Honda F1 project leader Yusuke Hasegawa has outlined a number of reasons why Honda has been struggling so badly in the beginning of the 2017 Formula One season. He confirmed that lots of problems were not discovered while running on the dynamo meter.
Fuel chemistry plays a huge role, in fact it plays the biggest role during combustion together with the chamber geometry. And it's also very important to get the combustion under control with good ignition timing algorithms and tables, which - in these current PUs - are probably not 3-dimensional, but I think 5 dimensional. However, you simply can't control a engine this precise. At a clock rate of 1 GHz, the CPU needs 1 nanosecond for every clock and even with a very fast engine control software, which for example only controls the injection and ignition timing for the ICE, you need multiple clock cycles to complete one timestep in the realtime ICE control. But the control unit isn't the only limiting factor, the latency between all the elements in the PU is much worse than the speed of the control unit and the variance of time it takes for the spark to jump over is bigger than in the nanosecond timescale.
What you say is true - more compute would be helpful.glenntws wrote: ↑11 Jun 2018, 21:55Fuel chemistry plays a huge role, in fact it plays the biggest role during combustion together with the chamber geometry. And it's also very important to get the combustion under control with good ignition timing algorithms and tables, which - in these current PUs - are probably not 3-dimensional, but I think 5 dimensional. However, you simply can't control a engine this precise. At a clock rate of 1 GHz, the CPU needs 1 nanosecond for every clock and even with a very fast engine control software, which for example only controls the injection and ignition timing for the ICE, you need multiple clock cycles to complete one timestep in the realtime ICE control. But the control unit isn't the only limiting factor, the latency between all the elements in the PU is much worse than the speed of the control unit and the variance of time it takes for the spark to jump over is bigger than in the nanosecond timescale.
During combustion model development of my own project I used a timestep of about .02 ms, but only during injection or ignition. .02 ms is very accurate and it provides you with great information about the way the engine works, but .02 ms is probably less than the combined timing error you have in real time control, even with tight-spec f1 components.
But that's not even a problem, even a very complex to control combustion system has no problem with these variances from cycle to cylce. It's more about the combination of the whole system - at which time each component has to do it's job and mainly the relation between every event in the chamber and the resulting timings - than the ultra accurate control - which is not possible, not even in f1.
Like I said, the fuel guys are looking at those time scales, because it honestly helps develop the fuel. That the engine in the car doesn't operate at those time scales is understood, but that doesn't mean someone isn't looking there, or that because the systems and the engine can't operate at that resolution doesn't mean analyzing those timescales is not useful for development.glenntws wrote: ↑11 Jun 2018, 21:55Fuel chemistry plays a huge role, in fact it plays the biggest role during combustion together with the chamber geometry. And it's also very important to get the combustion under control with good ignition timing algorithms and tables, which - in these current PUs - are probably not 3-dimensional, but I think 5 dimensional. However, you simply can't control a engine this precise. At a clock rate of 1 GHz, the CPU needs 1 nanosecond for every clock and even with a very fast engine control software, which for example only controls the injection and ignition timing for the ICE, you need multiple clock cycles to complete one timestep in the realtime ICE control. But the control unit isn't the only limiting factor, the latency between all the elements in the PU is much worse than the speed of the control unit and the variance of time it takes for the spark to jump over is bigger than in the nanosecond timescale.
During combustion model development of my own project I used a timestep of about .02 ms, but only during injection or ignition. .02 ms is very accurate and it provides you with great information about the way the engine works, but .02 ms is probably less than the combined timing error you have in real time control, even with tight-spec f1 components.
But that's not even a problem, even a very complex to control combustion system has no problem with these variances from cycle to cylce. It's more about the combination of the whole system - at which time each component has to do it's job and mainly the relation between every event in the chamber and the resulting timings - than the ultra accurate control - which is not possible, not even in f1.
I am happy to see the Honda PU having a huge bump in power, they deserve to be fruther up in the Power Unit ranking.
I also don't get it, if Gasly used the new PU in the race or not. In Qualifying he definetly used the old PU, which is the only explanation for his bad position. As I understand, he would have to start from the pits if he changed the car's specification after qualifying so he should have also used the old pu during the race. Call me dumb, but I think Honda did this all on purpose and Toro Rosso knew everything, I don't think they are able to switch back to the old PU between FP3 and Q1 and make a decent time in Qualifying, simply because there is no track-related PU setup for the old ICE and because of the time it would take to change the PU, so I think they already changed the PU before FP3 and I think they did it to show Red Bull the difference between old and new PU. But I'm not sure about that, just my theory.
Im not sure about that.... They will have software for combustion in a standard chamber to scientifically compare different properties... But simulation of combustion inside special racing engines with secret gometry? I doubt it. I am noy an expert but my intuition tells me they fuels are designed for desired traits and sent to the manufacturers for dyno testing and the best one is selected.godlameroso wrote: ↑12 Jun 2018, 17:17
These big petrochemical companies like Esso, and Petronas, and Shell have the computers and the equipment to carry out that kind of analysis.
Yes.roon wrote: ↑16 Jun 2018, 18:47https://uploads.tapatalk-cdn.com/201804 ... c60f49.jpganthonyfa18 wrote: ↑12 Jun 2018, 17:02HONDA UPDATE MOTOR FROM CANADA
http://www.f1i.fr/wp-content/uploads/20 ... _Honda.jpg
One air-to-air intercooler per side. Has it been confirmed if Honda maintained the same MGU-H/compressor arrangement as last year?
Photos? Heresay? Interviews? Press release?PlatinumZealot wrote: ↑17 Jun 2018, 02:09Yes.roon wrote: ↑16 Jun 2018, 18:47https://uploads.tapatalk-cdn.com/201804 ... c60f49.jpganthonyfa18 wrote: ↑12 Jun 2018, 17:02HONDA UPDATE MOTOR FROM CANADA
http://www.f1i.fr/wp-content/uploads/20 ... _Honda.jpg
One air-to-air intercooler per side. Has it been confirmed if Honda maintained the same MGU-H/compressor arrangement as last year?
A photo from circa two years ago.
