Trebuchet03's JettKamm (2000 VW Jetta)

[trebuchet03]

Originally Posted by ChrstphrR

Since Cd got you that model via my original digging ... I'm just curious now, trebuchet03 ... what format is that model saved as -now-?

Maybe I can help and do some of the trimming down on the model to help you cut down the size of it, so it won't bog your machine so. No ulterior motives for me, being that I get direct benefit from ANY work on Jetta models you get good results with, after all.

It's a solid works .sldasm and .sldprt

Most people don't have Solidworks.... Once I'm one cleaning up the model, I'm going to save it into a variety of formats... .stp, .stl, etc. and make it available for other people. Hopefully the converted file size will be smaller than the 80+ megs of Soldiworks file (it's shrinking as I clean it up)

you may want to have a look at The-Blueprints.com - 25990 blueprints online

Thanks I'm getting about the same... Except on the 2005, which seems to be a couple degrees steeper...

Cool stuff but not sure i understand the images, the different colours represent different air speeds, if so which colour is "best", do you want the air moving quickly over the surfaces?
Would be cool to see one done with a teardrop shape showing how a "perfect" aerodynamic shape looks in the program.

I can totally do a perfectly streamlined body

So what we're looking at.... In very simple terms, we're trying to minimize the blue zone in the colored pictures - those indicate areas where air velocity = 0. Ideally, everything would be moving at the wind tunnel velocity - 55mph.

When zoomed up close to the car - we're looking at the boundary layer. A thin boundary layer is a good boundary layer. Once the layer gets too thick, we call this separation. So on the colored images, we want a thin layer of blue/green/yellow starting at the beginning and extending to the very end.

In the black and white images, we can more closely see where separation is occurring... We don't want separation - but if unavoidable, we want it as far back as possible. When you see a bunch of black and white lines leave the vehicle, that's basically separation.

These are, of course, generalizations

[aerohead]

Originally Posted by trebuchet03

So, as of yet, I have not modeled directly against the Prius....

The last cross section I posted was a simple line, from the trailing edge of my trunk to the first tangent point on my roof line. I'm thinking, if I go this route - that'd be the way to manufacture it - because I'll need a false windshield and I'm ill equipped to make optical compound curved plastics...

I'd have CFD for you, but I accidentally killed the power to my laptop before going grocery shopping - my laptop died mid calculation

How's this picture?

Sorry,late to the party.I was looking at some of the "side-by-side"comparisons with the Prius and couldn't see the bottom of the tires very well.What I wanted to share,is that when scaling,it's important to match the height of the cars,as the cars fineness ratio is dependent upon it,and progressive roof curvature without separation,also a function based on height above the ground plane.

[ChrstphrR]

Originally Posted by trebuchet03

It's a solid works .sldasm and .sldprt

Most people don't have Solidworks.... Once I'm one cleaning up the model, I'm going to save it into a variety of formats... .stp, .stl, etc. and make it available for other people. Hopefully the converted file size will be smaller than the 80+ megs of Soldiworks file (it's shrinking as I clean it up)

Wow, it changed from a few MB as a DXF to 80mb under Solidworks?

I know the interior bits are superfluous, and could be removed, but... that's a huge change. I remember the old thread where you were having the machine crash loading the model, and I couldn't fathom why. 80mb of data, and then rendering it onscreen, now I can.

[trebuchet03]

Originally Posted by aerohead

Sorry,late to the party.I was looking at some of the "side-by-side"comparisons with the Prius and couldn't see the bottom of the tires very well.What I wanted to share,is that when scaling,it's important to match the height of the cars,as the cars fineness ratio is dependent upon it,and progressive roof curvature without separation,also a function based on height above the ground plane.

My car vs. Prius in reality - are within an inch in height In those pictures, the tires got cut off.... But, it would seem that those images weren't optimal for that sort of comparison (basically, I grabbed the first image that looked somewhat ok).

Wow, it changed from a few MB as a DXF to 80mb under Solidworks?

I know the interior bits are superfluous, and could be removed, but... that's a huge change. I remember the old thread where you were having the machine crash loading the model, and I couldn't fathom why. 80mb of data, and then rendering it onscreen, now I can.

I didn't even include the interior features The reason for the increase in size is because SW converted everything into surface features... It's actually a fair bit bigger than 80 at the moment :/

-----
Fresh Batch of Results....

This is a Fusion like model - deck lid was extended to twic the original length and I followed the lines of my trunk










So, is it better than stock?



Can't say... For sure, the wake is longer with the modified deck lid. And the extended lid doesn't seem to change where separation occurs. Is it better? I can't say for sure - but I'm leaning towards no, not for my car.

From CD-adapco's website

"Drag reduction of the Ford Fusion Hydrogen 999 was a key element of the successful world record attempt. By taking advantage of CFD analysis, physical test times could be reduced and multiple configurations of the vehicle studied. Other benefits included insights such as adding a rear spoiler for stability. Eventually the drag was reduced from 0.34 to 0.21, helping the Fusion to become the first fuel cell powered car to pass the 200 mph barrier. Partnering with CD-adapco's engineering services, and using its advanced family of STAR software solutions, allowed us to draw from a greater pool of expertise spanning CFD and aerodynamics." John W. Zaleski, Program Manager, Roush Industries.

CD-adapco Press Room - CD-adapco helps Ford design the world’s fastest fuel cell car

Stability obviously means it's doing something aerodynamically.....


In other news - I took a long stick to my car - to see how much would be added to kamm that went from trailing edge of trunk to tangent on roof as seen in a CFD model posted yesterday.... It's not big at all - in fact, it basically extends from the top of my rear glass down to my trunk.

Additionally, I have figured out a taper line that is easiest to manufacture - basically, following the line my trunk seam makes. If all goes well, I'm shooting for a rough prototype next weekend.




Next in the pipe

• Flat Deck Extension
• Boudnary Layer Trip
• Streamlined Foil

Still taking requests.

[Deezler]

Originally Posted by trebuchet03

Next in the pipe

• Flat Deck Extension
• Boudnary Layer Trip
• Streamlined Foil

Still taking requests.

Awesome! Thank you for doing this! This thread is rad.

I want to build my deck lid extension soon, and can't decide on whether to stay with the trunk line or bend it up to extend flat behind the car. I would think that staying with the declining angle of the trunk would be best, IF the flow remains attached. Hence some VG's on the roof to trip the flow.
Also much easier to manufacture, thinking just a flat plate of aluminum or maybe plastic (whatever's cheaper, ha.). With the VHB you speak of for attachment (where does one find industrial grade VHB tape?).

Where are you going to locate your "trip" in the model? I would think just before the roof line starts to bend downwards would be most beneficial...?

What do you mean by streamlined foil? A raw shape to test? (i.e. not something done to the jetta?)

Questions about your model:
- How did you obtain the geometry to build it? - It looks great btw. I would LOVE a copy once its all cleaned up!
- Did you model the curvature in the 3rd dimension? Specifically the gentle arc across the trunklid rear edge:



So you are running all these CFD simulations. What output are you getting other than a visual flow separation analysis for each iteration? Can you compute a Cd number of some kind also?

Seriously though, its rare find a study being done like this, with open data sharing, let alone for my own car... Sweet!

[roflwaffle]

What are the specs of the rig you're using? I tried messing around w/ openfoam/some stuff from gmsh and my computer could not handle it. Oh, and I second the stuff Deezler mentioned. I'm interested to see how curving the sides of the spoiler would change the wake.

[modmonster]

hi

great post keep the simulations coming. i wish someone would do this for a hatch back.

Question: what objective function are we trying to minimize here? speaking in 3D terms -is it important to just minimize the VOLUMEof blue stuff or is it important to minimize the REAR CROSS SECTION of car that separation occurs from?

[trebuchet03]

Originally Posted by Deezler

Awesome! Thank you for doing this! This thread is rad.

I want to build my deck lid extension soon, and can't decide on whether to stay with the trunk line or bend it up to extend flat behind the car. I would think that staying with the declining angle of the trunk would be best, IF the flow remains attached. Hence some VG's on the roof to trip the flow.
Also much easier to manufacture, thinking just a flat plate of aluminum or maybe plastic (whatever's cheaper, ha.). With the VHB you speak of for attachment (where does one find industrial grade VHB tape?).

I'll add the VG with deck lid to the queue I first need to figure out a better placement for the VG - I did one run with the VG right about where the rear glass starts.

VHB... I haven't looked yet... The place I was interning at last summer had a large roll of it - so I imagine it can be ordered I don't know how much it will cost, however.

Originally Posted by Deezler

Where are you going to locate your "trip" in the model? I would think just before the roof line starts to bend downwards would be most beneficial...?

First test was around the top of the glass... Separation doesn't occur until about midway down the glass, so I'm wondering it they'll need to be placed on the rear glass itself...

Originally Posted by Deezler

What do you mean by streamlined foil? A raw shape to test? (i.e. not something done to the jetta?)

I modeled a NACA 0012 foil - nothing to do with the Jetta

Originally Posted by Deezler

Questions about your model:
- How did you obtain the geometry to build it? - It looks great btw. I would LOVE a copy once its all cleaned up!
- Did you model the curvature in the 3rd dimension? Specifically the gentle arc across the trunklid rear edge:

At the moment, I'm using a 2D model sketched off an image (as seen in my posted images). The fancy 3D model you've seen has not been used yet - it still needs processing to get it usable.

Nice tail lights

Originally Posted by Deezler

So you are running all these CFD simulations. What output are you getting other than a visual flow separation analysis for each iteration? Can you compute a Cd number of some kind also?

As of this instant, I'm mostly just getting visulization results. BUT, I have setup most of the models such that I can grab cD etc. with a little extra work... I don't have/know the equation off hand, but after I spend a few minutes to figure it out, it's a matter of implementing it into the model. Then, I can grab cD I'm not going to do that just yet as the equation is variable depending on geometry.

Originally Posted by Deezler

Seriously though, its rare find a study being done like this, with open data sharing, let alone for my own car... Sweet!

I'm not exactly sharing data - not that I don't want to, it's just I don't have anything... I am sharing the pretty pictures - open for discussion and interpretation We can compare a few things and then decide what is better, or if we can't determine so...

What are the specs of the rig you're using? I tried messing around w/ openfoam/some stuff from gmsh and my computer could not handle it. Oh, and I second the stuff Deezler mentioned. I'm interested to see how curving the sides of the spoiler would change the wake.

P4 3.2 GHz
896MB RAM (1GB total - 128 goes to video)
Whatever speed hard drive that comes in a hpzv5000 laptop
WinXP Home

I'm using Cosmos Floworks in Solidoworks... It's not the best solver - but it's not terrible... As of right now, I don't think it can do LES (Large Eddy Simulation) - so we won't see much in the way of eddies/vorticies.

I'm using a half scale model too - to reduce on computing time... However, I'm stupid and forgot to look up weather or not solidworks could do 2D cfd, which it can... Run time will be much faster now - while I'm still in 2D mode

I've never used OpenFoam, so I can't say too much.

Question: what objective function are we trying to minimize here? speaking in 3D terms -is it important to just minimize the VOLUMEof blue stuff or is it important to minimize the REAR CROSS SECTION of car that separation occurs from?

I can't say for sure at the moment... Optimally, there's no blue zone... What's better? Longer versus thinner? I'm not entirely sure...

As of right now, just take everything as general visualization Later, after finding the better iterations, I'll go back and see about getting barometric numbers for comparison. I'm not doing that now because of the time involved



------
In other news, my use of my laptop is going to be cut a little short soon... My girlfriend's laptop took a crap, and she needs it for work...Her machine is identical to mine, so we're going to swap hard drives depending on who's using it. If I can't fix her old laptop tonight, she's ordering a new one that should be in by the end of the week.... Progress will slow down, temporarily

It's also getting colder, which makes working with fiberglass resin a bit more difficult :/

[trebuchet03]

Images From Flat Deck Lid









I'm not convinced that it's beneficial for lowering drag... Maybe reducing lift, I can see that, but I'm not so optimistic otherwise...

For reference... Here's the baseline



And one more baseline



EDIT:
I just measured my model and the Fusion 999

Rear Angle
My Model: ~27 degrees
Fusion 99: ~20 degrees

Profiles look similar, but that subtle 7 degrees is making a big difference...

[trebuchet03]

Okay, so it's bothering me - comparing the fusion to the jetta....

So I grabbed an image, made a quick rough model, and threw it through the solver....

Red= 55mph or higher
Blue = 0mph



From this, I'd say the truck deck can be useful as flow near the vehicle on the rear end hasn't lost much velocity. Just a guess

I'm still curious as to why that slide (earlier picture) doesn't attribute the extended deck lid to any gains.... Perhaps it just does a great job at straightening the flow, to avoid nasty lift conditions at 200mph