Mercedes W15

[Vanja #66]

Agreed. Very poor, or low quality racing scanarios combined with the CFD mesh of the car from front to back as it goes through these scenarios it whats I think they behind in.

Imagine the car in roll and yaw through a high speed turn, with all the tyredeflections, tyre squirt, heat from the brakes etc... I bet their mesh is not capturing all that!

Not really mesh, more to do with scenarios, as you call them. I think they are usually called car attitudes. I don't think Merc forgot how to mesh the car.

Floors are trully massive and they generate 75-80% of the overall downforce now. They are also more susceptible to suspension movements generated by bumps, kerbs, roll, pitch and of course they are more suscpetible to yaw losses. Unlike all cars from 1983 to 2021, these floor losses are now bigger compared to optimal performance level and the floor itself is 50% more important now than a few years ago. So some cases of the car on track may now cost 10-15% of overall downforce instead of 3-5% like before.

I believe there are many such cases now, especially with raised floor edges since last year, and I believe teams may now need to validate one single design in about 50-100% more different initial CFD cases before going for further, more detailed cases and finally WT validation before approving a design for manufacture. This consumes a lot of CFD resources and teams need to be smart about using them, especially if there's a geometry model that doesn't offer substantial improvement, or worse. It's easy to fall into a trap of doing 5-6 initial CFD cases and waste an update (or a new car bodywork package) because you hoped corner curting won't go unpunished.

[izzy]

What I was assuming was that Mike Elliott did an aero-guy floor that assumed a certain suspension, and the suspension wasn't there. Now with James Allison I was thinking he's quite good with suspension as well as aero, and is more a whole-car person like Adrian is.

But it looks like their wind tunnel, modelling and all that, didn't predict what's happening on the actual track. But don't Aston use the same tunnel still? As well as the same rear suspension geometry. And even Lance was alright in that.

Looking on f1tempo, Lewis seemed to have consistently lower minimum and higher maximum speeds than for example LeClerc, and on T8 he was off the throttle more even than Lance. T8 with the change of direction was much the worst, so, I don't know but I'm thinking it's fixable even if it's not a Red Bull class car.

[maygun]

What I was assuming was that Mike Elliott did an aero-guy floor that assumed a certain suspension, and the suspension wasn't there. Now with James Allison I was thinking he's quite good with suspension as well as aero, and is more a whole-car person like Adrian is.

But it looks like their wind tunnel, modelling and all that, didn't predict what's happening on the actual track. But don't Aston use the same tunnel still? As well as the same rear suspension geometry. And even Lance was alright in that.

Looking on f1tempo, Lewis seemed to have consistently lower minimum and higher maximum speeds than for example LeClerc, and on T8 he was off the throttle more even than Lance. T8 with the change of direction was much the worst, so, I don't know but I'm thinking it's fixable even if it's not a Red Bull class car.

I think the consensus is that you cannot model the floor performance very well in the wind tunnel and need different/improved CFD tools to design efficient floors. Probably Fallows transferred some of the knowledge on how to get on top of this issue from RB to Aston but not the finalised improved version.

[denyall]

Assuming they could build a rig, would they be able to test the floor as it progresses through a corner by having the model rotate through an imaginary apex as windspeed drops/increases? Or are they only allowed static models?

[Matt2725]

Assuming they could build a rig, would they be able to test the floor as it progresses through a corner by having the model rotate through an imaginary apex as windspeed drops/increases? Or are they only allowed static models?

That would be very illegal.

[izzy]

I think the consensus is that you cannot model the floor performance very well in the wind tunnel and need different/improved CFD tools to design efficient floors. Probably Fallows transferred some of the knowledge on how to get on top of this issue from RB to Aston but not the finalised improved version.

Oh Dan Fallows yes. Well if he's the difference, then hopefully the aero nerds at Mercedes can work it out with the same gearbox.

[PlatinumZealot]

Agreed. Very poor, or low quality racing scanarios combined with the CFD mesh of the car from front to back as it goes through these scenarios it whats I think they behind in.

Imagine the car in roll and yaw through a high speed turn, with all the tyredeflections, tyre squirt, heat from the brakes etc... I bet their mesh is not capturing all that!

Not really mesh, more to do with scenarios, as you call them. I think they are usually called car attitudes. I don't think Merc forgot how to mesh the car.

Might be over my head, but I recall that you can have different types of physics in superimposed meshes... lets say you have an CFD mesh, then you can layer a radiation mesh in there, and a heat transfer mesh, and perhaps some other newly invented mesh that RedBull created... Sounds like an enormous computing task, but just something that comes to my mind as being used in F1 now? What do you think about this?

Floors are trully massive and they generate 75-80% of the overall downforce now. They are also more susceptible to suspension movements generated by bumps, kerbs, roll, pitch and of course they are more suscpetible to yaw losses. Unlike all cars from 1983 to 2021, these floor losses are now bigger compared to optimal performance level and the floor itself is 50% more important now than a few years ago. So some cases of the car on track may now cost 10-15% of overall downforce instead of 3-5% like before.

Hmm possible they aren't mimmicking the floor stiffness properly throughout CFD to Windtunnel, to race car... is it really that sensitive though?

I believe there are many such cases now, especially with raised floor edges since last year, and I believe teams may now need to validate one single design in about 50-100% more different initial CFD cases before going for further, more detailed cases and finally WT validation before approving a design for manufacture. This consumes a lot of CFD resources and teams need to be smart about using them, especially if there's a geometry model that doesn't offer substantial improvement, or worse. It's easy to fall into a trap of doing 5-6 initial CFD cases and waste an update (or a new car bodywork package) because you hoped corner curting won't go unpunished.

Only the best setup rigs seems to be reaping benefits of a balanced car. I can tell Aston, Ferrari, RedBull are the most balanced cars by the looks of things. McLaren is a bit better than Mercedes...

[PlatinumZealot]

Assuming they could build a rig, would they be able to test the floor as it progresses through a corner by having the model rotate through an imaginary apex as windspeed drops/increases? Or are they only allowed static models?

Yes they can but it would not be as accurate as their windtunnel model! I believe the windtunnel model has all those features you mentioned. Yaw is only up to a certain point though.

[Vanja #66]

Might be over my head, but I recall that you can have different types of physics in superimposed meshes... lets say you have an CFD mesh, then you can layer a radiation mesh in there, and a heat transfer mesh, and perhaps some other newly invented mesh that RedBull created... Sounds like an enormous computing task, but just something that comes to my mind as being used in F1 now? What do you think about this?

What I know is you can have all the physics regarding fluids (statics, dynamics, wall motion, heat transfer, compressibility, etc) taken into account and calculated within the same simulation. I'm not an expert in CFD overall, I use it for external aero mostly, but I haven't heard about superimposed meshing for those uses. Fluid Structure Interaction yes, but regular fluid physics not really. Which doesn't mean it's not used, just that I haven't heard about that

Hmm possible they aren't mimmicking the floor stiffness properly throughout CFD to Windtunnel, to race car... is it really that sensitive though?

That might have been an issue in early 2022 and especially with big exposed zeropod-floor, but I don't think floor stiffness is an issue since W14, and especially W14B. I still believe they had this issue with floor stiffness in 2022 and also Ferrari in early 2023. It's not coincidence stays have evolved from cables (tension only, so floors were free to flap around a bit) to bars (no flapping anywhere).

Only the best setup rigs seems to be reaping benefits of a balanced car. I can tell Aston, Ferrari, RedBull are the most balanced cars by the looks of things. McLaren is a bit better than Mercedes...

Remember, Ferrari was having issues with rear floor instability last year. Even with final, Japan-spec floor, although it was a lot better. This made them unable to exploit a well balanced car and it was always inherently understeery. Aston was not well balanced later in the year, seems to be a bit better now but at the expense of tyre deg - which suggests to me they may still have some minor floor instability which hurts them in races.

This is by no means an easy task, but the issue with Merc is they are failing to solve the floor conundrum for 3 straight years (counting early 2022 development started this time in 2021 already). Like I said, I think they will need to spend some time with correlation and establish better design verification procedures - and without any sandbagging, you can't do that with these cars - in a few weekends this season, otherwise they might carry over this issue all the way to 2026 and beyond.

[Martin Keene]

What's peculiar about Mercedes is that they have undergone a rather significant philosophy shift from the W13 and W14, and yet the high speed weakness + bouncing is still haunting them.

This hints towards a deeper fundamental issue that somehow Mercedes is missing the third year in a row. It just seems to me, that from a conceptual point of view, their aerodynamics is not flawed in itself since we've seen RedBull incorporate design elements of the W14 into their own car and see benefits from it. But there must be something about the Mercedes' platform which is not allowing them to exploit the full potential they see from their design in CFD / wind tunnel.

We've had Ferrari sticking to their own suspension configuration, with the technical team there claiming suspension design is overrated, words which I must admit, I thought would come to bite them in the ass. However they have kind of proven their point with how that car is performing.

That's why I don't think the suspension configuration is to blame for whatever is troubling Mercedes at the moment. It's bizarre how this car was losing ~0.6s to McLaren in S1 at Jeddah. That's an enormous amount of laptime which you can't attribute to wing/downforce levels alone.

I think it is their suspension. But not in the layout, but how the suspension can control the platform. It is worth remembering that Mercedes were the leaders in interconnected suspension, intnerters, trick hydraulics, etc. It seems like they have forgotten the basics in suspension. I believe they have got access to enough downforce, they just can't control it when they get it.

Of greater concern is they are still talking about correlation issues in the third year of the rules!

[izzy]

I think it is their suspension. But not in the layout, but how the suspension can control the platform. It is worth remembering that Mercedes were the leaders in interconnected suspension, intnerters, trick hydraulics, etc. It seems like they have forgotten the basics in suspension. I believe they have got access to enough downforce, they just can't control it when they get it.

Of greater concern is they are still talking about correlation issues in the third year of the rules!

They seem to be saying they have low drag but not much downforce in the high speed. So they only have enough downforce when they're porpoising, basically, too low to the ground. And then yes suspension, that they're running stiff for that reason but then it's edgy and the drivers can't trust the back end, as in T8 Jeddah.

[ispano6]

[PlatinumZealot]

They havw not forgotten the basics. This is not about the basics anymore! This is about knowing what the suspension needs to be doing in a veeeerry special scenario. It's about insight and ingenuity ON BOTH suspension and area. Mercedes is lacking that since Aldo costa left.

[organic]

They havw not forgotten the basics. This is not about the basics anymore! This is about knowing what the suspension needs to be doing in a veeeerry special scenario. It's about insight and ingenuity ON BOTH suspension and area. Mercedes is lacking that since Aldo costa left.

It seems everyone except mercedes can get this working though. AMR, McLaren, Ferrari, RB. Even Haas and Sauber have better ride than Mercedes

[F1ern]

Isn't Aston using the same suspension at the back as Mercedes?