Aerodynamic Streamlining Mini-Course
 

*With sharp leading edges separation is immediate,leading to high drag.
http://i1271.photobucket.com/albums/...titled-5-2.jpg
*With an 'optimum' nose,separation is eliminated and kinetic energy is conserved for downstream conversion to pressure rise
http://i1271.photobucket.com/albums/...titled-7-2.jpg
*With subtle softening of the leading edges,the same flow attachment is achieved as with the 'optimum' nose.
http://i1271.photobucket.com/albums/...titled-6-2.jpg
*With minimum softening,flow attachment is achieved as long as the roof,sides,and underfloor respect certain contours.
http://i1271.photobucket.com/albums/...titled-3-2.jpg
http://i1271.photobucket.com/albums/...titled-4-2.jpg
http://i1271.photobucket.com/albums/...ad2/Scans3.jpg
--------------------------------------------------------------------------
*The challenge for streamlining,is to reduce,or eliminate the low base pressure turbulent wake,who's eddies and turbulence cannot be converted to a pressure increase.
http://i1271.photobucket.com/albums/...ad2/Scans4.jpg
--------------------------------------------------------------------------
*Observation of the wake region itself suggests a flow contour which would support attached flow.
http://i1271.photobucket.com/albums/...ad2/Scans1.jpg
--------------------------------------------------------------------------
*With 'boat-tailing' we achieve the elongation necessary to reduce,or eliminate the separated wake.
http://i1271.photobucket.com/albums/...titled-2-2.jpg
http://i1271.photobucket.com/albums/...d2/06-2818.jpg
http://i1271.photobucket.com/albums/...Untitled18.jpg
http://i1271.photobucket.com/albums/...d2/06-1315.jpg
http://i1271.photobucket.com/albums/...d2/06-1317.jpg
http://i1271.photobucket.com/albums/...d2/06-1321.jpg
http://i1271.photobucket.com/albums/...d2/06-1323.jpg
http://i1271.photobucket.com/albums/...d2/06-1324.jpg
http://i1271.photobucket.com/albums/...ntitled1-2.jpg
*This is a 2D flow image,however you can see how the flow comes apart when the aft-body contour is too 'fast.'
http://i1271.photobucket.com/albums/...d2/06-2835.jpg
*The action is at the rear.This is prime real estate for drag reduction.
*Any automobile manufactured after the mid-1980s is a good candidate for the elongation and aft-body mods necessary for significant drag reduction.:)
*Here is the 'golf ball' effect of the turbulent boundary layer's ability to 'hold' the air against the aft-body,even though it would like to flow backwards,upstream,towards the low pressure region preceding the position of max camber as shown with the laminar boundary layer.The contour,between the 12:00 o'clock position and 4-seconds after,or 6:00 o'clock and 4-seconds before,is essentially the aerodynamic streamlining template.
http://i1271.photobucket.com/albums/...d2/06-2833.jpg
Here is the same thing happening underwater with a bowling ball.US NAVY researchers have glued sand to the forward stagnation area to induce an immediate transition to TBL,and as you can see,the separation point is moved back 4-seconds of clock sweep.
http://i1271.photobucket.com/albums/...d2/06-2834.jpg
*Any analog watch or clock face could be used to pattern a successful 'template.' Compare this to the 2014 VW XL1.
*Here is Mair's 1969 'Template' for a 'best' boat tail.He never goes beyond 22-degrees,and he takes awhile to get there.If you remove the constant-velocity portion of Mair's 'Gavre artillery projectile',simply combining the nose and tail,you get a pretty fine 'Template.' It's already in mirror-image ground reflection.Just cut away for ground clearance and add wheels.
*Estimating Mair's 'Template' as a simple car model,we'd have Cd 0.448 with no boat-tailing,Cd 0.284 at Mair's recommended truncation (the hashed vertical line),and Cd 0.166 with full tail.
http://i1271.photobucket.com/albums/...ad2/WAMair.jpg
Here is a drag breakdown provided by HOT ROD Magazine in 1962.The 'Form' drag component is what streamlining is about
http://i1271.photobucket.com/albums/...rohead2/50.jpg
Here is a aero power relationship table from Walter Korff in 1963,denoting the cube function of aero power requirements as a function of velocity,for a car of 25-sq-ft frontal area
http://i1271.photobucket.com/albums/...ntitled1-4.jpg
Here is a HOT ROD table depicting the relationship of Cd,velocity,as it pertains to aero hp load.
http://i1271.photobucket.com/albums/...ohead2/100.jpg
Here is R.G.S.White's drag table for a vehicle of 22-sq-ft frontal area
http://i1271.photobucket.com/albums/...ntitled2-4.jpg
Here is a comparative chart illustrating the low-drag nature of the streamlined body of revolution.All 4 bodies have identical drag.When messengers commented in the 1970s,that we could have large vehicles with less drag than some tiny econoboxes,this is what they were referring to.
http://i1271.photobucket.com/albums/...titled11-1.jpg
In this table,I've taken the 'golf ball' from above,and added aft-bodies of increasing radius to the forebody.In every case,flow always separates at 4-seconds seconds after 12:00 on top and 4-seconds before 6:00 on the bottom.At some radius,the separation point is lost at the longitudinal bifurcation line,leaving only the drag do to the boundary layer torus which cannot be recovered.This is the reason that streamline bodies cannot have a lower Cd than 0.04.It's all skin friction.
http://i1271.photobucket.com/albums/...ntitled9-2.jpg

I could use some tips and info on the finer aspects of aeromodding , Its easy to miss little details that matter.

I am curious of the cars side aerodynamics at the moment , I am assuming they should get the same attention as the roof , by adding a camber point like it does for the roof in the boat tail template ?

The last image in Phil's post deals with side taper - and yes I think the sides are similar to the top. But since they are adjacent to the bottom - and the ground, they are a bit different than the top. If they slope "too fast" (to use Phil's term) then they will pull air down from the top more than up from the bottom and change the air flow over the top, and possibly cause swirling at the back.

If the sides are not tapered enough, then the air lifts up to the top, causing a swirl in the other direction. (On the image of Phil's boattailed truck, you can see this slightly.)

I think the ideal is to have as even air pressure as possible all around the rear of the vehicle.

Aerodynamic Streamlining Mini-Course :thumbup:

You will be graded at the pump. :D

My thought is , the front wheel flares are the camber point , if I trim the rear fenders it would provide a smoother transition from front to rear.


The Beetle has these pronounced side skirts that could be exaggerated on , to give the car more side air stability , I think ?

I could use these $175 aftermarket Xenon side skirts help keep the side air cleaner ... its more of a guess at this point.

http://www.bodykits.com/productImage...4244031060.jpg

Opinions on if the aftermarket side skirts are a good addition or a waste of material / mpg ? for a vehicle modded with mpg in mind.

sides
 

Here is a plan-view of the 'template.'
http://i1271.photobucket.com/albums/...ad2/Bottom.jpg
It's based upon a symmetrical wing/strut section of minimum drag (approx.4:1 length/thickness).
http://i1271.photobucket.com/albums/...rohead2/-2.jpg
Here is the 1957 MG EX 181 land speed record car,with Cd 0.12.It has a similar plan-view profile for the rear.
http://i1271.photobucket.com/albums/...ntitled2-1.jpg
--------------------------------------------------------------------------
Since the side flow is closer to road speed,it does not have as much kinetic energy to impart into the turbulent boundary layer,it cannot 'carry' as fast a curvature as over the roof.Many low-drag cars have used this more relaxed profile and I believe that,short of having a wind tunnel to experiment with,that this profile is a sure-thing.

Two schools of thought:
Keep the air topside at all costs, which the body kit tries to do, vs the 'banana car' which is low front to back, but arched to allow crosswinds underneath.

Attachment 15298
Here it is in stock form

They are actually pretty beefy.

http://www.bodykits.com/productImage...4244031060.jpg

Where the stock runningboard leaves 2-3" of fender exposed, the body kit exends all the way to the fender opening and is still 1-1 1/2" wider there. It establishes a line that is picked up by the airdam and carried across the car.

http://i1271.photobucket.com/albums/...ntitled2-1.jpg

The other strategy shows in the MG EX 181 aerohead posted. The large-radiused rocker combines with the tight-fitted spats.

Which is best? :shrug: The MG is likely better in a cross-wind but the design isn't really compatible with doors. Consider a rocker that leans out 30° vs one that leans in 30° (similar to your body kit). Intuitively, the 'fillet' shape would offer less interference drag between the ground and the car's side.

I see 24 different complete body kits for the New beetle advertised , which includes front & rear spoiler /bumper covers and side skirts. VW Beetle Body Kits I can purchase pieces of the kits separately.
Although I think I will be staying with home fabricated parts at the moment , I should be able to make what I need.

I ordered a anemometer / wind gauge today from china.
I want to use it to do a small scale wind tunnel for the Beetle 1/32 kammback.

I need to know the wind speed, for scale model smoke testing.

1/32nd seems a little small-scale for wind tunnel testing. I've got a 1/18th New Beetle (in Cyber Green) somewhere, and the pros generally go to 1/4th or 1/5th.

http://i1.ytimg.com/vi/E1IIMFOFpjU/mqdefault.jpg
have you looked at the work of Yoshi Suzuka? He uses 1/24th.

Its all I have , Its going ahead once I get my wind gauge I expect to gather some information and photos from it, despite it being just 5" inches long.

If I see a bigger one, I will get it, they come in large sizes for kids toys, The Barbie New Beetle is about 1/5th scale (28" long) A friends daughter had one at one point.. I will see what I can turn up in the months to follow at the second hand junk stores.


edit:

I just checked eBay , the biggest it I seen was a 1:18th scale for $30

I think I will check a couple of the 2nd hand stores in the next city over , when I'm there visiting tomorrow. It would be nice to find a 1/5 or 1/8th scale Beetle. I have a 1/8th scale RC truck ,it's about 2ft long , it would be a nice scale to work with for the Beetle..

Freebeard You could (should) make a wind tunnel to test the 1:24 Beetle , You could test your designs and whatnot. Once the tunnel is built it could be used for some years testing scale models.

They have lots of really cool older cars in the 1:18 scale (11"x 5" ) Mercedes , fiats , Porsche's and so on, they would be interesting to wind test.
They also have Classic Beetles in many 1:18 forms / designs.

I am going to buy a 1:18 New beetle from ebay if I do not find something today.

eBay has a few Bburago 1/18 New Beetle's , made in Italy with opening doors, hood , hatch back and the steering wheel turns the front wheels. Its a very accurate cast model, even the mirrors look correct.

Possibly so, but then I'd never get re-finishing my new kitchen cabinet done. I have made progress with 3D modeliing. Something I couldn't do before works now. ??? I can remove the bumpers and change tire sizes. As soon as I merge the Beetle and boattail geometries, I'll have something to show.

Found this, it seems on-topic for this thread.

To Predict Turbulence, Just Count the Puffs - Issue 15: Turbulence - Nautilus

It then proceeds to a lucid description of Renynold's Number.

http://static.nautil.us/3682_898aef0...a8e9903991.png

It must work at *very* high velocity, in order to compensate for the small scale. The sensors look to be pretty sensitive. Pretty darn cool!

A discussion on the Reynolds Number and windtunnels:
Wind tunnel tests and Reynolds Number

The Reynolds number is New to me , it is now making my head spin/ grind/stop when trying to apply it to a design.

I was watching this video (from the Yoshi link) when Yoshi mentioned the wind speed he would be using is 40 to 50 mph to simulate a full size car going 100 mph.
EDIT: I WAS WRONG , YOSHI ADDED THAT IT EQUALS 2MPH
It appears that you reduce speed to simulate a real car.
I had also thought I would need to increase the wind speed.. it just got easier to construct by a few fans at least and several mph.

http://www.youtube.com/watch?v=E1IIMFOFpjU

The rabbit hole goes a long way down. Can anyone explain the scaling on the Re axis to me?
http://www.aerospaceweb.org/question.../drag-disk.jpg


The link you posted says no.

No, he said it would simulate a real car going about 2 mph.

Ideally, you want the Reynolds number of the model to be the same as the vehicle you are simulating. But since the car is 1/32 size, you would have to run the wind speed up to 32 times the real car speed that you want data for, which is not practical for most people.

But you can find out what the relative impact is for changes to the shape. Just try to make the model as big as possible and the wind speed as fast as possible.

FYI, a quick approximation for Reynolds number for objects in air at sea level is
Re = Length in feet X Speed in mph X 10,000.

So a 15 ft long car at 50 mph would have a Reynolds number of 15 X 50 X 10,000 = 7,500,000.

A 1/32 scale model in a 50 mph wind tunnel would have a Reynolds number of 15 X 1/32 X 50 X 10,000 = 234,375, or 1/32 of the full-size car. If you double the wind speed in the tunnel to 100 mph, the Re would be 1/16 of the full-size car. See the relationship?

O.k I need to study this more..

I had a hard time understanding Yoshi at that point in his description, I lost his meaning at that point.

It's a logarithmic (Log) scale. Each major division increases by a power of 10. Where the 1 is would be Re = 1; where the 10 is would be Re = 10; the next one is 10 squared (see the 2 to the right just above it?), or 100; the next is 10 to the third power (cubed) or 1000; next is 10 to the fourth, 10,000; next is 10 to the fifth, 100,000; then 10 to the sixth, 1,000,000; and last is 10 to the seventh, or 10,000,000.

The other numbers in between the big divisions show how far that vertical line is to the next big division. Take that number and multiply it by the nearest big division before it to get the Re for that line. For instance, the vertical line to the left of the word Sphere is close to the line marked 4 at the bottom, but that line is to the right of the 10 to the fifth major line. 10 to the fifth is 100,000, so we multiply 100,000 by 4 to get 400,000 for the Re when the drag of the sphere drops way down.

Log scales are used to squish the graph down to a reasonable size.

Thanks for the insight.

I do see the relationship , thanks for that and the formula for scaling.

How the Re interacts with vortexes is still unclear , If you could simplify that , It would be much appreciated.

Once you've got a handle on that, all you need is an ample air source, and an effective diffuser to straighten the flow. It's job is easier if the fan is behind the model.

I'll just point back to here:
http://ecomodder.com/forum/showthrea...tml#post433339

That mans research is what I need to study alright.

I came across this student thesis on a scale wind tunnel.

http://www.davidoprevatt.com/wp-cont...ors-thesis.pdf

It's interesting that his focus was on buildings in hurricane force winds. It was a consideration in the dome I designed in 1980 for the Oregon coast.

I haven't read the student paper yet, but I will. Just glancing at the Contents it looks like what I was calling a de-fusser he calls a settling chamber.

Patrick -- Thanks. I knew it was logrithmic; looking at it today I see the superscripted exponents are interleaved with the subdivisions. It wasn't making no sense yesterday.

You forgot "induced drag", which is caused by lift on a body. Trying to direct all the air on the trailing edge near the ground creates lift...especially if you don't have a 20 foot tail.

struts are at least 20x wider than their chords. End effects don't matter. Car bodies are relatively narrow, and the above rules of thumb do not apply.

wind drag *begins* to have a greater effect than friction at 50 mph. more streamlined the vehicle, the higher that speed. there is a point diminishing returns, where fuel economy doesn't increase, and the manufacturability, and style, go out the window.

GM seems to design its cars to a CdA of 7.0. Germans are a little better, at ~6.0. Engines are designed to operate with a load, and 2500-3500 rpm. A super low drag car, cruising at 1500 rpm, is not optimum. imo the aero civic owes more to its lean burn engine (wouldn't pass emissions in certain areas) than to its hideous boat tail.

Can OEM lower CdA to 5.0-6.0 without increasing cost? Yes. will it improve hwy mileage at 60 mph? not much, if at all. If the government, like aerohead wants, raises speed limits to 100 mph, then yes.

Well that's no fun is it.

I think your forgetting that work or force is not done without energy , adding a wind load will not benefit fuel economy , removing wind load will. Driving at 1500rpm will save fuel over driving at 2500-3500 because it takes more fuel to drive at higher rpm's.

My understanding is very different then yours.
The load theory does not trump all other conditions , not by far , the load theory is more about acceleration then cruising purposely with a wind load to strain your motor , or make it do more work while cruising then necessary..

quoted myself .. its catchy :)


So I ended up buying one of the above mentioned models for $20 with free delivery. I had a few Beetle choices but choose a 1/18 New Beetle Cup car.

So I can test it in its race configuration then add a Kammback for further testing.

I will wind test with what I can , I remember that my grand dad made a shop ventilation fan by attaching metal blades to a 2hp electric motor - it was powerful.
I also want to make a water test system up , a fish tank to test with a inlet and outlet, yet to be designed or thoroughly looked into ..

I do not like the graphics or color much.. I bought it with no intent on reselling it as I will be damaging the paint in short order in the name of science ... ouch to the collectors with there neatly displayed boxes.

A photo of the "Cup" car and the body kit , basically it allows bigger tires and is lower to the ground.

Okay, so:
[Google]

So, to simulate a car going 55 MPH, you need a model that is 761.207051 / 55 = 13.8401282:1

Otherwise, you would be creating a sonic boom indoors.

my point was that operating an engine far from its peak volumetric efficiency...is NOT as efficient. Even if it burns less fuel overall, but it's burning it LESS efficiently. rpm is not the determining factor (although there is slightly less engine friction to overcome). A car needs a certain amount of HP to cruise at a given speed and load. A downsized engine tuned to make just enough HP at cruise rpm, and peak VE, is the most efficient option. engine tech and driving style have more impact on consumption than a few counts of Cd.

My other point was, that OEMs know that aero doesn't make much difference at 60 mph...that's why they don't try. In city, it makes no difference at all, and hardly impact EPA mixed cycle ratings. Low Cd is more a marketing tool than anything.

You are convinced of your opinion even if it makes no logical sense from start to finish.

Your statements support each other only because they are both wrong.

I tried to explain it to you but your bent on peak performance not economy

I 'll paint a picture for you

Say you as in You are driving in your peak volume efficiency all day long.

But me I'll be cruising at 1000rpm all day, who do you think would consume the largest amount of fuel ?

How about we dispense with pictures and use numbers and actual logic?

Using the same car, you may use less fuel, and you'd be much slower. But that nowhere near what I'm saying. I say "efficiency", you understand "performance". ?! RPM does NOT automatically mean less fuel. It's about the load on a vehicle (drag and friction) and how efficiently the engine produces the power to balance the load. Newton's second law.

Haven't you ever compared highway consumption of a mid sized car with the four and six cylinder engines? The Malibu 2.4 gets 33 mpg highway vs the 3.6 26. That's 27% better. It cruises at a higher rpm. That is typical. Still feel I'm wrong?

There's an optimum engine size for a given car, at a given speed, which minimises consumption. Eg The 2014 Stingray is best at 5.7L per GM's SAE paper. The 6.2 was still marginally better than the 5.3 because it ran in V4 mode longer.

To take it a step further, design a smaller engine with the same power at part throttle, and lower pumping loss, and you get better economy.

This is the rationale to OEM downsizing engines. It's not "my logic". I'm done arguing.

I mean I guess you're technically correct, make an engine that runs slightly lean, 30:1 static compression ratio with a super long stroke, Atkinson cycle cam that closes 130 degrees after BDC, tuned headers and intake for VE below 2000rpm, and you'll have super high efficiency. If you need 15hp to cruise, then such an engine would only be capable of producing maybe like 25hp peak, and you wouldn't be able to accelerate.

An engine producing more power used at peak VE would be equally efficient possibly more, but it would also be sitting there wasting fuel more of the time.

VE and efficiency are very different anyhow, VE is a function of all the breathing on the engine, header, exhaust, valves, combustion chamber/piston, rod/stroke, and most importantly cams. Peak VE is very rarely where peak efficiency is, because peak VE on mild cams tends to be achieved where the intake and exhaust resonant frequency matches the engine speed, but that number is quite arbitrary.

The higher VE=higher efficiency argument only works when the whole system is designed for peak efficiency, because usually more VE doesn't increase efficiency downstream. For example past 60% load more VE means more energy wasted on the exhaust stroke. People who think higher VE means higher thermal efficiency are just practicing wishful thinking, BSFC is way harder to figure out and measure.

Exactly !! you had MAJOR breakthrough just there , now hang on to the thought.

And go drive and save yourself some gas - instead of saving some time !

So you're the wonderful guy who does 90 kph on the highway. It all makes sense now. Thank you for a thoughtful and insightful discussion.

A clean car is a happy car!

So simulate it at LESS than 55 mph. DUH!!!

some new images
 

On page-1,in the 1st series of images I've added a 'golf ball',bowling ball,and Mair's boat tail.They're at the bottom.
If you compare the curvature on the balls leading up to the separation point with Mair's and the 'Template' you'll find a close correlation.
It is the gradual body convergence and pressure rise created by it, which allows the turbulent boundary layer to continue to receive kinetic energy,strafed in from the inviscid outer flow,critical for holding the TBL attached to the aft-body in spite of its 'desire' to flow backwards,against the current,towards the windshield headers low pressure region.