Honda Power Unit Hardware & Software

[Craigy]

The H could only ever receive 120kW (plus a small efficiency factor) from the K, as the K is limited to 120kW in either motor or generator mode.

Quite why you would want the H to be able to generate 240kW is beyond me, unless that is the power required for spooling the turbo. Which I doubt.

But the spooling power requirement is likely to be the sizing factor for the H, not the recovery power.

The reason why you want the H to be able to generate at >120kW is for the condition where it's offloading inertia and turbine power into the ES, and isn't being driven by the K.

You want the inertial energy out of the H to the ES to happen as fast as possible so that once it's depleted the flywheel down to the lower rpm limit, you can get back to receiving energy from the K at the regulated maximum 120kW rate.

Some examples:
If the H receives energy from the K at 120kW and dumps it to the ES at 60kW, it spends 2/3rds of its time unavailable to receive energy from the K, and can spend only 1/3rd of the time spooling up off energy from the K.
If the H receives energy from the K at 120kW and dumps it to the ES at 120kW, it spends 50% of its time unavailable to receive energy from the K, so spends half the time charging and half discharging energy from the K.
If the H receives energy from the K at 120kW and dumps it to the ES at 240kW, it spends 1/3rd of its time unavailable to receive energy from the K and can spend 2/3rds of the time spooling up off energy from the K.

You want the H spending as little time as possible dumping flywheel energy into the ES, because the faster it does that, the more time it's available to store from the K, and the more energy you can collect outwith the 2MJ K->ES rule.

To show a ridiculously extreme, non-real-world example, if you could transfer from the H to the ES at 1200kW, you could spend 91% of the MGU-H time collecting energy from the K at 120kW and only 9% of the time forwarding it to the ES at 1200kW.

Obviously, if you're talking about energy transfers of over 4MJ a lap from the ES to the K using the H as a go-between the same logic applies: the more time you are transferring energy from ES to H, the less time the H can be driving the K.

[stevesingo]

Question

1, Will the increase in MGU-H rpm also increase the amount of work done by the compressor?

2, Will energy inputted from MGU-K be equal to increased compressor work + inertia?

3, Will the added compressor load assist the reduction of rpm of the MGU-H during recovery of stored inertial energy.

4, Electrical losses aside what will the ratio of % energy in to % energy out?

[Craigy]

Question
1, Will the increase in MGU-H rpm also increase the amount of work done by the compressor?
2, Will energy inputted from MGU-K be equal to increased compressor work + inertia?
3, Will the added compressor load assist the reduction of rpm of the MGU-H during recovery of stored inertial energy.
4, Electrical losses aside what will the ratio of % energy in to % energy out?

My opinion, each question in turn:
1. We're talking about varying the compressor speed by <1% either side of what you'd otherwise have for a fraction of a second each time, so the additional effect of this is negligible on the compressor or turbine.
2. Yes, although as stated in (A1), the effect is negligible.
3. The compressor load, averaged out over even a short amount of time like 0.1s, is the same as without the flywheel system. We don't vary the compressor speed by more than 1% either side of optimal, and the compressor spends most of its time in the middle half of that 1% range, either side of the optimal amount. The effect in terms of "additional pumping losses" will therefore be quite small. I suppose it is possible that the compressor loses some efficiency through "chop", if poorly designed, but if that were the case, the compressor would also be inefficient even if we weren't using it as a flywheel.
4. Good question.
For context, I think K direct to ES and back to K is probably in the region of 90-95% roundtrip efficiency.
I further think that the flywheel method won't be as efficient, but perhaps in the region of 80-85% at a guess?
Edited to add: the first 4MJ of energy deployed via the K will be via the direct route of ES->K which is likely >95% efficient.

How you calculate the roundtrip efficiency of energy which nominally went the K->H->ES->K route is quite a complex topic, even if you ignore that it is pulsed in one direction, constant in another, and there's compounding in there too.

[dren]

Compressor work shouldn't increase due to either blow off valve operation and/or the H being de-clutched.

[stevesingo]

Question
1, Will the increase in MGU-H rpm also increase the amount of work done by the compressor?
2, Will energy inputted from MGU-K be equal to increased compressor work + inertia?
3, Will the added compressor load assist the reduction of rpm of the MGU-H during recovery of stored inertial energy.
4, Electrical losses aside what will the ratio of % energy in to % energy out?

My opinion, each question in turn:
1. We're talking about varying the compressor speed by <1% either side of what you'd otherwise have for a fraction of a second each time, so the additional effect of this is negligible on the compressor or turbine.
2. Yes, although as stated in (A1), the effect is negligible.
3. The compressor load, averaged out over even a short amount of time like 0.1s, is the same as without the flywheel system. We don't vary the compressor speed by more than 1% either side of optimal, and the compressor spends most of its time in the middle half of that 1% range, either side of the optimal amount. The effect in terms of "additional pumping losses" will therefore be quite small. I suppose it is possible that the compressor loses some efficiency through "chop", if poorly designed, but if that were the case, the compressor would also be inefficient even if we weren't using it as a flywheel.
4. Good question.
For context, I think K direct to ES and back to K is probably in the region of 90-95% roundtrip efficiency.
I further think that the flywheel method won't be as efficient, but perhaps in the region of 80-85% at a guess?
Edited to add: the first 4MJ of energy deployed via the K will be via the direct route of ES->K which is likely >95% efficient.

How you calculate the roundtrip efficiency of energy which nominally went the K->H->ES->K route is quite a complex topic, even if you ignore that it is pulsed in one direction, constant in another, and there's compounding in there too.

Seems reasonable. I can't see it being efficient in terms of fuel usage, but it is not really about that. It is about lap time!

Is there not scope throttle the compressor inlet to allow the rpm to rise whist still delivering the appropriate amount of air?

[Nonserviam85]

My point is this. Why have the 4 MJ and 2 MJ limits on K transfers then allow those limits to be breached by sending the energy - to the same place - by a different route? If this is happening and the FIA knows about it (and they would because it would be obvious in the data) they would ban it! It is a device/strategy designed purely to circumvent a rule. It isn't happening IMO.

Why you say that? F1 is full of examples of following the letter of the law and not the spirit of the law...Unless someone proves or hints that a specific manufacturers gains a huge advantage compared to the others I don't believe it should be banned, especially if all manufacturers do the same.

[Craigy]

The "unusual energy flow" is legal according to my reading of the rules. What rule would it breach, according to your reading of the rules?

My point is this. Why have the 4 MJ and 2 MJ limits on K transfers then allow those limits to be breached by sending the energy - to the same place - by a different route? If this is happening and the FIA knows about it (and they would because it would be obvious in the data) they would ban it! It is a device/strategy designed purely to circumvent a rule. It isn't happening IMO.

Why you say that? F1 is full of examples of following the letter of the law and not the spirit of the law...Unless someone proves or hints that a specific manufacturers gains a huge advantage compared to the others I don't believe it should be banned, especially if all manufacturers do the same.

I think your quotes are a little messed up there nonserviam. They suggest I am saying things GG said and vice-versa.

[restless]

But we KNOW that Honda engine is not enough fuel-efficient.
in order to fully utilize all these advanced routes they need to nail the ICE performance & efficiency.

Am I right that until then all these tricks will have small effect on overall performance - simply because they need to save fuel during race?

[dren]

But we KNOW that Honda engine is not enough fuel-efficient.
in order to fully utilize all these advanced routes they need to nail the ICE performance & efficiency.

Am I right that until then all these tricks will have small effect on overall performance - simply because they need to save fuel during race?

Yes, improving ICE efficiency is better than electrically accumulating fuel. This is just a means of side skirting fuel flow and MGUK regen to ES regulations.

[Webber2011]

The 40 pulses of K spinning up the H would be interleaved with 40 pulses of K direct to ES, so the net result at the K is a constant 120kW of retardation, not pulsed at all.

Hi Craigy

I don't understand why the K would be used to spin up the H?

There's an energy limit of 2MJ on the direct connection from K->ES per lap per FIA F1 technical regulations.
From ES->K, the equivalent limit is 4MJ, so it's not a symmetrical limitation in the regulations.
If you want to deploy 4MJ (or more) of energy per lap out of the K, then you need a way to harvest it.

There's no regulated limit on the energy per lap the K can send to the H, and no regulated limit per lap on the energy the H can send to the ES.

Consequently, you can elect to send >2MJ from the K to the ES, so long as the H is used as a go-between for any amount over 2MJ per lap.

Since the H can't both speed up (receiving energy from the K) and spin down (sending energy to the ES) at the same time, you would switch between sending energy K->H then stopping that, and starting to discharge energy H->ES, and back to K->H again, and so on, many times per second.

The switching would speed up and slow down the H only a small amount each time it is spun up by the energy from the K or spun down by sending energy to the ES, keeping it (and all the rest of the turbo/compressor) in -or very close to- the range the ICE needs.

Because the H is spending some of its time sending energy to the ES, the K can be used during that time to send energy directly to the ES, inside the 2MJ "direct" route regulated amount. Thus the K can actually be harvesting all the time, even when the H is busy dumping energy into the ES.

The K is switching from sending energy direct to the ES, then to the H, then to the ES, then to the H, about 40 times per second in the Honda docs.
All this energy ultimately gets into the ES, but the 2MJ limit only applies to direct K->ES transfers.

It's a way to get 4MJ or more into the ES per lap, without relying on the MGU-H's other job, which is compounding energy out of exhaust gases.

Fantastic explanation !

Cheers Craigy

[Nonserviam85]

My point is this. Why have the 4 MJ and 2 MJ limits on K transfers then allow those limits to be breached by sending the energy - to the same place - by a different route? If this is happening and the FIA knows about it (and they would because it would be obvious in the data) they would ban it! It is a device/strategy designed purely to circumvent a rule. It isn't happening IMO.

Why you say that? F1 is full of examples of following the letter of the law and not the spirit of the law...Unless someone proves or hints that a specific manufacturers gains a huge advantage compared to the others I don't believe it should be banned, especially if all manufacturers do the same.

I think your quotes are a little messed up there nonserviam. They suggest I am saying things GG said and vice-versa.

You are right, I deleted an extra line of quote there...I'll try and fix it

[PlatinumZealot]

...The basis for this is an energy flow diagram which depicts 1MJ being sent from the MGUK to the MGUH and then 1MJ from the MGUH to the MGUK, and a set of squiggly lines on a performance plot.

I'm afraid you're wrong there. The basis for this is an idea I had about 2 years ago. Happily, I posted on F1technical about it at the time so I can prove it.
The timestamp for that post is Tue Feb 09, 2016 2:36 pm. It's in the Mercedes PU thread if you want to read it.
viewtopic.php?f=4&t=23688&p=617506

My idea at the time wasn't fully-formed as I was missing the "switch it on and off fast to keep the compressor in the correct working range" part of the solution.

I didn't keep talking it through at the time, because of the reaction from gruntguru and others.

It was nice to see the Honda article though.

Yeah.
Yeah very creative thinking there. You deserve a gold star for that.

[Wazari]

Some great stuff here. Unfortunately this is one area where I am definitely prohibited from speaking on specifics.

However, a few weeks ago, I attended a sailing symposium (that's a another story) and a fellow engineer from Mercedes was there. I thought he came up with a great analogy when trying to explain some of the broad parameters of the relationship between the MGU-H, K and ICE to some other sailors. He said think of an hydroelectric plant and the effect of different water levels behind the dam turning the turbines and how to maximize output of the generator at all times. Also how much electrical output would you sacrifice for the ability to refill the dam to lengthen the overall duration of output.

One thing I learned about the MGU-H, K relationship is bigger is not better and many times less is more.

[henry]

.....

My point is this. Why have the 4 MJ and 2 MJ limits on K transfers then allow those limits to be breached by sending the energy - to the same place - by a different route? If this is happening and the FIA knows about it (and they would because it would be obvious in the data) they would ban it! It is a device/strategy designed purely to circumvent a rule. It isn't happening IMO.

If you look at the energy flow diagram the route from K to H is specifically allowed by the arrowhead on the line. So if you follow the lines and arrows the flow K -> H -> ES is envisioned in the rules and the mechanism proposed would be 100% legal.

[godlameroso]

Some great stuff here. Unfortunately this is one area where I am definitely prohibited from speaking on specifics.

However, a few weeks ago, I attended a sailing symposium (that's a another story) and a fellow engineer from Mercedes was there. I thought he came up with a great analogy when trying to explain some of the broad parameters of the relationship between the MGU-H, K and ICE to some other sailors. He said think of an hydroelectric plant and the effect of different water levels behind the dam turning the turbines and how to maximize output of the generator at all times. Also how much electrical output would you sacrifice for the ability to refill the dam to lengthen the overall duration of output.

One thing I learned about the MGU-H, K relationship is bigger is not better and many times less is more.

Does this mean that MGU-H harvesting is an afterthought and more important is to focus on combustion? As that would be the damned water in your analogy?