Honda Power Unit Hardware & Software

[diffuser]

ok thousands!

[Craigy]

ok thousands!

In hours, how many do you think each test engine/test article runs for?
How many test engines or components do you recon Honda will have run from last season to the first test?

[godlameroso]

Apparently they've been working on the 2017 power unit for the better part of last year, and ramping up development as the season went on. I don't think we'll ever really know how many engines they run on the dyno and for how long, nor will we ever learn how the power unit evolves. I remember Mercedes saying that they started with a very large and overweight power unit, and that the engineering process was much like a sculptor, in that all the inessentials were trimmed away to reveal the final product. In other words they started large, much like the design of computer chips, and then worried about downsizing components making them stronger lighter, and more compact.

[Craigy]

Apparently they've been working on the 2017 power unit for the better part of last year, and ramping up development as the season went on. I don't think we'll ever really know how many engines they run on the dyno and for how long, nor will we ever learn how the power unit evolves. I remember Mercedes saying that they started with a very large and overweight power unit, and that the engineering process was much like a sculptor, in that all the inessentials were trimmed away to reveal the final product. In other words they started large, much like the design of computer chips, and then worried about downsizing components making them stronger lighter, and more compact.

The process used to develop these engines is probably more secret than the resulting engines themselves, so I also doubt we'll ever find out about that. As an aside, a while ago I was at an IMechE lecture involving the Mercedes F1 team, and a question popped up from an aerodynamicist in the crowd about the sample frequency of Mercedes wind tunnel laser-scanned smoke during their aero development work. The Mercedes PR people jumped in to stop the complete answer from being given (I'm sure the aero guy would have wanted to give it and also been too professional to give it..). The process itself is part of their closely-guarded secret sauce, especially given the limited tunnel and CFD time available through the regulations. The PU process is just as closely kept.

For these complex PUs, I think that it's quite likely that lots of the component work is done in isolation, otherwise there are too many variables to capture data from, even when establishing the basics that each part is doing what it needs to (turbo generating correct pressures at expected RPM, that sort of thing). It would mean you could develop each part to the breaking point, make it run hotter/higher pressure, or indeed redevelop it to be lighter or smaller, and iterate as Godlamerso suggests. Rolls Royce's technique was (and remains) to do this: run the thing until it breaks, then redevelop it to be better. Iterative development.

Obviously, some things will be a fundamental shift in concept (HCCI/TJI, whatever) and will require probably all of the components to change at the same time: two separate branches of development. Even so, the overall process is likely to involve separation to stop the thing becoming a gordian knot of complexity. The components need to be tightly cohesive but loosely coupled (borrowing a term from software development there).

I think the fundamentals of the turbo, or intercooler(s), or (and especially) the combustion concept can be done more quickly and in isolated teams of developers with external simulation of the environment they need to work in - eg boost pressure, gases/fluids at the relevant temperatures/pressures, and so on.

Obviously they'll all be brought together to form the complete PU eventually, and integration tested on a bench somewhere, then integration tested in a chassis on a rig, then track tested, and ideally all of that will happen on time, instead of the first integration testing happening at a track during the preseason as we've unfortunately seen before due presumably to delays elsewhere.

I'm one of those sad people these days: a fan of McLaren. I want to see a McLaren test day (100km filming day) somewhere before the first test. It has worked in the past for Mercedes to get these minor things sorted out, so that in the initial testing the cars hit the track absolutely niggle-free, and run huge mileages so that they can do advanced things about understanding the real life of each of the components in track testing. It's the only time of the year this is possible.

[PlatinumZealot]

They can use empirical data from the previous years engines to shorten testing times.

[Brian Coat]

Craigy, I agree with your post but I offer the following remarks about a couple of points.

"The process used to develop these engines is probably more secret than the resulting engines themselves, so I also doubt we'll ever find out about that. "

There's no mystery. The engineering world is a small world so we already know F1 teams use state-of-the-art-but-normal engineering tools and apply them in much the same logical order as other engineers. But their process can be much faster and more effective due to the skill and teamwork concentration achievable in a specialist team of only a few hundred people.

"rig test ... each part to the breaking point ... Rolls Royce's technique was (and remains) to do this: run the thing until it breaks, then redevelop it to be better. "

Rig testing is an essential part of the engineering process but this statement is not completely correct. Testing to destruction is too costly/impractical/time consuming in many cases (especially on a jet engine!) and there are often better choices available.

[bauc]

Honda changing F1 engine concept for 2017 a risk, Hasegawa admits

"The concept is completely different," he told Autosport.
"It's very high risk, we don't know a lot of things about that new concept.
"We know it will give us a performance advantage but the biggest risk is whether we can realise that potential this year."

http://www.autosport.com/news/report.ph ... -high-risk

[-o<

[Thunder]

Well, another Year of unlocking Potential it is...

God i hope they can make it work, i'm (very) cautiosly optimistic.

[bauc]

Well, another Year of unlocking Potential it is...

God i hope they can make it work, i'm (very) cautiosly optimistic.

This is how I read into it.

Honda: We have a new concept, it will give us performance advantage but will it work properly when installed in the car and will it have the needed durability if we push it 100% is the main question.

[godlameroso]

I just hope that they have a good direction to go in, if they have to backtrack and head in another direction mid season it could prove costly.

[GoranF1]

McLaren has the Honda in the MCL32 engine started this morning in Woking

https://twitter.com/AlbertFabrega/statu ... 8468207616

[Craigy]

There's no mystery. The engineering world is a small world so we already know F1 teams use state-of-the-art-but-normal engineering tools and apply them in much the same logical order as other engineers. But their process can be much faster and more effective due to the skill and teamwork concentration achievable in a specialist team of only a few hundred people.

I guess the length and grasp of each iteration, the split of FEA/CFD simulation to prototype item to production part, the amount of QA and the tolerances involved at each step are all part of this, even within state of the art processes (so, how you use the tools). Equally the amount of resources (ultimately money) at each step will vary by development and organization; how much in-house content is there, versus supplier-provided stuff.
If F1 teaches me much, it's that the machinery is important but so is the technique of the operator in using the tool. That's applicable to drivers as well as development engineers.

"rig test ... each part to the breaking point ... Rolls Royce's technique was (and remains) to do this: run the thing until it breaks, then redevelop it to be better. "

Rig testing is an essential part of the engineering process but this statement is not completely correct. Testing to destruction is too costly/impractical/time consuming in many cases (especially on a jet engine!) and there are often better choices available.

I have just finished re-reading Hooker's book "not much of an engineer" for the umpteenth time so perhaps my head is full of how things were rather than are. I won't split hairs though.

[KingHamilton01]

No Risk, no reward! that idea of starting with an engine that could be fast but unreliable. They can fix the unreliable part but can't make a fundamentally slow engine Fast!

[bauc]

No Risk, no reward! that idea of starting with an engine that could be fast but unreliable. They can fix the unreliable part but can't make a fundamentally slow engine Fast!

Agreed =D>

[hemichromis]

Talking about a 'performance advantage'?

This could be a similar layout to now but much better or more likely a split turbo of some description.
I'm really hoping that they can match mercedes in power off the bat and sort out reliability through the season.

I think we also need to remember that Honda did not have TJI last year whereas all other teams did.
I am very hopeful