Drag Bucket Paradox, sometimes less aero = less drag

[aerohead]

Originally Posted by Tesla

Ok,
As I understand it you may be suggesting a taller vehicle may need an even longer proportional taper profile.

Would this be the same whether that vehicle had a longer body or shorter one with the same frontal area?

This then also contradicts the use the narrower section to govern the taper as has been mentioned on numerous posts in regard to boattail design.

Got the 22-23° rule, never break it under any circumstances.

Got the Laminar thing, not relevant to road going vehicles.

Got the VG thing, many questions still remain unanswered and no conclusive evidence of benefits shown except in rare specialised cases.

But things still don't add up:
The Dryden van boattail, now they reported significant benefits and yes they did get seperation further down the on the original tail before they truncated it, but they most likely went way past the 22° point then as well.

As I understand the template theory, they should have had seperation on that tail in the first 6", yet flow remained attached for a good 4'.



from this thread:
http://ecomodder.com/forum/showthrea...n-21952-7.html

And then there is all the designs, patents and fitted boattails to Semi Trailers, none of these are anywhere near the template, and most are in the same range as the Dryden Van 10-15°, I assume they have had some proven testing over the last 10years, otherwise we would not be seeing an increase in their application.

I don't dispute the effectiveness of the template profile in boattail design, but am just trying to answer some of the contradictions.

I do know now that in fluid motion the rules change with size, materials and conditions, infact it seems we really don't know what the rules are.
We have a set of rules that can effectively describe an outcome under a fixed set of conditions, then as conditions change we need to change the rules.
Doesn't seem to be far from quantum mechanics, everything is just a model for an observed effect, but actual governing laws are ever elusive.

*I don't recommend anything other than the 'Template.'
*I offered the plan view 4:1 section profile only as an illustration of something 'different' which has been used to good effect.
*NASA's Ford Econoline boat tail lowered the drag from Cd 0.47,to 0.238.That's really good,but they left 50% of the drag remaining on the van.
*In fluid mechanics the shape defines the drag.Size makes no difference as long as Reynolds numbers are equal.
*We do know what the rules are:
*1 you can base a design around the 'Template' aft-body contour for low drag in a vehicle you can see out of.
*2 you can base a vehicle on a 'Reverse-Template' and achieve drag within 15% of the 'Template',but you won't be able to see out of it.
*3 you can base a vehicle on an ellipsoid and achieve identical drag as the 'Template' but you won't be able to see out of it,park it in a garage,go over a hump,around a corner,etc..
*4 you can base a vehicle on arbitrary legal length requirements and do your best to optimize for that length only,knowing you'll never actually achieve the lowest drag.

[Tesla]

just testing my posts are out of sequence, not seeing aeroheads recent reply.

[Tesla]

Showing 3 pages but can't view page 3.
Edit: Just showed up, but I'm still one post ahead.
Had some internet issues the other day when posting, I think that post has got stuck in the ether, I see it, but don't think anyone else does, so it's put the count out of sequence.
Oh well, whatever.

[Tesla]

Originally Posted by aerohead

*I don't recommend anything other than the 'Template.'
*I offered the plan view 4:1 section profile only as an illustration of something 'different' which has been used to good effect.
*NASA's Ford Econoline boat tail lowered the drag from Cd 0.47,to 0.238.That's really good,but they left 50% of the drag remaining on the van.
*In fluid mechanics the shape defines the drag.Size makes no difference as long as Reynolds numbers are equal.
*We do know what the rules are:
*1 you can base a design around the 'Template' aft-body contour for low drag in a vehicle you can see out of.
*2 you can base a vehicle on a 'Reverse-Template' and achieve drag within 15% of the 'Template',but you won't be able to see out of it.
*3 you can base a vehicle on an ellipsoid and achieve identical drag as the 'Template' but you won't be able to see out of it,park it in a garage,go over a hump,around a corner,etc..
*4 you can base a vehicle on arbitrary legal length requirements and do your best to optimize for that length only,knowing you'll never actually achieve the lowest drag.

For the rules, I was refering more so to the fluid dynamics equations, but those 4 points above do help clear some things up for me,
I think I'm stuck with #4 and trying to see what I can draw from 1,2 &3 to optimize that position.
Thanks.

[aerohead]

Originally Posted by Tesla

For the rules, I was refering more so to the fluid dynamics equations, but those 4 points above do help clear some things up for me,
I think I'm stuck with #4 and trying to see what I can draw from 1,2 &3 to optimize that position.
Thanks.

The computational fluid dynamics guys are working with the full Navier-Stokes Equation with spherical 3-D coordinate system and utilizing turbulence models and a host of ancillary computer sub-routines.They require a super-computer to model 3-D wake flow which absolutely must be understood.Aircraft don't have wakes per se,so they're a relative breeze to model numerically.
Guys like Hucho would actually put a car in the wind tunnel at Volkswagen and blow smoke all over it to reveal the wake.
Then a 'wake' would be constructed in 3-D,based on the smoke flow,and then be digitally panelized discreetly into the wireframe model of the virtual car for further CFD analysis.
Any iteration of the body would require starting again from scratch! The millions of dollars spent is much cheaper than developing a new powertrain for fuel economy though,so it turns out to be a big bargain.

[Tesla]

Originally Posted by aerohead

*1 you can base a design around the 'Template' aft-body contour for low drag in a vehicle you can see out of.
*2 you can base a vehicle on a 'Reverse-Template' and achieve drag within 15% of the 'Template',but you won't be able to see out of it.
*3 you can base a vehicle on an ellipsoid and achieve identical drag as the 'Template' but you won't be able to see out of it,park it in a garage,go over a hump,around a corner,etc..
*4 you can base a vehicle on arbitrary legal length requirements and do your best to optimize for that length only,knowing you'll never actually achieve the lowest drag.

I went back and had a look around for some ellipsoid drag data, mainly found info on airships, the one thing was that all these shapes seem to centre on the 2.5 - 4:1 Fineness ratio.

So what's going on there?
If you can take the template or reverse it or use an ellipsoid with the same Fineness ratio and basically get the same drag.

We have a sphere, which I think I read had a Cd of 1, if we put a cone on the front or rear it comes close to the template drag or if we split it and insert a cylinder between the two halves, we again get a drag figure approaching the template.

This has to be related to the boundary layer and it's developement over a distance, the front and rear angles don't change but drag does.

So this suggests to me that in some cases the frontal design may have a significant bearing on the overall drag.

I know that being in the impact/pressure zone there is almost guaranteed attachment with almost any frontal design, but seems to me there is something a bit more subtle there than just plain attachment.

[freebeard]

Originally Posted by Tesla

Last edited by Tesla; 02-01-2013 at 02:35 PM..

There's a Delete button too. Just sayin'.

And thanks for that video. I liked the real-world examples of things we would get simulations of today.

Originally Posted by aerohead

*1 you can base a design around the 'Template' aft-body contour for low drag in a vehicle you can see out of.
*2 you can base a vehicle on a 'Reverse-Template' and achieve drag within 15% of the 'Template',but you won't be able to see out of it.
*3 you can base a vehicle on an ellipsoid and achieve identical drag as the 'Template' but you won't be able to see out of it,park it in a garage,go over a hump,around a corner,etc..
*4 you can base a vehicle on arbitrary legal length requirements and do your best to optimize for that length only,knowing you'll never actually achieve the lowest drag.

I get this in general, but:

• What's a 'reverse template'? Why can't you see out of it?
• What's an 'ellipsoid'. Is that like a Dymaxion shape?

[Tesla]

Originally Posted by freebeard

There's a Delete button too. Just sayin'.

And thanks for that video. I liked the real-world examples of things we would get simulations of today.



I get this in general, but:

• What's a 'reverse template'? Why can't you see out of it?
• What's an 'ellipsoid'. Is that like a Dymaxion shape?

Yeah, I know about the delete, but I figure leave well enough alone, something is a bit out of whack on my view.
Anyway

The vid was good, there are three more from the same guy, he's not bad in terms of presentation and helps me to get a handle on some of this stuff.

The reverse template is just that, swing it around 180° and use the pointy end as the front, only 15% worse drag than the right way around. The vision issue is when you place your windscreen at a 10-15° angle to horizontal, the distortion, refraction and internal reflection make it almost impossible to see out.

The elipsoid is just that, an elipse, highest point of camber at halfway point along the length and basically identical front and rear.

Aerohead is right in that the template profile is probably the most practical to apply to most existing vehicles, but understanding that the same Cd can be achieved with significantly different shapes suggests there is a lot more to take into consideration than just throwing a boat tail on everything.

As said elswhere, it is already hard enough for me to get into a std parking space without adding 50% more length to it, so I am drilling down, trying to understand the detail more and seeing how some of these factors may be applied more so to my vehicle.

Not suggesting this is a viable option, but if a reverse template can get within 85% of a full boattail, then it may be, in my case with a blunt, slightly rounded rear, that some attention modeled on a reverse template may allow me to gain some benefits without the excessive length penalty.

[freebeard]

The elipsoid is just that, an elipse, highest point of camber at halfway point along the length and basically identical front and rear.

The Dymaxion isn't like that.

You'll see that when I said "what's an ellipsoid" I was being disingenuous. I didn't respond promptly because I've been under the weather (I'm not complaining because I'm down 10 pounds and a week ahead on my grocery budget )

I did show this before, a symmetrical, prolate 4v octahedron.


I hadn't run the software that generates the primitive geometry for a year or two and I'm sure it ran in the Terminal on Mac OSX but it doesn't want to now; but finally I found my old laptop running System 9.2.1 and managed to wrestle a 6v octahedron out of it. I tried something more exotic, a octahedral Bucky ball but had no further success.

I pipelined it through a 2nd program to convert file types and took it into Wings 3D.


Half the sphere was stretched 230% to get 0° camber at 30/70 forebody/afterbody ratio.


The bottom half was selected and reduce to 25% as a rough approximation of an underbody. It looks like I missed a step where I scaled on the X axis to get a 2.5:1 fineness ratio.


The entire object was reduced on the Z axis to 62% (Golden Ratio).


4 is 2.3x2.5, so I scaled in X by 230% to get a 4:1 fineness ratio.

Finally I added some materials for better visualization.


Additional modification aft the 22° camber point could even more closely match the template. The point would be that this method can provide any level of accuracy required, or attainable. You can see the improvement in contour between 4v and 6v. At 16-20v you'd have a monster data cloud, but it could be passed directly to a 3D printer.

[Tesla]

Originally Posted by freebeard

The Dymaxion isn't like that.

You'll see that when I said "what's an ellipsoid" I was being disingenuous. I didn't respond promptly because I've been under the weather (I'm not complaining because I'm down 10 pounds and a week ahead on my grocery budget )

I did show this before, a symmetrical, prolate 4v octahedron.


I hadn't run the software that generates the primitive geometry for a year or two and I'm sure it ran in the Terminal on Mac OSX but it doesn't want to now; but finally I found my old laptop running System 9.2.1 and managed to wrestle a 6v octahedron out of it. I tried something more exotic, a octahedral Bucky ball but had no further success.

I pipelined it through a 2nd program to convert file types and took it into Wings 3D.


Half the sphere was stretched 230% to get 0° camber at 30/70 forebody/afterbody ratio.


The bottom half was selected and reduce to 25% as a rough approximation of an underbody. It looks like I missed a step where I scaled on the X axis to get a 2.5:1 fineness ratio.


The entire object was reduced on the Z axis to 62% (Golden Ratio).


4 is 2.3x2.5, so I scaled in X by 230% to get a 4:1 fineness ratio.

Finally I added some materials for better visualization.


Additional modification aft the 22° camber point could even more closely match the template. The point would be that this method can provide any level of accuracy required, or attainable. You can see the improvement in contour between 4v and 6v. At 16-20v you'd have a monster data cloud, but it could be passed directly to a 3D printer.

Just lost the entire post somehow.
will try to repeat in brief:
I missed your ref. to the Dymaxion , but it looks like the first couple of images posted.

My main insight was that the fineness ratio was much more than a boattail, that you could take template or reverse it, or anywhere in between and get virtually the same drag.

So what does this mean, it suggests that the front profile and the overall length is just as important as the rear, yes we do need to have compound curvature, but the options are a bit broader than just boattailing.
It suggests if I build a boattail on the front of my vehicle I will get to within 15% of the drag of a boattail at the rear.

The other thing that it questions to me is the influence of the thickness of the turbulent boundary layer, this takes time to develop, and if we have longer surface contact at the front then the layer is thicker at the rear and will negotiate a steeper curve without separation, vs a short abrupt front, then a more gentle boattail curve is required.

So it’s not just about the rear departure angles, the front plays nearly as much in the overall drag, if we can reduce the intensity and depth of the pressure / bow wave at the front, then we can closer the rear much more quickly with more aggressive curvature.

The Dymaxion shape seems to be just slightly biased towards a rear boattail and maybe that is the ideal shape to be functional in the real world.

I have read many threads about the front curvature, but they are usually dismissed under the banner that the big gains are in the rear, but this information seems to suggest this is not quite the whole truth.