.

We take the CdA, divide by the CSA of the acutal car to get the Cd then multiply it by the new CSA of the car we are going to build. Why not just cut the step to comparing a similar 'wetted area' and just compare the Cd values. Obviously this only holds for similar conditions (same Re same Mach etc etc).

Now when you are dealing with streamlined bodies such as planes/wings. Wetted area is more useful becuase the flow stays attached and the majority of the drag comes from skin friction which is area dependent. Which is why I like the fact you used a helecopter aerodynamics book to prove a point about car aero. I know in general aero is aero, but the method about comparing cars always tends to use Cd becuase the flow seperates.

One textbook example I can think of that demonstrates this is you have a small diameter circe and a large aerofoil i think the characteristic length of the aerofiol was 10x that of the circe (i'll try to find this example from my books and fill in the details)
They both have the same drag force (so the same CdA as they were tested under the same conditions) but the streamlined aerofoil has the lower Cd.

Now its obvious that when you scale the aerofoil down to the same charactersic length as the circle it will have a much lower drag.I'd like to apologive for the tone of my last post, it was hastily typed. Dont get me wrong Cyrus I like you and your direct approach, but sometimes you piss me off with needlessly condecending or combative posts. I wouldn't mind if they were wrong and talking bollocks but the guys post was clearly anecdotal. I pointed out that it didnt matter and he used the phrase 'cool points', that more than anything else shows he wasnt tryin to make a serious point.

 

[Astronuc]

In aerodynamics, the product of some reference area (such as cross-sectional area, total surface area, or similar) and the drag coefficient is called drag area. In 2003, Car and Driver adapted this metric and adopted it as a more intuitive way to compare the aerodynamic efficiency of various automobiles.

Hmmm - back when I studied the fundamentals of fluid dynamics, we looked at both frontal area and surface area and we had to calculate the normal dynamic force and friction (shear) force. Then we'd have to consider flow separation, which put a static force on an object due to the differential pressure between front and back.

Of course, a car is designed for utility and operation/environment. The utilities of a race car, sports car and family passenger car are all different, and so are their environments and operation. The utility (function) is just one of the constraints with which a designer must deal.

And there is the coolness factor for sports car and family car. A race car just needs to be fast, reliable and safe at the higher speeds at which it performs, and the designer isn't constrained by the comfort or appeal of the driver.

When I looked at the ultimatecarpage site, the banner at the top had a picture of couple of the Gulf Porsche 917's from back in the early 70's. Fantastic car, but it was quite uncomfortable for the driver. The space for the driver was minimized to reduce frontal area. Visibility to the rear was a problem too, because external rear-view mirrors would project into the airflow and the rear window was small. The only time one needed it was when someone was trying to pass (not too often) or the driver was passing another car (most of the time).

http://www.khulsey.com/stockphotography/porsche_917k_le-mans.jpeg

 

[Cyrus]

Cyrus we arent comparing cars directly only the shapes. And although you need the same reference area to directly compare shapes. Cd gives a better indication of how good the shape is when you know nothing about the car than CdA. CdA is a better metric if you want to compare cars themselves and not just the geometry.

If you are going to compare cars on nothing but shape, then you need to make sure both cars have the same frontal area for a direct comparison of Cd. Otherwise, CdA is the only thing that make sense as a comparison.

You have also got to be really careful Cyrus, because you blatantly come from an aerospace background and are used to dealign with streamlined bodies. Car's as you probably konw are bluff bodies. This is why comparison methods differe slightly, wetted area doesn't really mean much if the flow has separated and the drag is coming from the pressure.

We always use frontal area for a car.

The problem is that the frontal area is not a constant for every car.

The reason for this is that CdA can give misleading ideas when talking about bluff bodies rather than spteamlined. A low Cd car with a high CSA can give a higher CdA when compared to a high Cd car with a low CSA.

I don't know what "CSA" stands for: I think you mean frontal area for CSA. We call this S in aerodynamic standards. If that's the case, then what you wrote in the second sentence is exactly my point. The influence of area cannot be ignored for exactly that reason. When comparing two cars a given dynamic pressure, q, I care about which one has a higher drag force because it will require more power. CdA tells me this, Cd does not. Cd only helps for similitude of scale models.

For example:

Ok we have a small family hatchback shaped car with a drag coefficient of 0.3 with a frontal area of 0.5m^2
We have much larger modern GT car with a drag coefficent of 0.25. With a frontal area of 0.7m^2.

The CdA for both is:
Hatch: 0.15 m^2
Sports: 0.175 m^2


Now if we knew nothing about the Cd of these cars, but wanted to pick one of the shapes we could have to conclude the hatchback is better when the GT car produces more drag simply becuase it is bigger doesn't lead to a good comparison as if the Gt car were the same size it would give a lower CdA. Basically the shape of the GT car goes through the air better than the hatchback.

No, it says that the Hatch does go through the air better at a given speed than the GT. The hatchback will absolutely require less power to stay at a fixed speed. If you want to compare the two cars at a purely geometric standpoint in terms of the aerodynamics, give them both the same refernence area. Then you can compare the Cd. In fact, this is what you are doing implicitly in your argument of "just compare Cd"!

As you stated Cd is there for simlitude, which allows scale models to be used. Imagine two cars of different CSA are just scales models of the car that you want.

We take the CdA, divide by the CSA of the acutal car to get the Cd then multiply it by the new CSA of the car we are going to build.

Eg above. We know that out frontal area of the car is going to be 0.3m^2 for example.

CAR 1:CdA 0.15
CAR 2:CdA 0.175

Ok so we divide by the current CSA and times by new CSA.

CAR 1: 0.15*(0.3/0.5) = 0.09
CAR 2: 0.175*(0.3/0.7) = 0.075

As the above is simply the Cd multiplied by the area we want. Why not just cut the step to comparing a similar 'wetted area' and just compare the Cd values. Obviously this only holds for similar conditions (same Re same Mach etc etc).

Because when you "cut the step" that is an assumption of having the same reference area value for both design shapes.

Cd is a funcion of Re and Ma and flow direction. Under the same conditions the Cd will be the same. So i'll repeat what I said above so you don't miss it. "Obviously this only holds for similar conditions (same Re same Mach etc etc)".

Yes, that needs to be held constant, but that isn't the issue here.

Of course when you scale both shaped down to the area you want the one with the lowe rvalue od Cd is the one that gives the lower drag. This is why at the start of designs, when you know the rough size of the car you want but want to focus totally on the shape its more useful to compare Cd values of other cars NOT CdA.

Exactly, becuase they both have the same area!

Now when you are dealing with streamlined bodies such as planes/wings. Wetted area is more useful becuase the flow stays attached and the majority of the drag comes from skin friction which is area dependent. Which is why I like the fact you used a helecopter aerodynamics book to prove a point about car aero. I know in general aero is aero, but the method about comparing cars always tends to use Cd becuase the flow seperates.

Attached or unattached flow doesn't matter, that is accounted for in the value of Cd. Helicopters are by no means streamlined, they are a highly unsteady aerodynamic environment due to the rotors. Any unsteadiness is captured by time varying aperiodic terms inside of Cd, i.e Cd = Cd(Re,Ma, time, and other explanitory variables).

I'll leave with this. For cars in the wind tunnel, you can cause them to have side slip angle and get a side force. What is interesting is that you can return the car back to centerline of the wind tunnel, but you will still measure a strong side force despite no yaw angle. A flow condition occurs where the car starts sheading a vortex that doesn't go away when the car gets back to center. Very counter-intuitive.

I will also note: I am glad the OP stated his car has a targed wetted area < 0.1 for the competition.

 

[famousken]

Quote:[People don't put full size cars inside wind tunnels because its very, very expensive. Instead, they put scale models. In order to put a scale model, you must have dynamic similitude, which occurs through the buckingham-Pi theorem of nondimensionalization. You can then SCALE UP the value of Cd from the model to the full size production car by scaling the REFERENCE AREA.]

People don't put full size cars in wind tunnels? They most certainly do! There are some things that just don's scale up well, what do you think the apparent air density is for a 1/16th scale model is compared to a full size car. Air is not going to behave the same way on a model as it will on a full size vehicle. vortexes may not appear on a smaller model, It is damn near impossible to account for air that may be flowing THROUGH the car in areas like the radiator, under bumpers, undercarriage, etc. Another reason they do it is to take measurements of road noise and drafts in the vehicle. Please do a little research before you make such a false statement, people are on this forum are trying to understand new concepts and they don't need to be misled. Here are some links I turned up in about 60s of googling.

http://www.gm.com/experience/autoshows-events/racing/technology/aerodynamics/index.jsp
http://exoticcars.about.com/od/guidedtours/ig/Mercedes-Benz-SLR-McLaren/SLR-McLaren-wind-tunnel.htm
http://trucks.about.com/od/2007fordtrucks/ig/2009-Ford-F-150-Truck-Pictures/09-Ford-F-150-Wind-Tunnel-Test.htm
http://www.carbodydesign.com/gallery/2009/05/14-volkswagen-polo/22/

 

[Cyrus]

Quote:[People don't put full size cars inside wind tunnels because its very, very expensive. Instead, they put scale models. In order to put a scale model, you must have dynamic similitude, which occurs through the buckingham-Pi theorem of nondimensionalization. You can then SCALE UP the value of Cd from the model to the full size production car by scaling the REFERENCE AREA.]

People don't put full size cars in wind tunnels? They most certainly do! There are some things that just don's scale up well, what do you think the apparent air density is for a 1/16th scale model is compared to a full size car. Air is not going to behave the same way on a model as it will on a full size vehicle. vortexes may not appear on a smaller model, It is damn near impossible to account for air that may be flowing THROUGH the car in areas like the radiator, under bumpers, undercarriage, etc. Another reason they do it is to take measurements of road noise and drafts in the vehicle. Please do a little research before you make such a false statement, people are on this forum are trying to understand new concepts and they don't need to be misled. Here are some links I turned up in about 60s of googling.

http://www.gm.com/experience/autoshows-events/racing/technology/aerodynamics/index.jsp
http://exoticcars.about.com/od/guidedtours/ig/Mercedes-Benz-SLR-McLaren/SLR-McLaren-wind-tunnel.htm
http://trucks.about.com/od/2007fordtrucks/ig/2009-Ford-F-150-Truck-Pictures/09-Ford-F-150-Wind-Tunnel-Test.htm
http://www.carbodydesign.com/gallery/2009/05/14-volkswagen-polo/22/

I should have said, don't always put a full size car in a wind tunnel, that was a bit careless on my part. Don't lecture me on air density about scale cars when I was in a wind tunnel doing testing and saw a scale car sitting there from previous testing. It was a ford tarus. In fact, back in the 90s GM did so much testing they had their own private section of the tunnel. Road noise is from the RMS signal of the vibrations on the road. That's not something a wind tunnel tells you.

 

[xxChrisxx]

Ken not to bash you mate but you are wrong about the air not acting the same way.

If the people just put a scale model in a flow without changing it then, yes it wouldn't behave the same way.

In scale wind tunnel tests they use simlitude, (look up reynolds and mach numbers) to ensure that it does act in the same way as on a full sized vehichle.

This is the reason why people can use water tunnels to test (air)flow over cars.

Google doesn't give you all the answers I am afraid.

 

[famousken]

In model testing though; doesn't the air appear much more dense than it would with a full size vehicle? I know you could scale up the drag, but it seems to me that you would be missing out on quite a bit of important info by just using a model because there is a possibility that things like turbulance simply won't form on a smaller model because of the much greater apparent density of the air. That's just my reasoning though, I could be wrong I am not an expert.

 

[xxChrisxx]

Like I said, all that matters is that the nondimensional coefficients are the same. If thet is true the fluid will act in exactly the same way. All the information is there.

Now obviously sometihngs may be missed purely becuase of the size. All the phenomena are there, but if you are interested in a specific thing such as a vortex (that may be too small to capture by probes/smoke+camera) they sometimes run overscale models. So the model is X times larger than the actial thing.

 

[Cyrus]

In model testing though; doesn't the air appear much more dense than it would with a full size vehicle? I know you could scale up the drag, but it seems to me that you would be missing out on quite a bit of important info by just using a model because there is a possibility that things like turbulance simply won't form on a smaller model because of the much greater apparent density of the air. That's just my reasoning though, I could be wrong I am not an expert.

A few things need to be cleared up here.

[1] You were absolutely correct about full size testing. Some points though: the characterization of air through the engine scoops, moon roof, etc are very much final stage design items. It is WAYYYYYYYYYYYY to expensive to keep putting in new full size cars with different designs on them for something like engine intake scoops during the initial design of the car. The small models (and now CFD) are used for major body changes in the car. Once a more-or-less finalized design is reached a prototype is made. This full size prototype is VERY EXPENSIVE. That's why people use models first. Then the kinks are worked out in the full size tunnels you linked to earlier.

[2] The use of a small car is possible through similitude, which means (in the case of wind tunnels that have air as a working fluid), you need to make a scale model of the car. You can get away with this (pretty well) because the Reynolds dependency is relatively flat over a large range. It won't be 99% right, but it won't be very horrible. It will be pretty darn close. But that doesn't matter for a car! You can always put it in the full size wind tunnel when you make the actual prototype! Notice how you CANT do the same thing for an airplane though, which makes it much more of a challenge in terms of interpreting the data from a scale model. Provided the Reynolds number stays similar, the transition points and boundary layer should stay the same. Thats the entire point of Reynolds dependencies in the flow.