Aerodynamics in passenger car

pszczelaszkov · June 12, 2019, 12:59am

Hi.
I’m new in CFD and need some help before I will make wrong assumptions.
I’m driving my cars without a rear spoiler, but since I will be mounting it soon I was curious if there are any benefits of a stock spoiler except better look.
I found some 3D project of car I’m using IRL, and decided to make some tests at 125kph for a general overview.
First case is stock plain rear end.

Second one is small lip on rear.

And the last one is stock spoiler

If I’m correct “nut” in case without spoiler is sucking rear side of car causing drag?
Smallest “nut” is on version with tiny lip, but it creates a jet of air striking upward and more dirty air at the bottom for a car behind, while stock spoiler looks like it creates more steady airflow behind. Which one of those two spoilers helps better eliminate drag and could be more useful on track use?
Here’s the link for a project-> Project

ROHIT_SR · June 12, 2019, 5:44am

hi @pszczelaszkov,
your stock spoiler and small lip don’t make differences because both generate height low-pressure zone at the back of your car and a larger wake was generated. remember spoiler work at only high speed and use for generated large downforce. in your case spoiler are flat i think its should be like aerofoil for high downforce.

if your aim was to eliminated drag only then i think the stock plain rear end was good.

thank you,
Rohit.

pszczelaszkov · June 12, 2019, 12:36pm

@ROHIT_SR
Thanks for response, actually as reference I was using this blog entry.

Was focused on this part

Spoilers are barricades to undesirable flows, and thus are able to reshape airflow streams around the vehicle. This can help keep the rear of the vehicle down and decrease drag by changing the effective vehicle shape.

Clearly those two do something with air but, are they making more good or bad things with air behind?
Although sedan rear end is not so tragic as coupe.

ROHIT_SR · June 12, 2019, 1:05pm

hi @pszczelaszkov,
thank you for this link. really some good information about aerodynamic(wing and spoiler)
#redbull_gives_you_wings_NOT_SPOILERS.
ROHIT.

Get_Barried · June 13, 2019, 1:04am

Hi @pszczelaszkov,

The answer is not straightforward. Spoilers on top of generating downforce, also generate drag no matter what. The issue is whether the influence of the spoiler will help in reducing the overall drag of the vehicle. If you’re talking just a spoiler itself, the net gain in drag will always be positive regardless of how well designed your spoiler is. You can’t expect a device that manipulates the flow of air to not have drag.

While the specifics of how to control the flow of air with specific spoiler designs in order to reduce drag is not something I am familiar with, off the top of my head I would assume that the key concept is to ensure that flow manipulated by the spoiler is of minimal turbulence. Turbulent flow typically contributes significantly to drag, the spoiler helps reduce that, hence overall drag is reduced. I’m sure I am also missing other things like the influence on the rear wake of the car or something like that. @ROHIT_SR is much more experienced in these concepts. Maybe he can share some insights.

Cheers.

Regards,
Barry

pszczelaszkov · June 13, 2019, 12:52pm

Hello again @ROHIT_SR.
Since you were talking about pressure and the wage I decided to change some fields for better understanding pressure rather than “nut” which is as I suspect an acceleration of particles(m^2/s)?
Also i put seed on the rear, just noticed, that should be a correct way.

For nonspoiler pressure and particles swirl looks like this.

For lip version i have

and final stock spoiler looks like this

Can we predict according to shoots of swirls and data from “nut” which one is more turbulent?
Also take a look at wage’s at non spoiler and stock spoiler. They are not so different now, while lip is creating some underpressure pocket.
I tried to generate drag coefficient chart but I failed, im getting negative values, probably because of mesh and i dont know if i can trust those values. Maybe for sake of comparision they can be useful?

ROHIT_SR · June 14, 2019, 5:34am

hi @pszczelaszkov,

all post-processing image gives the answer to your question.

non-spoiler create the turbulent almost lower side of the car. this kind of turbulent by the lower part of the car.when flow separated from the car part(just like a diffuser.)

if you notice both cases create different turbulence. in a spoiler case has two turbulent zones.first, one generated by spoiler itself and the second one was the underbody part.
more you restrict the flow more turbulent was created.

your spoiler has high turbulent (high wake ) and also have a high low-pressure zone.

i agreed with the @Get_Barried comments.

can you explain step by step procedure for a cd.?

Thank You,
Rohit.

pszczelaszkov · June 14, 2019, 12:18pm

Hi @ROHIT_SR
Since velocity is Y=-35m/s and roof is on +Z. I setup drag vector at Y=1 and lift vector at Z=1
Which gives me respectively nonspoiler, lip and spoiler Drag coefficients as follows:
-0.94,-1.02,-0.98
Lift coeficients:
0.017,-0.49,-0.29
Pressure forces Z:
11,-367,-215
I think I can trust on lift values since it should work for the whole body anyway.
But drag is messed up, I think it’s because of large open space in mesh, and spoiler sharing same face as rest of the body. However I can’t separate spoiler from hull on this geometry, it seems the only way it works is by importing as STL. When I’m trying with STEP, meshing process is ignoring open bodies leaving me with decals. I think i will recreate geometry from a blueprint as a more solid block, maybe then STEP will work and let me separate drag modifiers from rest of the body.
At this point, if I’m about to believe in lift values it seems like lip spoiler more likely helps push body towards ground making it more stable at high velocities. But I’m still curious about drag changes just for sake of clarity of results.
Please take a look at the modification section:
http://www.iosrjournals.org/iosr-jmce/papers/vol13-issue3/Version-1/R130301114118.pdf
There is something im missing, maybe its a brain
I think i will come back with new geometry.
Cheers,
pszczelaszkov

pszczelaszkov · July 5, 2019, 1:10am

Hi @ROHIT_SR
I recreated from blueprint new geometry, this one is watertight and I’m finally able to import it as multiple faces.
I also created some more scenarios.
Which are:
-Nonspoiler
-Lip
-Stockspoiler
-Frickinwall(Nascar like)
-Trapezlip(Similar to one from the document I posted above, but slightly moved towards the glass.)
-Dashspoiler(Small dash which makes rear end longer)
-Windowdash(Just like above but on the glass behind roof)
-Windowdash&lip(Just a combination of two from above)
-Bigspoiler(Something like Toyota Supra)

The plots are configured as follows.
Forces:

I’ll tried to configure Cd properly as positive values, but when I paired it with forces plots it makes sense for me even in negative so for sake of comparision i left it as it is.
Which is
-Reference length 5m
-Reference Area 0.5m^2
-Drag direction on positive Y, velocity is (-35)
-Lift direction on positive Z, which should be upwards
Selected faces excludes drag modifiers(dash is only modifier included). Since in theory spoilers shouldn’t even be hit by main flow, differences were marginal. However bigspoiler shows decreased uplift when included, but its high and is in path of main flow.

Pressure planes looks as follows.
1.Nonspoiler

2.Lip

3.Spoiler

4.Frickinwall

5.Trapezlip

6.Dash

7.Windowdash

8.Windowdash&lip

9.Bigspoiler

Data from plots looks as follows:

Name	Cd	Cl	Cl(f)	Cl®	ForceY	ForceZ	MomentY	MomentZ
Nonspoiler	-0.32	0.27	0.15	0.11	-102	146	-9	-58
Lip	-0.265	0.02	0.01	0.01	-76	-61	-80	-63
Spoiler	-0.245	0.127	0.07	0.05	-74	56	-36	-60
Frickinwall	-0.345	-0.10	-0.05	-0.04	-112	-133	-101	-41
Trapezlip	-0.25	0.045	0.02	0.02	-77	7	-53	-56
Dashspoiler	-0.26	0.04	0.02	0.02	-81	9	-53	-57
Windowdash	-0.29	0.27	0.13	0.13	-90	121	-15	-61
Windowdash&lip	-0.274	0.03	0.18	0.18	-81	-31	-66	-52
Bigspoiler	-0.32	0.30(0.18)	0.15(0.09)	0.15(0.09)	-102(-105)	146(101)	-12(-21)	-58(-45)

It looks like stock spoiler is some kind of sweet spot between creating a stagnant pocket which slows down air on rear end which as result reduces uplift and creating lesser wake than other cases because of space underneath.
Various additions in form of lips or slightly facing upwards dash, are creating bigger pocket which reduces uplift even further at cost of bigger wake region, although slightly moved lip leaves a clear path on rear edge, which maybe after some adjustment will give similar results to stock spoiler.
I think wall greatly stabilizes rear end in cornering, maybe I made it too big.
Window dash and bigspoiler are in my opinion useless (values in parentheses are after including spoiler faces into account).
I don’t know how to read forces moment, is it in N*m?
What do you think, does it makes sense?
Maybe there is something on your mind which I should include in those tests?
I’m out of ideas for now .
Cheers

jousefm · July 5, 2019, 9:17am

Very detailed description of your project! Let me tag our expert and PowerUser @yosukegb4 here who might give you some additional tips on your setup and the post-processing of forces in general.

Cheers!

Jousef

ROHIT_SR · July 5, 2019, 10:39am

HI @pszczelaszkov
you are working very hard on your design and simulation.
in your, all 9 design non-spoiler and the big spoiler was created more downforce than other its means it also generated more drag which slows down the car. the downforce was necessary for go faster at the corner but here i don’t know what our purpose for this cfd simulation.
Capture

totally agree with you. but here your spoiler generated average downforce its might be good provided average cornering speed with better fuel economy. spoiler change air flow and increases downforce on the rear of the vehicle to help improve traction.
but it’s difficult to predict. @Get_Barried

(OPTIONAL)
you can validate your results with your car with any method…?
thank you,
Rohit.

pszczelaszkov · July 5, 2019, 11:47am

Actually thats a little problem too, if i set lift vector upwards, does it means that positive values is pushing rear end up or down? Take a notice that there are negative numbers aswell.

I should be able to recreate non spoiler, spoiler and maybe lip version, but that’s in month or so, since now im preparing car for painting.

People are advertising a lot of miraculous devices which in theory should help save fuel.
But in the same time, many people think that spoilers are just visuals.
I want to find best option which can help save fuel during long highway cruises.
But since its common to feel wobbly like behaviour of rearend during changes of road height, i’m trying to find something that can weaken this effect, without adding additional drag like with usage of airfoil wings.
Also on high speed cornering my rear axle have tendency to slip, i’m trying to improve that aswell, but thats bonus task.
Cheers

ROHIT_SR · July 5, 2019, 12:11pm

spoiler split the air for better air flow all around the car with less downforce(because it is not WING).
more downforce means more drag and required more power for the wheel to achieve required speed so, more fuel consumption.

i think you go with less downforce design which ultimately having less drag for with better fuel consumption.
Thank you,
Rohit.

pszczelaszkov · July 5, 2019, 1:03pm

I know that spoilers are not wings I’m not expecting downforce. Isn’t it all about upforce?
I mean since without spoiler(nonspoiler) air is rushing towards bottom, in the same time imho rear end is rushing upwards. Which actually generates drag because it needs energy to be taken from somewhere to perform that up lifting. Isn’t it all about trying to make that rear end at 0 pressure force which somehow trapezlip scenario is closest one? Does it makes sense?
Cheers

yosukegb4 · July 10, 2019, 12:17am

Hi everyone,

I can not keep up with the topic, but I will point out some issues I noticed.

Force and moments & Force and moment coefficients settings
- Probably the drag direction should be [ dx, dy, dz ] = [ 0, -1, 0 ]
- Probably the Pitch axis should be [ x, y, z ] = [ -1, 0, 0 ]
- Put the Center of rotation at the center of the car wheel base
- Reference: Newbie. Forces and moments Help
  - Notice: The values depend on your coordinate system and your car demensions
Mesh
- I think it is better to
  - insert boudary layers on the car surfaces and the floor
  - make mesh smaller around the car
Lift
- Generally, a passenger car sometimes lifts without aero-devices for the downforce

I am sorry if you have already corrected them.

Regards,
Yosuke

yosukegb4 · July 10, 2019, 12:25am

HI @pszczelaszkov ,

The Pitch axis may be [ x, y, z ] = [ 1, 0, 0 ] correctly.
I’m sorry I’m not sure on your model.

Best,
Yosuke

pszczelaszkov · July 10, 2019, 12:30am

Hi @yosukegb4
Thanks, I forgot how rotation works. I’ll make a test with corrected values. By smaller mesh around car, you mean more detailed or refinement box smaller?
So drag direction is direction of velocity? I thought its opposite force.
Cheers

yosukegb4 · July 10, 2019, 1:34am

Hi @pszczelaszkov

Please refer my old f1 car mesh for the refinement.
https://www.simscale.com/workbench/?pid=8042688191781154255&mi=spec%3Af9d9d3cd-573e-409d-b380-4b10c6079831%2Cservice%3AMESHING%2Cstrategy%3A11838&sh=92

Drag is the same direction as the velocity of the flow.

Best,
Yosuke