2014-2020 Formula One 1.6l V6 turbo engine formula

[63l8qrrfy6]

Yes! Exactly! I wonder how much experimentation happens along these lines. The trial and error to arrange all these compromises to get the most net benefit must be both exiting and frustrating.

In today's world, wouldn't this be solved with TensorFlow controlling the variables during the dyno runs?

I would expect machine learning to be well advanced in all aspects of F1 design and operation.

modeFrontier and Dassault iSight are typically used in the powertrain world

Surely trial and error is a thing of the past.

[godlameroso]

Yes! Exactly! I wonder how much experimentation happens along these lines. The trial and error to arrange all these compromises to get the most net benefit must be both exiting and frustrating.

In today's world, wouldn't this be solved with TensorFlow controlling the variables during the dyno runs?

I would expect machine learning to be well advanced in all aspects of F1 design and operation.

Your comment got me thinking, maybe to save time they use a variable valve timing system on the dyno to save time from making constant adjustments to the valve timing, as the regulations require fixed timing and lift. Commonly VVT is controlled by oil pressure, thus the cam gear typically has an oil filled chamber. Perhaps that VVT system on the dyno dampens some of the engine vibrations which would otherwise show up with fixed cam gears/shafts. Maybe that's something that caught Honda off guard.

[godlameroso]

Yes! Exactly! I wonder how much experimentation happens along these lines. The trial and error to arrange all these compromises to get the most net benefit must be both exiting and frustrating.

In today's world, wouldn't this be solved with TensorFlow controlling the variables during the dyno runs?

I would expect machine learning to be well advanced in all aspects of F1 design and operation.

modeFrontier and Dassault iSight are typically used in the powertrain world

Surely trial and error is a thing of the past.

There's always a place for trial and error, if it's good enough for nature, it's good enough for us.

[63l8qrrfy6]

Yes! Exactly! I wonder how much experimentation happens along these lines. The trial and error to arrange all these compromises to get the most net benefit must be both exiting and frustrating.

In today's world, wouldn't this be solved with TensorFlow controlling the variables during the dyno runs?

I would expect machine learning to be well advanced in all aspects of F1 design and operation.

Your comment got me thinking, maybe to save time they use a variable valve timing system on the dyno to save time from making constant adjustments to the valve timing, as the regulations require fixed timing and lift. Commonly VVT is controlled by oil pressure, thus the cam gear typically has an oil filled chamber. Perhaps that VVT system on the dyno dampens some of the engine vibrations which would otherwise show up with fixed cam gears/shafts. Maybe that's something that caught Honda off guard.

You are correct - hydraulic phasers do have a major damping effect.
However though so called "camshaft swing" tests are done early in the development stage while dyno durability validation is done relatively late in the programme and must run representative engine hardware otherwise it is useless. I doubt any team would commit such an oversight.

[godlameroso]

Yes! Exactly! I wonder how much experimentation happens along these lines. The trial and error to arrange all these compromises to get the most net benefit must be both exiting and frustrating.

In today's world, wouldn't this be solved with TensorFlow controlling the variables during the dyno runs?

I would expect machine learning to be well advanced in all aspects of F1 design and operation.

modeFrontier and Dassault iSight are typically used in the powertrain world

Surely trial and error is a thing of the past.

https://www.motorsport.com/f1/news/hond ... irculation

"The biggest thing is for sure the learning from the past three years.

"We did so much trial and error and we studied a lot from that. From that we can have some steps forward, and that can be connected to next year."

[erikejw]

What is the reason for the harsh engine restrictions? 3 per season?

My belief is that building two extra engines per team is cheaper than the extensive quality procedure and testing to have this long life engine.

The engines must be ready anyway. I think the engines run during the season is a small fraction of all produced.

Am I wrong and in that case why ?

[djos]

What is the reason for the harsh engine restrictions? 3 per season?

My belief is that building two extra engines per team is cheaper than the extensive quality procedure and testing to have this long life engine.

The engines must be ready anyway. I think the engines run during the season is a small fraction of all produced.

Am I wrong and in that case why ?

I agree, even if the PU's cost a million bucks each to make, an extra 2 million per year for 5 PU's has got to cost less than the R&D required to make them last an extra couple of races each.

[Tzk]

It's the costs. The customer teams pay per engine. So 3 engines is by far cheaper for the customers than 5 engines. For the manufacturer however it's cheaper to develop a not-so-long lasting engine because dyno time is by far shorter.

To me they went a step too far with 3 engines. Mercedes and Ferrari admitted that meeting this target is hard. They should've sticked to 5 or 6 engines per season and just reduce the upper price limit to save money on the customer side.

[Zynerji]

It's the costs. The customer teams pay per engine. So 3 engines is by far cheaper for the customers than 5 engines. For the manufacturer however it's cheaper to develop a not-so-long lasting engine because dyno time is by far shorter.

To me they went a step too far with 3 engines. Mercedes and Ferrari admitted that meeting this target is hard. They should've sticked to 5 or 6 engines per season and just reduce the upper price limit to save money on the customer side.

6 engines, 3 specs would be perfect.

[wuzak]

It's the costs. The customer teams pay per engine. So 3 engines is by far cheaper for the customers than 5 engines. For the manufacturer however it's cheaper to develop a not-so-long lasting engine because dyno time is by far shorter.

To me they went a step too far with 3 engines. Mercedes and Ferrari admitted that meeting this target is hard. They should've sticked to 5 or 6 engines per season and just reduce the upper price limit to save money on the customer side.

6 engines, 3 specs would be perfect.

6 ICE, 3 of everything else, perhaps?

Currently it is 3 of ICE and Turbo, and 2 of the electrical side, IIRC.

[Zynerji]

I would expect the control electronics to be relatively inexpensive, and battery cost is in assembly.

6 ICE, unlimited batteries and control electronics, unlimited turbos.

Turn them up to 11!

[Tzk]

6 engines, 3 specs would be perfect.

Well, currently only the number of engines is limited, but not the number of different specs. They teams could run a single spec for the whole year if they wanted to... So raising the number to 6 engines would mean it's possible to run 6 specs. And i'd keep it this way. Also you might get away without penalty, if one of the 6 engines fails/dies, as you could still make it through the season on 5 engines.

To me raising the engine count again is the way, together with introducing an even lower price limit. This way the customers still have the advantage of lower cost and the engine manufacturer doesn't have to run excessive dyno sessions to test reliability.

[Zynerji]

6 engines, 3 specs would be perfect.

Well, currently only the number of engines is limited, but not the number of different specs. They teams could run a single spec for the whole year if they wanted to... So raising the number to 6 engines would mean it's possible to run 6 specs. And i'd keep it this way. Also you might get away without penalty, if one of the 6 engines fails/dies, as you could still make it through the season on 5 engines.

To me raising the engine count again is the way, together with introducing an even lower price limit. This way the customers still have the advantage of lower cost and the engine manufacturer doesn't have to run excessive dyno sessions to test reliability.

3 specs would decrease cost. But honestly, the cost of just administering and checking compliance might be high.

6 ICE. Good number.

[Big Tea]

6 engines, 3 specs would be perfect.

Well, currently only the number of engines is limited, but not the number of different specs. They teams could run a single spec for the whole year if they wanted to... So raising the number to 6 engines would mean it's possible to run 6 specs. And i'd keep it this way. Also you might get away without penalty, if one of the 6 engines fails/dies, as you could still make it through the season on 5 engines.

To me raising the engine count again is the way, together with introducing an even lower price limit. This way the customers still have the advantage of lower cost and the engine manufacturer doesn't have to run excessive dyno sessions to test reliability.

3 specs would decrease cost. But honestly, the cost of just administering and checking compliance might be high.

6 ICE. Good number.

Another restriction could be 'use as many as you want, but each team supplied gets the same, for the original seasons cost'

If they are then prepared to supply an update to everyone, it has to be a worthwhile update or genuine failure, and will not increase the cost to 'lower' teams.

There would obviously have to be facility for a supplied team to purchase an extra unit incase of unforeseen damage

[Zynerji]

I agree.

I expect the real customer cost to be wrapped up in the human capital, not blocks of metal running through computerized machines.

Flat rate engine supply? Great idea!