2017 front wings downforce compared to 2010

[Dipesh1995]

Ok so I'll first run the wing in isolation to make sure there's no fundamental issues with it after which I'll run it with a rotating wheel in a set up identical to what Pegrum used. One question I have is that a rotating wheel is fundamentally an unsteady problem however would I be able to get away with running with a steady solver like CD-adapco have done for their FSAE car even though they're running with rotating wheels?
@ 6.00 mins approx and 10.55 mins : https://www.youtube.com/watch?v=f1UL4o9rdJs

If not, would a implicit solver be better than a explicit solver for this problem? I'm leaning towards implicit as I can run with larger time steps and tends to be more stable.

I would say a steady solver is fine, Vyssion may disagree. To run a properly transient case would need to run without a symmetry plane, doubling your cell count. My philosophy with CFD (certainly at university level) is to run what your machine is capable of running, if your cell count isn't as high as you'd like or your solver is a compromise, if you can critically explain the results then thats fine. Ultimately CFD is just a simulation, there are studies from Chinese facilities running tens of billions of cells and the accuracy (to a wind tunnel simulation) was not really any better, drag was better predicted but lift was worse or whatever.

Yeah I agree, running steady won’t give the most accurate answer but it should give a good indication of the wing’s performance.

[Vyssion]

Nope not yet. I'll hopefully be able to do that after I do the ride height sensitivity study. At the moment, I'm just working on getting spec 2.1 ready. Like I said, I reckon there's still more performance to unlock with this wing.

You may find yourself unlocking that performance in isolation then losing it with the presence of the wheel. The wing and wheel have opposite circulation, so the wheel will work to cancel the circulation of the wing and vice versa, considering your wings span this will be especially prevalent towards the edge vortex you're working so hard to develop. This is also very true with a ride height study, different ride heights will change the way the wing/wheel system interacts, with positive or negative results. Likewise the gap between the trailing edge of your wing and the front point of your tyre - in F1 this is fixed by the legality boxes not optimal performance.

Check out theses and papers from Pegrum, van den Berg, Heyder-Bruckner, Diasinos, and Kellar to see how the wing cannot be considered without a wheel. I know Pegrum is now very senior at McLaren, Diasinos worked for Toyota when they were in F1, and the Kellar paper included Pearse who I think is now head of aero performance at Mercedes so they're top sources. I would also try to use a more representative tyre profile, check out Sprot's SAE paper which has scans with coordinates of the contact patch and upper deformation of a F1 tyre on a 13" rim.

^ This... For pretty much the same reason that jjn9128 suggested you extend your nose cone much further downstream. The shear "presence" of the wheel will affect the aerodynamics both downstream AND upstream (perhaps not all that much for a little winglet or something, but definitely a fair bit for a wheel; which is THE biggest annoyance of mine to mesh and simulate)

I would say a steady solver is fine, Vyssion may disagree. To run a properly transient case would need to run without a symmetry plane, doubling your cell count. My philosophy with CFD (certainly at university level) is to run what your machine is capable of running, if your cell count isn't as high as you'd like or your solver is a compromise, if you can critically explain the results then thats fine. Ultimately CFD is just a simulation, there are studies from Chinese facilities running tens of billions of cells and the accuracy (to a wind tunnel simulation) was not really any better, drag was better predicted but lift was worse or whatever.

Nah, agree - at least for this level of simulation we are going for; KISS = Keep It Simple Stupid

[Dipesh1995]

Ok so I'll run a mesh sensitivity study for the rotating wheel in isolation once I get round to it before combining it with wing and nose setup; atm, I'm just working on spec 2.1.

[Dipesh1995]

So I've done a mesh sensitivity study for a rotating wheel. Dimensions of the wheel are 630mm diameter and 289 mm width, and the radius on the edges of the wheel is 31.5mm. It is sunk into the ground plane by 20mm. It is therefore a bit smaller than a F1 tyre but the car is designed around those wheel dimensions so I'm not going to change them.

https://goo.gl/photos/8BMqeLA49XvwZygt8

I've included mesh scenes (Y+ is between 25 and 75 for nearly all of the wheel), a contour plot showing the flow around the wheel (trailing region of the wheel looks like an angry octopus ), isosurface plots with Q criterion (10000 /s^2) which show the two vortices forming from the edges of the wheel, and a vector plot showing showing the wake of the wheel (20 cm from the wheel).

Drag of the wheel 222 N and lift is 227 N.

Hopefully, I'll be able to put the wing and wheel together later on this week or next week which will show the true colours of the wing to a large enough extent.

[Dipesh1995]

Ok so I've ran the wing with rotating wheels behind it in a similar set up Pegrum used. I've not added suspension to keep it relatively simple for now.

https://goo.gl/photos/T97xMJejQsTH8G4eA

Isosurface messiness due to quite low isosurface value of 250000 /s^2.

Unexpectedly, my edge vortex seems to prefer to go round the inside of the wheel than the outside of the wheel. I'm sure that would change when the car is in yaw (something I could check later). I think McLaren had a similar trait with their wing back in 2014 hence they came up with this solution for their vanes: http://imgur.com/GA5woRp.

In terms of force values, my downforce has been reduced from 3871N (spec 2.1 value) to 3676N and my drag has also reduced from to 727 N to 640 N (significant reduction in pressure drag) leading to surprising wing efficiency increase (L/D) from 5.32 to 5.74.

I'm going to be doing a ride height sensitivity study to find a optimal range of ride heights for the wing.

[jjn9128]

Ok so I've ran the wing with rotating wheels behind it in a similar set up Pegrum used. I've not added suspension to keep it relatively simple for now.

https://goo.gl/photos/T97xMJejQsTH8G4eA

Isosurface messiness due to quite low isosurface value of 250000 /s^2.

Unexpectedly, my edge vortex seems to prefer to go round the inside of the wheel than the outside of the wheel. I'm sure that would change when the car is in yaw (something I could check later). I think McLaren had a similar trait with their wing back in 2014 hence they came up with this solution for their vanes: http://imgur.com/GA5woRp.

In terms of force values, my downforce has been reduced from 3871N (spec 2.1 value) to 3676N and my drag has also reduced from to 727 N to 640 N (significant reduction in pressure drag) leading to surprising wing efficiency increase (L/D) from 5.32 to 5.74.

I'm going to be doing a ride height sensitivity study to find a optimal range of ride heights for the wing.

I would say you still need to boat-tail the back end of your 'chassis', though maybe the images are clipping off the rear of the geometry.

I don't find the ultimate direction of your edge vortex all that surprising. The purpose of those tunnels as I understand them is more to build up the vortex to burst on the front face of the tyre, reducing drag on the wheel, while also getting a bit of downforce. The turning of the flow comes more from the upper surface as you have. IMO you need to change your tyre geometry to be more realistic, this will affect the forces and wake direction. It doesn't necessarily need spokes, brake ducts and blown axles but a more realistic tyre and contact patch - with camber. Check out papers by Saddington and Knowles to see how much camber can affect the wake of a wheel. I would also generally look at total/stagnation pressure in the wake between 0<Cpo<1 or -0.5<Cpo<1 rather than velocity.

Those are forces just on the wing? What about the wheel? Did you see a drop of drag and lift? Why do you think you lost some drag from the wing?

[Vyssion]

I don't find the ultimate direction of your edge vortex all that surprising. The purpose of those tunnels as I understand them is more to build up the vortex to burst on the front face of the tyre, reducing drag on the wheel, while also getting a bit of downforce. The turning of the flow comes more from the upper surface as you have.

Pretty much, at least to my knowledge as well. Here is a good example picture which shows you two slices before and aft of the wheel showing Total Pressure (a.k.a. lossely, the amount of energy in the flow). You can see the main large tunnel vortex heading straight into pretty much the centreline of the tyre. You can I'd say that in that image, the vortex travels outboard, with the three smaller vane-generated vortices traveling inboard around the wheel. Air is a fickle thing, and so sometimes in order for you to divert it where you want it, you REALLY have to give it a violent kick with a flap element or endplate feature.

IMO you need to change your tyre geometry to be more realistic, this will affect the forces and wake direction. It doesn't necessarily need spokes, brake ducts and blown axles but a more realistic tyre and contact patch - with camber. Check out papers by Saddington and Knowles to see how much camber can affect the wake of a wheel.

I think that if you had some sort of simple detailed wheel like in the picture above, you'd prety much be okay for what youre trying to go for. Camber is the big issue though as jjn9128 said - if you don't want to bother with it, at least take a look at those papers so you get some sense of what you may be missing.

Here is a direct link to his paper on near wake structure, which also has his two other papers as appendices; both of which reference camber:
https://www.google.co.uk/url?sa=t&rct=j ... fDRBBZlYvg

One other little tid-bit is that your wing looks to be ever so slightly wider than your wheels - not the case for 2017 - but *shrug* if thats what you wanted, all's good!!

[Dipesh1995]

Ok so I've ran the wing with rotating wheels behind it in a similar set up Pegrum used. I've not added suspension to keep it relatively simple for now.

https://goo.gl/photos/T97xMJejQsTH8G4eA

Isosurface messiness due to quite low isosurface value of 250000 /s^2.

Unexpectedly, my edge vortex seems to prefer to go round the inside of the wheel than the outside of the wheel. I'm sure that would change when the car is in yaw (something I could check later). I think McLaren had a similar trait with their wing back in 2014 hence they came up with this solution for their vanes: http://imgur.com/GA5woRp.

In terms of force values, my downforce has been reduced from 3871N (spec 2.1 value) to 3676N and my drag has also reduced from to 727 N to 640 N (significant reduction in pressure drag) leading to surprising wing efficiency increase (L/D) from 5.32 to 5.74.

I'm going to be doing a ride height sensitivity study to find a optimal range of ride heights for the wing.

Those are forces just on the wing? What about the wheel? Did you see a drop of drag and lift? Why do you think you lost some drag from the wing?

Yep those are the forces just on the wing. The drag of the wheel reduced from 222N to 116N and the lift reduced from 227N to 134N. The best explanation I can come up with for the reduction in drag is that within the wing's vortex tunnel, the "diffuser" effect is reduced i.e the adverse pressure gradient is not of the same magnitude with the rotating wheel as it was without it. Essentially what is happening is that flow in the pressure recovery region of the wing is better extracted by the rotating wheel hence the flow's kinetic energy in that region never reduces as much as it did without the wheel (the velocity of the flow no longer stagnates or reverses; dynamic pressure to static pressure is not as severe as before). Therefore, this reduces pressure drag. Basically, this no longer occurs: https://goo.gl/photos/QkLEhbBD6wFaQCet6

This also leads to a loss in downforce as the peak suction region of the underside of the wing is not worked as hard as the adverse pressure gradient isn't as strong as it was before.

[Dipesh1995]

I don't find the ultimate direction of your edge vortex all that surprising. The purpose of those tunnels as I understand them is more to build up the vortex to burst on the front face of the tyre, reducing drag on the wheel, while also getting a bit of downforce. The turning of the flow comes more from the upper surface as you have.

Pretty much, at least to my knowledge as well. Here is a good example picture which shows you two slices before and aft of the wheel showing Total Pressure (a.k.a. lossely, the amount of energy in the flow). You can see the main large tunnel vortex heading straight into pretty much the centreline of the tyre. You can I'd say that in that image, the vortex travels outboard, with the three smaller vane-generated vortices traveling inboard around the wheel. Air is a fickle thing, and so sometimes in order for you to divert it where you want it, you REALLY have to give it a violent kick with a flap element or endplate feature.
http://i.imgur.com/YuP5pYO.png

IMO you need to change your tyre geometry to be more realistic, this will affect the forces and wake direction. It doesn't necessarily need spokes, brake ducts and blown axles but a more realistic tyre and contact patch - with camber. Check out papers by Saddington and Knowles to see how much camber can affect the wake of a wheel.

I think that if you had some sort of simple detailed wheel like in the picture above, you'd prety much be okay for what youre trying to go for. Camber is the big issue though as jjn9128 said - if you don't want to bother with it, at least take a look at those papers so you get some sense of what you may be missing.

Here is a direct link to his paper on near wake structure, which also has his two other papers as appendices; both of which reference camber:
https://www.google.co.uk/url?sa=t&rct=j ... fDRBBZlYvg

One other little tid-bit is that your wing looks to be ever so slightly wider than your wheels - not the case for 2017 - but *shrug* if thats what you wanted, all's good!!

Yep I'll try a few ideas to the get the vortex to go around the outside of the front wheel. In all honesty, I don't think having the vortex go around the inside of the front wheel (like back in the old days) is such as bad thing, I think it would work well depending on the design of the turning vanes and barge boards.

Thanks for the link, I'll include a camber of -3.5 degrees when I do my next CFD run. I think it'll have an effect on the direction of the edge vortex hopefully persuading to go outside the front wheels.

I am maximizing the span of the wing with outer front track so yeah the span of the wing and outer width are pretty identical.

[jjn9128]

Ok so I've ran the wing with rotating wheels behind it in a similar set up Pegrum used. I've not added suspension to keep it relatively simple for now.

https://goo.gl/photos/T97xMJejQsTH8G4eA

Isosurface messiness due to quite low isosurface value of 250000 /s^2.

Unexpectedly, my edge vortex seems to prefer to go round the inside of the wheel than the outside of the wheel. I'm sure that would change when the car is in yaw (something I could check later). I think McLaren had a similar trait with their wing back in 2014 hence they came up with this solution for their vanes: http://imgur.com/GA5woRp.

In terms of force values, my downforce has been reduced from 3871N (spec 2.1 value) to 3676N and my drag has also reduced from to 727 N to 640 N (significant reduction in pressure drag) leading to surprising wing efficiency increase (L/D) from 5.32 to 5.74.

I'm going to be doing a ride height sensitivity study to find a optimal range of ride heights for the wing.

Those are forces just on the wing? What about the wheel? Did you see a drop of drag and lift? Why do you think you lost some drag from the wing?

Yep those are the forces just on the wing. The drag of the wheel reduced from 222N to 116N and the lift reduced from 227N to 134N. The best explanation I can come up with for the reduction in drag is that within the wing's vortex tunnel, the "diffuser" effect is reduced i.e the adverse pressure gradient is not of the same magnitude with the rotating wheel as it was without it. Essentially what is happening is that flow in the pressure recovery region of the wing is better extracted by the rotating wheel hence the flow's kinetic energy in that region never reduces as much as it did without the wheel (the velocity of the flow no longer stagnates or reverses; dynamic pressure to static pressure is not as severe as before). Therefore, this reduces pressure drag. Basically, this no longer occurs: https://goo.gl/photos/QkLEhbBD6wFaQCet6

This also leads to a loss in downforce as the peak suction region of the underside of the wing is not worked as hard as the adverse pressure gradient isn't as strong as it was before.

Good that your wheel forces are also reduced. I think the main explanation is a bit more simple than you suggest. The wheel, being bluff, has a high positive pressure field ahead which impinges on the negative pressure region behind the wing, increasing the pressure under the wing. In the same way the negative pressure under and behind the wing impinges on the positive pressure ahead of the wheel reducing the wheels drag. As you say this will probably reduce the adverse pressure gradient, so you may not get flow reversal in the wake, there may also be a blockage effect from the wheel. You also lost some downforce on the wing, I wonder if there is a way to show the effect of the wheel on the wings surface pressures and if that would help to determine the cause of the force loss.

It's good to start thinking about these things, especially when your result is surprising as with your improved L/D, it's what separates an average study from a good one.

[Dipesh1995]

Those are forces just on the wing? What about the wheel? Did you see a drop of drag and lift? Why do you think you lost some drag from the wing?

Yep those are the forces just on the wing. The drag of the wheel reduced from 222N to 116N and the lift reduced from 227N to 134N. The best explanation I can come up with for the reduction in drag is that within the wing's vortex tunnel, the "diffuser" effect is reduced i.e the adverse pressure gradient is not of the same magnitude with the rotating wheel as it was without it. Essentially what is happening is that flow in the pressure recovery region of the wing is better extracted by the rotating wheel hence the flow's kinetic energy in that region never reduces as much as it did without the wheel (the velocity of the flow no longer stagnates or reverses; dynamic pressure to static pressure is not as severe as before). Therefore, this reduces pressure drag. Basically, this no longer occurs: https://goo.gl/photos/QkLEhbBD6wFaQCet6

This also leads to a loss in downforce as the peak suction region of the underside of the wing is not worked as hard as the adverse pressure gradient isn't as strong as it was before.

Good that your wheel forces are also reduced. I think the main explanation is a bit more simple than you suggest. The wheel, being bluff, has a high positive pressure field ahead which impinges on the negative pressure region behind the wing, increasing the pressure under the wing. In the same way the negative pressure under and behind the wing impinges on the positive pressure ahead of the wheel reducing the wheels drag. As you say this will probably reduce the adverse pressure gradient, so you may not get flow reversal in the wake, there may also be a blockage effect from the wheel. You also lost some downforce on the wing, I wonder if there is a way to show the effect of the wheel on the wings surface pressures and if that would help to determine the cause of the force loss.

It's good to start thinking about these things, especially when your result is surprising as with your improved L/D, it's what separates an average study from a good one.

Yes you are correct, having looked at the pressure contours, the pressure in the pressure recovery region of the wing is greater with the wheel than without due to the reason you have given. I'll upload the pressure contours and velocity vector slice through the vortex tunnel comparison tomorrow or the day after once I've completed the ride height sensitivity.

[Dipesh1995]

https://goo.gl/photos/soHBtQSPoWXriA2u9

So here are the pressure contours and velocity vector slices. Key differences are the higher pressure in pressure recovery region with wheel compared to without it and lack of flow reversal leading to reduced drag and downforce.

I've done the ride height sensitivity and the height of 30mm between the ground and lowest element of the wing is the optimum in terms of wing efficiency and downforce although that is with a weak edge vortex.

h=50mm Downforce: 3676 N Drag: 640 N L/D: 5.74
h=40mm Downforce: 3922 N Drag: 641 N L/D: 6.12
h=35mm Downforce: 4086 N Drag: 644 N L/D: 6.34
h=30mm Downforce: 4153 N Drag: 638 N L/D: 6.51 (Weak edge vortex, terminates as soon it leaves the vortex tunnel)
h=25mm Downforce: 3372 N Drag: 511 N L/D: 6.60 (Burst edge vortex)

I'll continue to test at h=50mm knowing that the wing has a bit more in reserve.

[Just_a_fan]

It's interesting to hear you say that the results are best at 30mm but with a weak edge vortex. The interaction of the front wing's flow structures with the rest of the car is key to the overall performance of the car. Knowing that you get great results at a low ride height for the wing but with likely compromised performance downstream is a good thing to look at too, it seems to me, even if just as a discussion point.The loss of the vortex from the front wing will be something that would be avoided on a real car, I would suggest. Also, knowing that reduced ride height causes the vortex to break down, and knowing this would affect, detrimentally, the performance of the rest of the car, we can perhaps see why RedBull worked so hard to droop the front wings in to the ground. I bet there was a big overall drag benefit from doing that.

[Dipesh1995]

It's interesting to hear you say that the results are best at 30mm but with a weak edge vortex. The interaction of the front wing's flow structures with the rest of the car is key to the overall performance of the car. Knowing that you get great results at a low ride height for the wing but with likely compromised performance downstream is a good thing to look at too, it seems to me, even if just as a discussion point.The loss of the vortex from the front wing will be something that would be avoided on a real car, I would suggest. Also, knowing that reduced ride height causes the vortex to break down, and knowing this would affect, detrimentally, the performance of the rest of the car, we can perhaps see why RedBull worked so hard to droop the front wings in to the ground. I bet there was a big overall drag benefit from doing that.

For sure, running the wing at h=30mm would be detrimental to the performance of the aero components downstream of the wing; running the wing at h=35mm would be significantly better in terms of overall car performance. Obviously, in reality it is near impossible to keep the ride height of the front wing constant especially in the corners so hypothetically speaking, if this car was racing and the CFD simulation was 100% trusted (wing hasn't been tested when in yaw), I would set the static ride height of the front wing to around h=45 mm to retain a operating window for the wing without the wing stalling in the corners. The loss of the vortex system would be definitely be avoided by teams in the corners as its critical in terms diverting the front wheel wake away from the leading edge of the floor and sidepods.

I was just considering the standalone wing performance in terms of efficiency and force values where h=30mm is the optimum ride height, perhaps I should be clearer in terms of what I meant.

The Red Bull flexible front wings was a very effective solution since it gave two benefits from one technique. The outboard part of the wing flexed enough in the medium-high speed corners to increase front downforce (also kept the edge vortex at the desired strength) whereas down long straights, the wing flexed further to the point where the outboard region the wing effectively stalled hence reducing downforce and more importantly, drag. This was actually confirmed to me by a Red Bull aerodynamicist when I was invited to one of their assessment centres. If I remember correctly, one of GA's analysis of the front wing, he got the part about the increased downforce but missed the part about the reduced drag. Nowadays, teams struggle to fully replicate that function due to the increased stiffness required by the FIA tests.

[Dipesh1995]

Ok so I’ve spent the last couple of weeks trying to recover the downforce that I had lost from the rotating wheels. I’ve more or less done that and now the downforce is 3859 N and drag is 651 N so L/D has increased to 5.93.

I also tested the wing with camber of -3.5 degrees with a more realistic tyre profile and downforce increased to 3863 N and drag increased to 654 N reducing L/D to 5.91.

https://photos.app.goo.gl/by2qzyNTK2rIIFL32

However, I’ve come across a problem with the Y250 vortex where its core loses energy after which it bursts before a new weaker vortex forms again (Isovalue: 10000/s). It hasn’t happened with my previous simulations so I have no idea why it’s happened now. I haven’t changed the inboard tips of the wing, it happens with both tyre profiles and the Y250 vortex is produced exactly in the same way it did before from the inboard tips.

The Y250 vortex is an anti-clockwise vortex, the edge vortex is a clockwise vortex and the nose pylon vortex is also clockwise. Does a counter rotating vortex pair strengthen each other the closer they get or does the weaker of the two sap energy from the stronger one in an attempt to cause an angular momentum equilibrium and thus net momentum becomes 0?

I’ve read different things from different people so now I’m a bit confused.

I’m so close to getting this thing where I wanted it to be but now I've come across this