2024 Scuderia Ferrari F1 Team

[Vanja #66]

https://www.the-race.com/formula-1/mcla ... explained/

Thinking back 12 months ago, when Ferrari was quite openly admitting it wasn’t that fussed that Sanchez was leaving, the summary from various conversations The Race had was: ‘he’s a great technical mind, but he doesn’t play well with others’.

Interesting.

As we've seen last year, Ferrari finally had a major improvement of the chassis during the season since at least 2012. Sanchez arrived in October 2012. Not implying any direct correlation, but something clearly wasn't working as it should. He's also completely credited for Byrne's conceptual ideas for SF70H and F1-75.

[Sphere3758]

Who gets credit for this year’s car concept ? Would it be Cardile or Byrne again ?

[Xyz22]

Who gets credit for this year’s car concept ? Would it be Cardile or Byrne again ?

I don't think it there is a single guy/girl we have to credit for the car.

What i said multiple times is that before there wasn't a clear structure. Who was taking the most important decisions? Was it Binotto? Cardile? Sanchez? We don't know.

What we know is that under Binotto the technical direction was always trying to find the most amount of pure performance without considering other factors (at least on the same importance level).

"Strangely", the first car designed under a different Technical Director (Cardile) is completely different in terms of behaviour.

[Venturiation]

Who gets credit for this year’s car concept ? Would it be Cardile or Byrne again ?

everyone working at scuderia ferrari

[leblanc]

https://www.the-race.com/formula-1/mcla ... explained/

Thinking back 12 months ago, when Ferrari was quite openly admitting it wasn’t that fussed that Sanchez was leaving, the summary from various conversations The Race had was: ‘he’s a great technical mind, but he doesn’t play well with others’.

Interesting.

As we've seen last year, Ferrari finally had a major improvement of the chassis during the season since at least 2012. Sanchez arrived in October 2012. Not implying any direct correlation, but something clearly wasn't working as it should. He's also completely credited for Byrne's conceptual ideas for SF70H and F1-75.

I thought the same. If upgrades go well as planned perhaps he really was the clog in the drain lol.

Didn’t know he got all the credit for Byrne’s concepts.

[ing.]

https://www.the-race.com/formula-1/mcla ... explained/

Thinking back 12 months ago, when Ferrari was quite openly admitting it wasn’t that fussed that Sanchez was leaving, the summary from various conversations The Race had was: ‘he’s a great technical mind, but he doesn’t play well with others’.

Interesting.

As we've seen last year, Ferrari finally had a major improvement of the chassis during the season since at least 2012. Sanchez arrived in October 2012. Not implying any direct correlation, but something clearly wasn't working as it should. He's also completely credited for Byrne's conceptual ideas for SF70H and F1-75.

Sanchez effectively wasted almost a year of development last year by sticking with his flawed concept. Seems to me he didn’t fool anyone at McL either, which would explain his almost immediate (record-setting?) departure from Woking.

About SF70H and its trend-setting SP inlet design, wasn’t this car Cardile’s first F1 as aero chief (?) having come from the GT sector of the company?

[venkyhere]

Interesting discussion.
Fundamentally, all the 'modelling tools' (1) and 'data tools' (2) are there to :

(1) Modelling tools to imitate nature, the physics of car racing - mechanical, electrical, aero ; inorder to drive the engineering that goes towards making the car. The purpose is to come up with the 'least compromised' optimal car when there are myriad conflicting engineering restrictions ; the most basic set of which is defined by FIA as the 'formula for this rule cycle'.

(2) Data tools to find out what the changeable inputs -
(a) driver input
(b) engineering input (the 'settings' - engine mode, battery deployment mode, suspension stiffness, wing levels, brake bias etc)
(c) environmental inputs (tyre compounds, tyre pressure, track and ambient temperatures, air pressure, humidity etc)
affect the accuracy of (1) sought by the engineering team. For this, the data tools are just acting like 'distilleries' where they chew though thousands of variables, and come up with the most relevant "few hundreds" to perform surgery on (1), such that two things can happen :
-- how to make (1) immune to the different classes of inputs from (2) ; a.k.a how to refine the base-design of the car
-- how to make (1) adapt to the different classes of inputs from (2) ; a.k.a how to make the car suit different tracks and drivers in the most cost-effective and least engineering-intensive way.


Obviously, we can see all the points we are discussing :
- how accurate
- how quickly in terms of duration
- how efficient in terms of resources needed
- how many 'development/correction paths' in parallel possible
- how repeatable
is the hardware and software available to a team, when it comes to implementing (1) and (2). The scope for all the jargons we hear 'big-data' , 'distributed computing' , 'ML' , 'AI' is immense, since the game has changed from the crude 'I think this is how an engine should be put together like this, in a car that should be put together like this, with a body shape like this' - the only way to know how good it is, is from what the driver says and what the stopwatch says' (we know that for a large portion of F1's 100 year history, cars were made that way - I call them magicians, because the whole thing was based on intuition and nothing else) , into the world of simulation and correlation.

Now tell me, how to really separate the 'physics' from the 'data' ? Physics is just a 'model' of nature, we have at best, some mathematical equations. Data from sensors and actuators 'reveals' how nature is behaving. Now, assume we have the lastest AI capability and AI can itself figure out the physics of an engine or the physics of air or the physics of complex springs and dampers, purely from 5 years of 'data' from each track on the racing calender. Can we ignore human engineers with years of experience and design a car from scratch in the 6th year, purely via AI ? Now bring in words like 'engineering' , 'design' , 'algorithms' , 'what the driver feels on his bum and how effectively is he able to make the engineers understand it, whether he needs a language translator' etc into this, and this gets seriously intertwined into a complex web of various disciplines of expertise.

So I would rather refrain from making statements like 'but that is physics and this is just data' or vice-versa.
This is exactly why I was vehemently calling out (pardon me for out of topic mention) Toto Wolff's statement of 'we are just struggling with the physics of it' (as if everything else is in place) , in other threads on the forum - it's the stupidest thing that he could have said.

PS : things like material science, manufacturing process, logistics of manufacturing and transporting parts, tyre-strategy, optimizing cost-cap etc has not even been mentioned above. They are further more layers.

[Polite]

https://www.the-race.com/formula-1/mcla ... explained/




Interesting.

As we've seen last year, Ferrari finally had a major improvement of the chassis during the season since at least 2012. Sanchez arrived in October 2012. Not implying any direct correlation, but something clearly wasn't working as it should. He's also completely credited for Byrne's conceptual ideas for SF70H and F1-75.

Sanchez effectively wasted almost a year of development last year by sticking with his flawed concept. Seems to me he didn’t fool anyone at McL either, which would explain his almost immediate (record-setting?) departure from Woking.

About SF70H and its trend-setting SP inlet design, wasn’t this car Cardile’s first F1 as aero chief (?) having come from the GT sector of the company?

not Cardile but Resta with Sanchez under him

[Vanja #66]

Didn’t know he got all the credit for Byrne’s concepts.

It's not a public credit, but people online are completely unaware of the work Byrne still does. Not that he cares, of course He was in Maranello for a few months this winter, this can only mean he was deeply involved with some important aspects. His approach is to always go for usable downforce and make sure it works as expected in all conditions, even when early 2000s Ferraris were running in Q mode for half the race. You can literally see when he's being listened to more or less, depending on how the car behaves.

Sanchez effectively wasted almost a year of development last year by sticking with his flawed concept. Seems to me he didn’t fool anyone at McL either, which would explain his almost immediate (record-setting?) departure from Woking.

There was a very peculiar and unfortunate series of events that lead to SF23 being launched the way it was. On one hand, Ferrari's terribly strategies and poor PU reliability (of the updated PU that was supposed to be even more reliable and more powerful than launch spec) ensured there is no point updating the 22 car after France. Usually Binotto's decision to stop there and focus on 23 would have been ok - if there wasn't an incoming TD39 and it's questionable effects on Ferrari.

It would also have been very useful to use the 22 car as a test mule for higher floor edges with at least a minor rework done on one spec of the floor no longer in use. So going into 2023 design they had no idea what could wait on the other side. On top of that, you had Vigna pressuring to go for lower drag at all costs and this meant sacrificing some things. Neither Binotto nor Sanchez agreed with this decision and this was one of the things that pushed them both towards resignation. As far as I understood, it was this late-2022 period when Byrne was less involved.

Right now, Cardille and his team seems to be working very well with Byrne. In my view, Vasseur also understands how important it is to have such an experienced engineer, even as a consultant. Steering the project in the right direction at a crossroads where you can't really tell for sure which path is better is priceless. It's like a decision matrix I imagine, an experienced engineer has the factor-weighing method for every case

Now tell me, how to really separate the 'physics' from the 'data' ? Physics is just a 'model' of nature, we have at best, some mathematical equations. Data from sensors and actuators 'reveals' how nature is behaving. Now, assume we have the lastest AI capability and AI can itself figure out the physics of an engine or the physics of air or the physics of complex springs and dampers, purely from 5 years of 'data' from each track on the racing calender. Can we ignore human engineers with years of experience and design a car from scratch in the 6th year, purely via AI ? Now bring in words like 'engineering' , 'design' , 'algorithms' , 'what the driver feels on his bum and how effectively is he able to make the engineers understand it, whether he needs a language translator' etc into this, and this gets seriously intertwined into a complex web of various disciplines of expertise.

All models are inhretenly flawed, all data is inherently flawed, errors creep up and there's no changing that. However, this does not mean it's all useless - most of the time it's very useless since errors are now quite small and some are practically insignificant.

What "AI" is doing is taking huge amount of data, generating some patterns from it and is later capable of yielding some results for completely new input data. What's different from CFD in F1 for example, is that AI computing is not regulated. So even if it's only good enough to single out very bad results which would lead to wrong dev path, it's very usefull. And this is where those errors from before can make a real mess - if the original data fed to AI/ML tools is not good enough, you can perfect the process itself all you want. It will still yield unreliable results.

This winter there was a post somewhere that all teams are already using some AI/ML tools for initial CFD work and this isn't surprising. It takes a lot of data to sort out the CFD results to the levels required for F1 application. The corelated CFD results are a valid input and all teams have serious amounts of those. The wrong thing would be to only use 2022 and 2023 results for 24/25 car development, since you deprive the AI/ML tool of tons of diversified data. This diversification will likely ensure that proper geometry-flow structure patterns emerge, especially since previous generation of cars had an abundance of vortices interactig with turbulent wakes all over the car.

[leblanc]

Didn’t know he got all the credit for Byrne’s concepts.

It's not a public credit, but people online are completely unaware of the work Byrne still does. Not that he cares, of course He was in Maranello for a few months this winter, this can only mean he was deeply involved with some important aspects. His approach is to always go for usable downforce and make sure it works as expected in all conditions, even when early 2000s Ferraris were running in Q mode for half the race. You can literally see when he's being listened to more or less, depending on how the car behaves.

I misunderstood and thought you meant Sanchez got public credit. And, yep… 97-04 was fun; Seventy-one victories was a beautiful thing.

[Vanja #66]

Vasseur talked about the team work and managing the drivers and how they push each other

https://scuderiafans.com/fred-vasseur-o ... or-carlos/

“Last year there were sparks between our drivers at Monza. In Singapore, we asked Charles to sacrifice himself for Carlos, and he did it, favoring teamwork. He started on the red tires and blocked those behind him. Since the beginning of the year, they have understood that they need to ally to maximize Ferrari’s results, and they are doing it very well, even though Carlos is in a difficult position. Charles Leclerc also listened to us in Australia, when we asked him not to attack, or in Japan, where he let Carlos Sainz pass”

“This kind of competition can happen at every level, even for 18th place. It is inevitable because in our sport, your teammate is the only person you have the same references with. Some performance also comes from emulation. For example, in 2023 after the summer break, Carlos was very strong, and this pushed Charles Leclerc to go faster, so much so that he secured 6 or 7 front rows in succession. This competition is necessary, and my job is to keep it within limits”

[FDD]

Excellent video presentation of Ferrari's simulator model - math (ENG subs are available).
The most important thing in math modeling is to understand the terms, like "Artificial Intelligence" or "Machine Learning" etc and model must be not underfitted ie non precise, but also must be not overfitted ie not robust.
Regarding the terms, the first one is not "Intelligence" and the second one is not "Learning"!!!
The simulator model development, learning also is always made made by Humans, every single time.
Where is hardware and software in the game?? Huge number of iterations in short time, nothing more and nothing less then that.

[deadhead]

It's not a public credit, but people online are completely unaware of the work Byrne still does. Not that he cares, of course He was in Maranello for a few months this winter, this can only mean he was deeply involved with some important aspects. His approach is to always go for usable downforce and make sure it works as expected in all conditions, even when early 2000s Ferraris were running in Q mode for half the race. You can literally see when he's being listened to more or less, depending on how the car behaves.

That's how it seems to me as well from the outside but there are a lot of people online that say that he doesn't do much at Ferrari anymore, although I find that hard to believe.. why would he spend 2 months at the factory otherwise?

Every time he seems to be more involved it shows on track.

Do you think the F1-75 concept would've had more potential if the FIA hadn't intervened with the technical directive? Every "journalist" keeps saying how that was the wrong concept, and RB got it right, but I'm really not convinced that's the case at all, and I guess we would never know.

[KimiRai]

Skimming through, article claims;
First of three planned "big" upgrade, comes to Imola which will include inverted sidepod inlets, changes to sidepod, floor, front wing and cockpit? deflectors.

2,5 tenths expected. What I don't see is how they will invert the intake with the position of their SIS.

More than a month before the event (when that article came out) and already technical details and supposed tenths gained were leaked. I reiterate Ferrari has a real problem with leaking so much information, no other team comes close to this. It can create a bad dynamic with fans and the press if things don't go as hoped and probably just adds more pressure to the team.

[Vanja #66]

That's how it seems to me as well from the outside but there are a lot of people online that say that he doesn't do much at Ferrari anymore, although I find that hard to believe.. why would he spend 2 months at the factory otherwise?

Every time he seems to be more involved it shows on track.

Do you think the F1-75 concept would've had more potential if the FIA hadn't intervened with the technical directive? Every "journalist" keeps saying how that was the wrong concept, and RB got it right, but I'm really not convinced that's the case at all, and I guess we would never know.

The best and most experienced consultant on any subject is only as useful as the client allows them to be. As far as I understand, Ferrari "has" Byrne at their disposal however much they want (within reason and regular work hours ) and Byrne doesn't want to consult any other team, so it's always up to team management to decide how much they will include him. At the moment, seems Vasseur and Cardille are aligned on working with Byrne as much as he's prepared to work and coming back to Maranello for so long after a few years shows he's very eager - even at 80 years

F1-75 concept was using blunt undercut to generate strong outwash and this had a strong overall effect on keeping turbulent wake away from diffuser. It then used decent taper and inwash rear to generate some pressure recovery and reduce drag. Low floor edges made it possible to work, as we know from misfortunes of 2023. Alpine/RB downwash concept since launch 2022 works differently, pressurisation is introduced slowly at the front and this generates less drag. There doesn't seem to be any special trick around downwash rear and diffuser interaction, other than simply always bringing clean air to that area by extending the sidepod all the way down. This also generates a decent pressure recovery area for drag reduction.

In my view, Ferrari always had an option to evolve the front of the sidepod into what we see now, while inwash rear is maybe better for pressure recovery and it also generates some extra pressurisation on top of the floor - also known as downforce. Not a lot, but a decent amount. Who knows, maybe teams manage to find a way to improve the diffuser stability even with some inwash sidepod geometry. This kind of solution would be superior to anything we see know, provided it can be made to work