Aerodynamic Streamlining Template: Part-C

[ChazInMT]

Something I notice, the air on the side is moving Up...I'd think we'd wanna do what ever it takes to keep the air moving in a straight line in relation to where it starts. If this is achieved then the air will end up where it started before going around the shape and this will accomplish the best result. If the air is moving up, then it will meet the air coming from the top and create the swirl (AKA Vortex). I don't know what it will take to balance this out.

ERTW, This is GREAT stuff you are putting up here, I don't know what we did to deserve you but I feel lucky to have you around. A Huge THANKS from me!

[Cd]

Great thread .

[NeilBlanchard]

Yes, looking at that rear view makes me wonder if a flattened "beaver tail" or a vertical "fish tail" would result in a lower drag shape?

[drmiller100]

ERTW, THANK YOU!!!!!!!!!!!!

The best data I have seen on this site, bar none. And it models my experiences in the real world.

I believe the rear of the car is where the vast majority of gains can be found. Form drag is MUCH more significant then the rest of it all.

Looking at the car you modelled, you can see where the air separates at the back. In the real world, a tiny bit of side wind could easily cause the air to stay attached longer on one side, but not the other. I wonder if this could cause the car to pull to one side.
A solution would be to chop the car at the point the air trips.

FWIW, you are modelling the basic shape of the trike I am building - at least the back half!!!!!

[Frank Lee]

Originally Posted by ChazInMT

Something I notice, the air on the side is moving Up...I'd think we'd wanna do what ever it takes to keep the air moving in a straight line in relation to where it starts. If this is achieved then the air will end up where it started before going around the shape and this will accomplish the best result. If the air is moving up, then it will meet the air coming from the top and create the swirl (AKA Vortex). I don't know what it will take to balance this out.

You want the air to be displaced the minimum amount. If the form is tall and narrow (like a vertical airfoil) the displacement is least by shoving the air laterally. If the form is short and wide (like a Countach) the displacement is least by shoving the air more vertical than lateral. If the form is about the same height and width then a displacement vector of about 45 degrees up and out (as shown) would be optimal.

The air is going to pretty much figure out where to go for the least displacement around any particular form. It's up to us to provide nice slippery forms. I think that since crabbing into x-winds is generally a solution with too many practical difficulties, the low+wide form has aero advantages over square and tall+narrow forms.

[ERTW]

Although there is room for improvement, it's only 10 counts. Lengthwise it's at ~80% of the full template (0.083). I'd guess that a flat chop vs the diffuser would help a bit. Fences can help. Shark scales are basically mini fences to prevent vortices. I'm also impressed with how little lift the shape has.

Im in the process of analysing the template as a prism. Cd was 0.247 at 200 iterations, and is now hovering at 0.192 at 262 iterations. It's been 7 hours. I need a 6 cpu xeon . Btw I had an car shape long ago which I called the aerowog. This reminds me of a gobi or pollywog. Peak air speed is 42 m/s. i intend to lower this by beveling the front end.

[drmiller100]

why so wide? your frontal area is HUGE, and I wonder why you don't make the sides more vertical

I don't know what you mean by prism. ?????

[KamperBob]

ERTW, keep up the great work!

Food for thought:
1. Simulations converge toward solution through local minima without necessarily ever reaching the global minimum.
2. Computational results must always be validated externally to judge accuracy.
3. Sometimes perfect is the enemy of good enough.

Rock on!

[aerohead]

Originally Posted by ERTW

Phil, do you have a front view and a plan view of the template? I don't think I'm getting the shape right. I set ground clearance at 10% of the body thickness with sharp edges and the Cd is coming in at 0.275 :/ Do you have a solid model?

I've also taken a 4:1 rear body with a 1:1 front end (same length as a 2.5:1 ellipse) which comes out to Cd 0.247. Using a 6:1 rear body (used to model dolphins) only dropped it to 0.243. Both had a 10% fillet and 10% ground clearance. What do you think?

edit: I rounded the bottom edge (R5" or 10%)...Cd 0.191, CL ~ 0.055. Sharp leading edges are a no no! I want to protrude the nose a little into the wake, and sharpen up the trailing edge. I believe it "wants" a 20% radius on the leading edge, and ~12.5° max angle. I think 25% ground clearance is closer to your template (and helps get it above the earth's boundary layer). I'll report it asap.

I don't have the other views yet.I can't seem to keep with all the traffic here at EcoModder.
I don't have a solid model either.
Looking at your image,it looks like the front profile rises to the max camber point,but does not reside there before the downward slope begins to the rear.
This is the most important thing about the template.It must be exactly like it it shown.
Shorter and we have separation.
Longer,and surface friction goes up.
Essentially,we take a 2.5:1 streamline body of revolution,bring it right down into the ground,where only half is above ground,the other half submerged( ground reflection),then cut away for ground clearance with no alteration of the height.
The ground clearance I show in the template is the minimums for the 16-degree/10-degree/10-degree SAE approach,breakover,and departure angles.This clearance needs to be maintained as well.
The body must be narrowed,as it would be too wide,so some morphing is required.
I was to have minimum radii at the lower edge,blended into the wheel fairings.
Above the max vertical camber point ( in line with max protrusion of nose ) the whole body is rolling into curvelinear tumblehome as Jaray's pumpkin seed.As the University of Bochum Solar West GT.Ditto Schlor Car.
The front elevation is going to be very much like the Bochum car.
The plan view will be much like the 1957 MG EX 181.The is a auto blueprint of this car online.
Walter Lay's pumpkin seed car of 1933 has a perfect match with the template aft-body.It showed at Cd 0.12.Jaray's was hypo-template and came in at Cd 0.13.I truly believe that these are the numbers we should be seeing.And that's with ideal cooling systems,features drag,wheels,etc..
The 1987 GM Sunraycer is a dead match to the template,from nose,out to 25% aft-body,afterwhich it refllexes and goes hyper-template a bit out to 89% of the template aft-body length.At Cal Tech,with wheel fairings,the Sunraycer model registered Cd 0.089.
I love your images! I was looking at the Honda Dream paper about their CFD.They broke the car body into 850,000 panels and solved simultaneously with the Navier-Stokes Equation.Blows my mind!

[aerohead]

Originally Posted by ERTW

I'm already ahead of you drmiller...the solver finished this morning after 12 hrs and 373 iterations!

I learned a huge lesson this morning. The goal plot was pretty flat at 100 iterations. Even at 200...and then around 220 it became a step function It was flat out until 350 iterations - and I was ready to stop the solver there...and then another step. wtf! It goes to show that what I assumed was "substantially" accurate may not be. I'll just let the solver finish from now on!

A moving floor becomes useful below 5" of ground clearance. At 13", it's more free air, than ground proximity. If the under body is rough, then it may be of benefit to drop the car, and force more air over the body. If the under body is smooth, it's more beneficial to have higher ground clearance. It's a trade off that's best explored in a wind tunnel.

I have the tail at 15°, and it still appears to detach somewhat. I hypothesise that ~12.5° is max. I believe the largest loss is the vortices. I think we could stop streamlines from migrating from the bottom, around the sides by widening the tail...more like a duck tail than a point.

rho = 1.205 kg/m³ @ 273.15 K
v = 30 m/s (108 kph)
A = 2.138 m² (23.0 ft²)
Fx = 108.68 N

Cd = 0.094

Drag index = 0.20 m² (2.156 ft²)

did I do good papa?

We need to chip in and buy you a bigger computer!I hope your sleeping okay.I wouldn't be able to ! Amazing images!