Trebuchet03's JettKamm (2000 VW Jetta)

[Christ]

Originally Posted by roflwaffle

Or they could just use a VGT.

How would a single VGT create more airflow than a setup where one turbo is activating another one? VGT's don't serve that purpose... they are basically there to keep specific boost under varying conditions.

My idea involved spooling the first (smaller) turbo off the engine, and using it's spooled (compressed) air, obviously a much larger CFM than the exhaust flow, to spool another turbo's exhaust side, before going into the engine. The secondary turbo's compressed air would then be fed through ducting into the wake at the rear of the vehicle, hopefully eliminating part of the wake.

Through the feedback loop that standard turbos use, this isn't really possible though. The secondary turbo would never reach "full spool" and thus, wouldn't create the kind of airflow it could otherwise achieve.

Also, a turbo won't spool properly without more airflow than it's host engine can achieve, hence the reason for "turbo lag".

[trebuchet03]

New machine is much better

It came with Vista - the version of SW I'm using won't work with Vista. I added a second hard drive and used my laptop's copy of XP on it (which happened to be a regular XP install disk). I'm not using it anymore and it's got a bum screen

Finding drivers for the mobo's chipset have proven to be difficult... And I had to jump through hoops to get the video card working (nothing special - it's onboard ). This is why, in addition to finals, it has taken me some time to get things operational.

But that said - I could open the 3D model and it only took 5-10 seconds! I can move the model without issue and can make edits in a reasonable period of time.... BUT - I can't run CFD on it

Something I didn't anticipate.... Because the model is made of surface loft features (basically, 0 thickness) it's not recognized as a component. If it's not a component, it can't be meshed

So, until I sort that out... I took the side profile of that model, and made a new 2D model. There's some big differences and I've got more detail including windshield wipers).

-----------
Runs that took 3-4 hours before, is now taking 45-60 minutes at an even higher resolution

Old Setup
1 100GB HD
1GB PC2700 Ram
128MB Shared ATI 9000m graphics card
P4 3.2GHz with HTT

New Setup
1 80GB HD to run XP and holds the Vista pagefile
1 620GB HD for Vista AND holds the XP pagefile
6GB of much faster Ram (32bit XP can only address 2.87GB - .5 goes to display)
512MB Shared Geforce 6150 graphics card (old - but working pretty well)
Phenom 9500 Quad Core (2.21GHz)



And now you know what I've been up to

-------
LeOld
6-7 hours to run


LeNew
3.5 hours to run

[Deezler]

Yes! CFD back on line just in time for my 3 week vacation from work.
I need your guidance on what areas of aero-modding to focus on during my time off!

still think you should close off the bottom of the engine bay....

[aerohead]

Originally Posted by mfor4x4

I did some more research after I had posted about the fans sucking the car to the ground. I know this would not work in this format for FE, these race cars were using over 20hp to drive these motors/fans.

Here is another option though to get free Horsepower to drive a pump like I am suggesting. I know this may sound crazy so I we need someone to tell me this just to confirm that I am indeed crazy, mounting a turbo on the engine that would not put a single CFM into the engine, it would just put air out the back of the car? Some of the turbos out there put out over 750CFM, if there is no waste gate on this turbo and it developed say 30psi of boost, that is 750CFM going out the back of the car in the dead air zone, I think that it would smooth out the back aero of the car. The added bonus of this would would be the amazing whine the turbo would make going down the road.

Since the turbo is not an "over-unity"device,it cannot derive more energy from the exhaust than is contained in the exhaust.Simply allowing the tailpipe to "blow-down" into the wake,would maximize any "thrust" available.

[trebuchet03]

Originally Posted by Deezler

still think you should close off the bottom of the engine bay....

It's coming The current run of CFD has another 30 minutes or so

[wyatt]

Should your ground plane be extended more to the front and back of the car, or is that just a representation of where the ground would be? I am just seeing what appears to be pressure crossing the ground plane.

[trebuchet03]

Originally Posted by wyatt

Should your ground plane be extended more to the front and back of the car, or is that just a representation of where the ground would be? I am just seeing what appears to be pressure crossing the ground plane.

Sorry, should have explained that For each configuration, I run two solutions. One as seen and another with the mirrored body ground plane. The one you're looking at has no ground (I probably should have turned off that plane before grabbing that image). For all intents and purposes - there is no "ground" surface - just a plane that shows where the ground would be so I can mirror the body over it.

Once I have both solutions, I go back and compare the difference (dividing applicable forces in half for the mirrored version).

I should try adding a moving ground - with the new computing horsepower I have

[Hasbro]

Originally Posted by trebuchet03

I should try adding a moving ground - with the new computing horsepower I have

Would really like to see how that impacts your tests.

[Dieselguru]

Seeing how this is my first post, I'll try not to overstep my welcome.

Tre - It would be very interesting to see some of the runs you have done at a little higher speed, say 75 mph (the speed I drive my 110 mile round trip to work every day). If there are gains to be had, they they will show up faster at those speeds. Especially things like VG's.

As a side note: probably the reason most people don't see much improvement using VG's is that
1. VG effectiveness increases with speed (need a seperation point, otherwise its just drag)
2. VG placement probably depends very closely on speed (location of seperation point)

Look very closely at the Mitsu report on VG's. In the test setup section (page 2), they mention the baseline speed of 50 m/s (111 mph). if tre's settup is correct, will the seperation point move forward with increasing velocity to somewhat match that of the Lancer (wont be exactly the same, but simular), and will the VG's will become more effective?

Tre - if you still taking requests, through higher speeds into the mix and see what happens to some of your results.

[cmroseberry]

Some cold water:
I would not expect to see the separation points move around much at all for such a modest difference in Reynold's number. Furthermore, an increase in velocity (increase in Re #) would tend to promote a turbulent boundary layer which would tend to adhere further downstream on the surface.
One overriding consideration is that Volkswagen has almost certainly tuned-up this model in the wind tunnel. These Germans know their stuff - Hucho himself may have fine-tuned the Jetta shape. VW engineers are masters in making hardedge, almost boxy shapes aerodynamically efficient through detail optimization. Accordingly, I would bet the flow separates right off the trailing edge of the trunk lid.
As for vortex generators, I made some comments on the business intro thread about how they may help in some situations where the pressure gradient is close to the critical point for separation. Knowing VW's reputation, I would not recommend vortex generators on any of their cars except for possibly a Beetle.