2017 front wings downforce compared to 2010

[jjn9128]

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.

Welcome back. Looks like you've been busy. When you say it happens with both tyre profiles you mean with and without camber? Or with a simple cylinder and with the wheel in the pics both with camber? It's difficult to tell what's happening from 1 slice - I don't know if it's possible to create a video/series of slices like this?


What I will say though is this, why do you think it is that F1 teams work so hard to develop such intricate lift producing devices under the nose...

[Dipesh1995]

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.

When you say it happens with both tyre profiles you mean with and without camber? Or with a simple cylinder and with the wheel in the pics both with camber?

Yeah so it happens with both the cylinder tyre and the realistic tyre with and without camber. So essentially all 4 simulations.

It's difficult to tell what's happening from 1 slice - I don't know if it's possible to create a video/series of slices like this?
https://www.youtube.com/watch?v=ZGvYxlih1r4

I'll create a animation as soon as I get back to Uni early next week. I don't have STAR 12 at home which was used for the simulation, I just have STAR 11 so I can't view simulation right now.

What I will say though is this, why do you think it is that F1 teams work so hard to develop such intricate lift producing devices under the nose...

I know what you mean but the Y250 vortex is probably the easiest and simplest vortex to create, far easier than the cascade vortex or the edge vortex which seem to be fine. Also, the way the vortex core collapses and somehow another similar vortex is generated from its residue makes me wonder whether that would actually be the case if the model was tested in a wind tunnel.

[Dipesh1995]

I've got a question which I think I already know the answer but would like to hear your thoughts on it.

To simulate my wings, I import a half geometry of the wing i.e a split down the nose cone. I then use a symmetry plane boundary condition to mirror the wing so that I end up with a full wing. This obviously halves my cell count compared to what it would be if I was to import the full wing i.e saving me time etc. However, to obtain my force values, does STAR take into account the symmetry plane and double the values before displaying them or do I need to double them myself?

I think it only takes into account the part that has been meshed so it gives me half the force values. If this is case then my force values become for the entire wing becomes 7718 N (downforce) and 1302 N (drag) which I think is way too much at 150 mph. The coefficients of lift, drag etc would stay the same of course.

[jjn9128]

I've got a question which I think I already know the answer but would like to hear your thoughts on it.

To simulate my wings, I import a half geometry of the wing i.e a split down the nose cone. I then use a symmetry plane boundary condition to mirror the wing so that I end up with a full wing. This obviously halves my cell count compared to what it would be if I was to import the full wing i.e saving me time etc. However, to obtain my force values, does STAR take into account the symmetry plane and double the values before displaying them or do I need to double them myself?

I think it only takes into account the part that has been meshed so it gives me half the force values. If this is case then my force values become for the entire wing becomes 7718 N (downforce) and 1302 N (drag) which I think is way too much at 150 mph. The coefficients of lift, drag etc would stay the same of course.

I think that a symmetry plane in itself does not automatically mirror the solution or forces - it's fairly standard practice to use a symmetry wall for the far walls. However, I think in post-processing you are mirroring the solution - I don't remember star all that well, but as you say the forces make sense. One way to be sure would be to run again with a full wing and 2 wheels and see how the forces stack up.

[Dipesh1995]

I've got a question which I think I already know the answer but would like to hear your thoughts on it.

To simulate my wings, I import a half geometry of the wing i.e a split down the nose cone. I then use a symmetry plane boundary condition to mirror the wing so that I end up with a full wing. This obviously halves my cell count compared to what it would be if I was to import the full wing i.e saving me time etc. However, to obtain my force values, does STAR take into account the symmetry plane and double the values before displaying them or do I need to double them myself?

I think it only takes into account the part that has been meshed so it gives me half the force values. If this is case then my force values become for the entire wing becomes 7718 N (downforce) and 1302 N (drag) which I think is way too much at 150 mph. The coefficients of lift, drag etc would stay the same of course.

I think that a symmetry plane in itself does not automatically mirror the solution or forces - it's fairly standard practice to use a symmetry wall for the far walls. However, I think in post-processing you are mirroring the solution - I don't remember star all that well, but as you say the forces make sense. One way to be sure would be to run again with a full wing and 2 wheels and see how the forces stack up.

I’ve simulated a single element wing using a symmetry plane and the full wing and yes I do have to double the forces when using the symmetry plane meaning my wing apparently does generate those forces. The mesh is fine and convergence is good so either this simulation is correct or there’s something else going on.

[n_anirudh]

Symmetry condition may only dampen the lateral forces/wake in an unsteady simulation.

If steady sims are running, my not affect a lot.

[Dipesh1995]

It's difficult to tell what's happening from 1 slice - I don't know if it's possible to create a video/series of slices like this?
https://www.youtube.com/watch?v=ZGvYxlih1r4

https://photos.app.goo.gl/Rx5hov9RVmGw1nfg2

That's the best I could do. Click on the video for it play in higher resolution if viewing on a PC. Vortex core collapse starts towards the start of rear wheel wheel so it looks like its the wheel wake that is causing it? Although, this wasn't happening before.

[jjn9128]

It's difficult to tell what's happening from 1 slice - I don't know if it's possible to create a video/series of slices like this?
https://www.youtube.com/watch?v=ZGvYxlih1r4

https://photos.app.goo.gl/Rx5hov9RVmGw1nfg2

That's the best I could do. Click on the video for it play in higher resolution if viewing on a PC. Vortex core collapse starts towards the start of rear wheel wheel so it looks like its the wheel wake that is causing it? Although, this wasn't happening before.

Very good! That's what I was after, it would be good to fix your data contour levels so the colour doesn't change slide-to-slide.

That was my first inclination too - the vortex is pulling low velocity flow from the wheel wake inboard to circulate, so it is suddenly losing angular momentum and the core is collapsing. I would still look at total pressure rather than velocity in your slices as the vortex core should show as Cpo<0.

Why it is now happening when it was not before is a puzzle. Is there anything different about this case? have you changed the mesh, the boundary conditions, averaging interval, solver type?? You said there is a difference with what you are using at home vs at uni? I'm clutching at straws here.... 2 identical cases should produce the same result in CFD.

[Dipesh1995]

It's difficult to tell what's happening from 1 slice - I don't know if it's possible to create a video/series of slices like this?
https://www.youtube.com/watch?v=ZGvYxlih1r4

https://photos.app.goo.gl/Rx5hov9RVmGw1nfg2

That's the best I could do. Click on the video for it play in higher resolution if viewing on a PC. Vortex core collapse starts towards the start of rear wheel wheel so it looks like its the wheel wake that is causing it? Although, this wasn't happening before.

Very good! That's what I was after, it would be good to fix your data contour levels so the colour doesn't change slide-to-slide.

That was my first inclination too - the vortex is pulling low velocity flow from the wheel wake inboard to circulate, so it is suddenly losing angular momentum and the core is collapsing. I would still look at total pressure rather than velocity in your slices as the vortex core should show as Cpo<0.

Why it is now happening when it was not before is a puzzle. Is there anything different about this case? have you changed the mesh, the boundary conditions, averaging interval, solver type?? You said there is a difference with what you are using at home vs at uni? I'm clutching at straws here.... 2 identical cases should produce the same result in CFD.

The mesh is the same, the boundary conditions are the same, averaging interval is untouched and the solver is the same so I don't know what has happened. The general flow field apart from the Y250 is approximately the same. I made some minor changes to the wing around the tunnel vortex serrations and the leading two elements of the wing but none of those parts don't really influence the Y250. The vortex leaving the inboard tips is the same. I have STAR 11 at home and STAR 12 at Uni so I can't view STAR 12 simulations at home but that shouldn't influence the simulation results.

I suspect that if the model had bargeboards, this issue would disappear as wheel wake would be pushed outboards although so would the Y250, honestly, I don't really know whether it would solve it. I'm at 8.5 million cells just with this model so its not really feasible running the simulation with them as I would have to run with some of the chassis and bits of the floor and splitter to get a representative flow field for the barge boards.

My biggest issue at the moment is the excessive (I think) downforce I'm getting. I never thought I'd be saying that when I first started simulating wings. I reduced the blockage ratio to 1.28% from 3.36% today for the computational domain (the frontal area of the half-model is 0.32 m^2 and the inlet area of the domain is 25 m^2) hoping that I would get a reasonable reduction in downforce but that hasn't been the case. Downforce is now 3764 N and drag is 630 N for the half-wing so the numbers for the full wing are 7528 N and 1260 N which is again substantial. The problem is that I don't what F1 front wings generate in numbers. I've seen several estimations online ranging from 400 to 600 kg for cars that are more than a decade old so I don't really know. Even the two-element half-wing tested to get the mesh correct was generating 3500 N in half-wing form so 7000 N in total. If I am setting something wrong, I really don't know what .

[jjn9128]

The mesh is the same, the boundary conditions are the same, averaging interval is untouched and the solver is the same so I don't know what has happened. The general flow field apart from the Y250 is approximately the same. I made some minor changes to the wing around the tunnel vortex serrations and the leading two elements of the wing but none of those parts don't really influence the Y250. The vortex leaving the inboard tips is the same. I have STAR 11 at home and STAR 12 at Uni so I can't view STAR 12 simulations at home but that shouldn't influence the simulation results.

I suspect that if the model had bargeboards, this issue would disappear as wheel wake would be pushed outboards although so would the Y250, honestly, I don't really know whether it would solve it. I'm at 8.5 million cells just with this model so its not really feasible running the simulation with them as I would have to run with some of the chassis and bits of the floor and splitter to get a representative flow field for the barge boards.

My biggest issue at the moment is the excessive (I think) downforce I'm getting. I never thought I'd be saying that when I first started simulating wings. I reduced the blockage ratio to 1.28% from 3.36% today for the computational domain (the frontal area of the half-model is 0.32 m^2 and the inlet area of the domain is 25 m^2) hoping that I would get a reasonable reduction in downforce but that hasn't been the case. Downforce is now 3764 N and drag is 630 N for the half-wing so the numbers for the full wing are 7528 N and 1260 N which is again substantial. The problem is that I don't what F1 front wings generate in numbers. I've seen several estimations online ranging from 400 to 600 kg for cars that are more than a decade old so I don't really know. Even the two-element half-wing tested to get the mesh correct was generating 3500 N in half-wing form so 7000 N in total. If I am setting something wrong, I really don't know what .

I would say altering the mainplane wing profile would be the cause of this difference in the Y250, I had assumed the wing was the unchanged.

As I've said before 3.4 to 3.7kN is about where an F1 front wing will be at 240km/hr - it is difficult to find exact numbers to cite as you say, but you can ballpark. Producing more than that with your wing is not infeasible as it's bigger, but double does seem anomalous. I would estimate drag at about 650N too so your half wing values seem correct..

From what you've said I would say ride height may be a factor, the lowest point on a F1 wing is 50mm above the reference plane BUT assuming the car is flat that is a minimum ground clearance of 60mm, as the skid block is lower than the reference plane. At realistic rakes the wing may have a point lower than 50mm (especially now the plank starts later down the car), but the endplates will be angled so you may find the camber of the wing needs reducing. I don't think this accounts for a 3000N discrepancy but maybe it's a start... ?

[Dipesh1995]

The mesh is the same, the boundary conditions are the same, averaging interval is untouched and the solver is the same so I don't know what has happened. The general flow field apart from the Y250 is approximately the same. I made some minor changes to the wing around the tunnel vortex serrations and the leading two elements of the wing but none of those parts don't really influence the Y250. The vortex leaving the inboard tips is the same. I have STAR 11 at home and STAR 12 at Uni so I can't view STAR 12 simulations at home but that shouldn't influence the simulation results.

I suspect that if the model had bargeboards, this issue would disappear as wheel wake would be pushed outboards although so would the Y250, honestly, I don't really know whether it would solve it. I'm at 8.5 million cells just with this model so its not really feasible running the simulation with them as I would have to run with some of the chassis and bits of the floor and splitter to get a representative flow field for the barge boards.

My biggest issue at the moment is the excessive (I think) downforce I'm getting. I never thought I'd be saying that when I first started simulating wings. I reduced the blockage ratio to 1.28% from 3.36% today for the computational domain (the frontal area of the half-model is 0.32 m^2 and the inlet area of the domain is 25 m^2) hoping that I would get a reasonable reduction in downforce but that hasn't been the case. Downforce is now 3764 N and drag is 630 N for the half-wing so the numbers for the full wing are 7528 N and 1260 N which is again substantial. The problem is that I don't what F1 front wings generate in numbers. I've seen several estimations online ranging from 400 to 600 kg for cars that are more than a decade old so I don't really know. Even the two-element half-wing tested to get the mesh correct was generating 3500 N in half-wing form so 7000 N in total. If I am setting something wrong, I really don't know what .

From what you've said I would say ride height may be a factor, the lowest point on a F1 wing is 50mm above the reference plane BUT assuming the car is flat that is a minimum ground clearance of 60mm, as the skid block is lower than the reference plane. At realistic rakes the wing may have a point lower than 50mm (especially now the plank starts later down the car), but the endplates will be angled so you may find the camber of the wing needs reducing. I don't think this accounts for a 3000N discrepancy but maybe it's a start... ?

Hmm but that doesn't explain the excessive numbers I got for the two-element wing. The dimensions can be found on page 3 of this thread.

https://photos.app.goo.gl/s8lYWUda5ThgHBN02

Do the pressure coefficient and velocity numbers look reasonable?
Since it apparently generates 3533 N so the full wing generates 7066 N and was still increasing a little even though the main plane is running 80 mm off the ground.

I'm convinced I'm setting up wrong although I've resorted to following online tutorials yet still get the same nonsensical values.

[jjn9128]

Hmm but that doesn't explain the excessive numbers I got for the two-element wing. The dimensions can be found on page 3 of this thread.

https://photos.app.goo.gl/s8lYWUda5ThgHBN02

Do the pressure coefficient and velocity numbers look reasonable?
Since it apparently generates 3533 N so the full wing generates 7066 N and was still increasing a little even though the main plane is running 80 mm off the ground.

I'm convinced I'm setting up wrong although I've resorted to following online tutorials yet still get the same nonsensical values.

I mean Cp doesn't look ridiculous, the peak -7 at the leading edge I wouldn't say is desirable but it's certainly not unprecedented and definitely makes sense for your wing design. Over the majority of the surface is -1 to -2.5 which is reasonable. You could output a 2-d plot of the pressure distribution and integrate Cl yourself, see if it lines up to what star is telling you.

Could it be that you're including force on the wing twice or force from something else in the setup?! Again clutching at straws without being able to click through the setup myself.

[Just_a_fan]

Small thought about wheel wake polluting the Y250 vortex. The teams run various types of device on the inner face of the hub/brake duct that are obviously flow conditioners directing wheel wake. Perhaps one of these might be beneficial here too?

[Dipesh1995]

Hmm but that doesn't explain the excessive numbers I got for the two-element wing. The dimensions can be found on page 3 of this thread.

https://photos.app.goo.gl/s8lYWUda5ThgHBN02

Do the pressure coefficient and velocity numbers look reasonable?
Since it apparently generates 3533 N so the full wing generates 7066 N and was still increasing a little even though the main plane is running 80 mm off the ground.

I'm convinced I'm setting up wrong although I've resorted to following online tutorials yet still get the same nonsensical values.

Could it be that you're including force on the wing twice or force from something else in the setup?! Again clutching at straws without being able to click through the setup myself.

The simulation is half wing so I have to double the force value it get the value for the entire wing. I’ve only got a wing in setup so I don’t think I’m including the forces from anything else.

[Dipesh1995]

Small thought about wheel wake polluting the Y250 vortex. The teams run various types of device on the inner face of the hub/brake duct that are obviously flow conditioners directing wheel wake. Perhaps one of these might be beneficial here too?

Yeah, I think wheel hubs and barge boards would probably keep the vortex alive. I’m going to run the simulation again to see if it happens again but this time with some greater refinement in the Y250 vortex region. If that doesn’t help, I’ll probably run with wheel hub with some flow conditioning devices.