Trebuchet03's JettKamm (2000 VW Jetta)

[lunarhighway]

one thing to keep in mind is that on the lancer the vg's are there to direct the air towards the rear spoiler. VG's add drag to do this so while they may succeed in makeing the airflow follow too steep cuve the energy it takes to do this might offset any gains made by the smaller wake...

another thing that you might consider but might be difficult to simulate due to it's size is a small ridge just ahead of the trailing edge of the trunk. a good example is the reworked chevy volt, but i've seen similar setups on more conventional cars. on some cars a slightly more aggressive version of this is referred to as a gurney flap, although i'm unsure if it actually behaves like on an if it serves the same function as the ridge on the volt

some other low drag cars have a slightly up swept spoiler lip wich could also be related.

since its something relatively easy to fit it might be worth haveing a look at in case it's usefull

perhaps you'd like to test this idea also
http://ecomodder.com/forum/showthrea...html#post60227

it might be nothing but i'd sure like to see what it does compared to the rest

i wich i had suitable software for this sort of testing so i could try all my crazy ideas

[Deezler]

Originally Posted by lunarhighway

one thing to keep in mind is that on the lancer the vg's are there to direct the air towards the rear spoiler. VG's add drag to do this so while they may succeed in makeing the airflow follow too steep cuve the energy it takes to do this might offset any gains made by the smaller wake...

This is a commonly stated misconception about Mitsubishi's intentions.
http://www.mitsubishi-motors.com/cor...004/16E_03.pdf
If you don't want to read it all, they were able to reduce Cd by 0.006 with the addition of the VG's.

[lunarhighway]

perhaps my statement was a bit off, but still
there's two things wich bother me about the lancers basic geomerty and that's that the window is very steep and the trunk is relatively short. overall the design looks rather dated to me. most sedans i see today seem to have much more gentle curved windshields and better windsield/tunklength ratio. i wonder how effectice they are on cars where flow separation is minimal of where the flow can reattach to the trunk. especially when compared to a reshaped back where this sepparation would not occur and where there's no energy loss due to the energy needed to make te vortices

[trebuchet03]

Originally Posted by Deezler

Treb, thanks. Good Info. I will not deny that I'm a total aero newb.

When it comes to this like turbulence - no one really "knows" We can just make OK to decent guesses

Originally Posted by Deezler

You seem to imply that the Lancer, with its steeper angle, relies more on the latter. Being to steep for flow attachment, they were going for a clean break, so to speak. However, they apply VG's to trip flow down the rear glass! So my logic is this: If you have any flow detachment, and you can positively offset this with a simple trick like VG's, why not? I would think your ability to run simulations to examine this could be a perfect test of this idea.

I didn't mean to imply what the design goal was for the lancer. I just wanted to make sure that we don't directly compare one car to another because they are both sedans (or similar in that nature).

VG's are band aids. They are to be added as afterthoughts because you screwed up in designing (or compromised which resulted in less than optimal aero). As my design goals are to completely avoid the situation of flow detachment, VG's won't be really of any use to me UNLESS the vehicle is so long that the boundary layer just gets to fat and lazy

That said, I'm not abandoning doing cfd runs on VG's I think they're interesting and potentially useful for someone that isn't going as far as I am. Weather or not my CFD is of any help is a completely different subject matter. I'm working to the best of my knowledge/experience/tools

And I should say that I'm not for nor against VG's. My preference is for the "better" overall design. I am, however, against "blind" placement of VG's. I also feel that VG's, in some situations, can be placed properly such that there is no additional frontal area added. Which may be the case for the Jetta depending on the height

Originally Posted by Deezler

The Jetta angle is less, but from your own simulations we still see that the speed of air decreases as you move down the rear glass. Perhaps the Lancer angle being more severe made a trick like this necessary where it could have little benefit on a Jetta.

I really can't speculate on that - but that reasoning isn't unreasonable.

Originally Posted by Deezler

While its getting hard to keep track off all the pretty pictures thus far, a simple goal, in my mind, could be this: Minimize the total amount of blue behind the vehicle, and replace the green on the trunk lid with yellow.
I don't know the scales on the Mitsu pics as compared to yours, but they seem to have done just that with the application of the VG's.

When I'm done compiling the numbers - we'll have a better method of comparing subtle differences. It's taking some time because I'm adding more stuff to the results list

Originally Posted by Deezler

Whats up with the fusion model? Can you normalize that one to those done for the Jetta (i.e. same background flow color)? It seems to have perfect flow over the entire vehicle.

Do'H, I forgot to include a total velocity cut plot It does have pretty good flow over the top There is a small section where velocity slows down, but I wouldn't really call it separation (especially when compared to the Jetta)....

[trebuchet03]

Of the 7 things and their variants I've rerun, Only 2 (and their variants) show signs of success

While I figure out how to get tabular data up for everyone here's the sneak peak.....

What Worked
PriusKamm - It was a line connecting the trunk trailing edge to the closest tangent of the rear of my car. Not exactly the Prius, 18.x degrees versus 15.5 degrees. 3.85% over baseline @55
PriusKamm with Lip - Same as above, but with a flat lip at the end to straighten flow. Was slightly more effective than PriusKamm. 4.89% over baseline @55
Novel Spoiler - as seen on Green Car Congress - copied directly and pasted right in. Was beneficial, but not as much as the two above. Needs further testing. 2.3% over baseline @55

What didn't work
crxKamm - 10 degree slope, one of the first images I posted -26%
crazyLongKamm - well explain later, it produced a lot of lift -29.5%
trunkDeck - 7 degrees to follow trunk line -5%
flatDeck -16%
Three Iterations of VG's - my last one was getting close to promising, the "value" was better, but the "average value" was worse. Bonus points that this placement doesn't add to frontal area but alas, is placed on the rear glass in line of sight.


In the pipe....
A variety of flat spoilers at different angles... I think I'm going to make a whole bunch and do a batch run tonight. Initial results indicate that a flat spoiler in the configuration I modeled is slightly beneficial o.0


Still taking requests

I've also decided that I'm going to do a more official paper type write up, to consolidate everything here and allow the information to spread easier

[MetroMPG]

Whoops, forgot to answer this, re: the Prius pics...

Originally Posted by trebuchet03

Metro, where are you measuring the 9 degree angle? I'm getting ~15 degrees


When I measure from tip of spoiler to very top, I get something ~9 degrees....

I was measuring tip to top. Since the rear of the car is slightly curved, I didn't want to try to find a tangent on the glass area. I was just trying to show the relative distortion between the 3 images.

[tasdrouille]

I would be really interested in a more refined iteration of the novel spoiler. It appears to me as something that already offer a fair return all the while not being much invasive compared to other beneficial options. It would be relatively easy to mount.

[lunarhighway]

I've been comparing linedrawings of production cars with Cd's equal or lower than 0.29 and it seems on every single one of them the rear angle is between 15°<>16° for most of the lenght. it's also interesting to note that on some cars this angle goes even further down near the end of the windshield but at the end of the windshield theres either a spoiler lip or a trunk and if you continue this 16 degrees like it almost always meets with the edge of this straight piece.

this could be an indication on how long horizontal trunklid extentions should be.

here's a couple of them

they are not to scale and i've only looked at drag coeficients and not frontal area, obviouslt a good or bad cd is not only due to the rear angle

for example it was found that the citroen C4 coupé has a near identical rear slope than the honda insight although obviously the C4 is an otherwise more conventional car wich is less aerodynamic refined elsewhere

also it's always a gamble if values found on line are true

[i_am_socket]

Can we get sketches of the 2 lists for "what worked" and "what didn't"?

It seems like it's much easier to design for a hatch-back since they don't have to deal with trunks that may or may not re-attach flow.

[Deezler]

I am still wondering whether you tested the VG's WITH the deck lid extension....?

What is the difference between the "value" and the average value?

Would you please consider also running the simulations at a higher speed? say 70 or 75 mph?

The novel spoiler is very interesting..... I wonder how it might do if the trailing edge were angled slightly further downwards in line with the flow at that distance behind the car, instead of being more horizontal as tested here.