- Login or Register
No account yet? Sign up
Mercedes F1 chief technical officer James Allison is confident that his technical team has found the root cause of the German-Anglo squad’s recent gearbox failures and can make the necessary changes to avoid its reappearance.
We often judge cars in comparison to others (which creates the illusion of being objective). I believe it is just astonishing to many members here that merc shows up in melbourne with an even more refined car while others made much smaller steps even though they started with a simplistic approach. In fact we were all expecting RB to come up with a huge aero update this weekend, but they did not, they still look simple. Merc launched the most complex spec and came to Melbourne with an even more complex version. That is a fact and not just a subjective observation. And it is of importance as well: Merc drivers stated they had problems in finding the sweetspot of the car. Many experts like Mike Gascoyne were wondering how to handle this complex aero set up, they judged the car to offer a narrow window by it´s look only (very subjective , isn´t it). The car´s performance is a result of the sum of it´s parts and their interactions. So where should we discuss car hardware on a holistic level? The secret is told by the overall picture and people here are wondering if merc´s hardware and they way they developed it is the right path to take.turbof1 wrote: ↑23 Mar 2017, 15:40Right, but it still has nothing to do in here. You are supposed to talk in these car topics about the functionality of specific parts. Complexity is highly subjective. For instance I find the corner matrix of an orb hugely complex as it interlines infinite amounts of corners into a perfect curve, but somebody else might find that proposterous. Let's stay objective: discussion about parts and functionality of parts only.djones wrote: ↑23 Mar 2017, 15:36Its one argument and potentially valid.
But given the success of this team in recent years, I think their skills and approach is probably justified as the best.
I see details as refinement, so stretching out every single mm of workable space.
Just a different way of looking at it. If this was a low budget, back end team, then I'd probably agree its a brute force and throwing things at it to try and help approach.
In my opinion because they where so far ahead in 2014, they have been able to be proactive instead of reactive. In other words instead of planing for 3 or 4 races from now, they are planing for the second half of the season and next year. In technical fields, not being under a constant time crunch can really pay dividends.f1rules wrote: ↑25 Mar 2017, 01:39And thats what i just dont understand honestly, both rb and mclaren seem to have choosen a basic route from which they can develop, what i dont understand is how on earth merc can show up, after new aero rules, and deliver a car that looks so mature and developed, in the past not even 3-4 years of stable rules would result in this level of maturity. Are they spending gazzillions more then others? Is their windtunnel the most sophisticated tunnel ever to be build, so accurate and calibrated they without problems can verify complicated concepts like this? Do they have an unreal cfd technology that is so close to real life? Or how on earth they do this?
Agreed.dans79 wrote: ↑25 Mar 2017, 01:53In my opinion because they where so far ahead in 2014, they have been able to be proactive instead of reactive. In other words instead of planing for 3 or 4 races from now, they are planing for the second half of the season and next year. In technical fields, not being under a constant time crunch can really pay dividends.f1rules wrote: ↑25 Mar 2017, 01:39And thats what i just dont understand honestly, both rb and mclaren seem to have choosen a basic route from which they can develop, what i dont understand is how on earth merc can show up, after new aero rules, and deliver a car that looks so mature and developed, in the past not even 3-4 years of stable rules would result in this level of maturity. Are they spending gazzillions more then others? Is their windtunnel the most sophisticated tunnel ever to be build, so accurate and calibrated they without problems can verify complicated concepts like this? Do they have an unreal cfd technology that is so close to real life? Or how on earth they do this?
What I suspected when I saw 2016's W floor, I said to myself there is no way a human being could design such a detailed thing. I'm a computer science engineer and I can tell you by the looks of it, from what I've seen in some of the projects, the bargeboard area could not have been completely desgined from a human hand. I know we all know about CFDs, every team has them, but I think Merc's using something more sophisticated. Some extremely sophisticated Machine Learning stuff is going on. I can assure you, a combination of Deep Learning and Machine Learning combination leads you to such unimaginable shapes. With their resources, such algorithms could outperform brute force approaches of trial and error with CFD and Windtunnel. They could cut the time necessary to come up with different shapes, they could produce faster pools of parts and just go crazy test them. I know every team is doing something similar, but Mercedes is way ahead I think.f1rules wrote: ↑25 Mar 2017, 01:39And thats what i just dont understand honestly, both rb and mclaren seem to have choosen a basic route from which they can develop, what i dont understand is how on earth merc can show up, after new aero rules, and deliver a car that looks so mature and developed, in the past not even 3-4 years of stable rules would result in this level of maturity. Are they spending gazzillions more then others? Is their windtunnel the most sophisticated tunnel ever to be build, so accurate and calibrated they without problems can verify complicated concepts like this? Do they have an unreal cfd technology that is so close to real life? Or how on earth they do this?
https://www.youtube.com/watch?v=aR5N2Jl8k14. From 10:50 Conti starts talking about applying a neuro system, from 12:30 he gets to the part where vehicles are involved.F1Krof wrote: ↑25 Mar 2017, 12:07What I suspected when I saw 2016's W floor, I said to myself there is no way a human being could design such a detailed thing. I'm a computer science engineer and I can tell you by the looks of it, from what I've seen in some of the projects, the bargeboard area could not have been completely desgined from a human hand. I know we all know about CFDs, every team has them, but I think Merc's using something more sophisticated. Some extremely sophisticated Machine Learning stuff is going on. I can assure you, a combination of Deep Learning and Machine Learning combination leads you to such unimaginable shapes. With their resources, such algorithms could outperform brute force approaches of trial and error with CFD and Windtunnel. They could cut the time necessary to come up with different shapes, they could produce faster pools of parts and just go crazy test them. I know every team is doing something similar, but Mercedes is way ahead I think.
Remember they were the first team to upload real-time data through wifi, analyze them in real time and react accordingly since 2014 with Qualcomm. There must be something about this.
E.g. (https://www.youtube.com/watch?v=YZUNRmwoijw).
As far as I know TED talks are quite well regarded but TEDx are less so.diego.liv wrote: ↑25 Mar 2017, 13:46https://www.youtube.com/watch?v=aR5N2Jl8k14. From 10:50 Conti starts talking about applying a neuro system, from 12:30 he gets to the part where vehicles are involved.F1Krof wrote: ↑25 Mar 2017, 12:07What I suspected when I saw 2016's W floor, I said to myself there is no way a human being could design such a detailed thing. I'm a computer science engineer and I can tell you by the looks of it, from what I've seen in some of the projects, the bargeboard area could not have been completely desgined from a human hand. I know we all know about CFDs, every team has them, but I think Merc's using something more sophisticated. Some extremely sophisticated Machine Learning stuff is going on. I can assure you, a combination of Deep Learning and Machine Learning combination leads you to such unimaginable shapes. With their resources, such algorithms could outperform brute force approaches of trial and error with CFD and Windtunnel. They could cut the time necessary to come up with different shapes, they could produce faster pools of parts and just go crazy test them. I know every team is doing something similar, but Mercedes is way ahead I think.
Remember they were the first team to upload real-time data through wifi, analyze them in real time and react accordingly since 2014 with Qualcomm. There must be something about this.
E.g. (https://www.youtube.com/watch?v=YZUNRmwoijw).
On a side note, is TED* something well structured and established?
