Regenerative systems (KERS)

[safeaschuck]

You know I hadn't thought of the change of speed xpensive (hence the 1:1 comment in my previous post) That would kind of force the use of a clutch unless the CVT varies it ratio perfectly in line with engine speed change when holding the full 60kw.

I suppose it wouldn't be too hard to have a fixed ratio to step 64k rpm down to roughly 18k but as you say you have the opposing destinations of the two parts of the powertrain.

For some reason I thought the KERS would be applied at a steady RPM (i.e. flat out in top) but given the way they gear the engines to max out for optimum time per lap I can't imagine that happening, unless some changes go on with that strategy.
And obviously there was the talk earlier on this thread about whether the boost would push the engine over the rev limit.
At the mention of strategy, does anyone think the teams might abandon the push to pass system and use this technology for fuel saving with the incoming re-fueling ban? Less fuel weight means a faster car at all points on the curcuit, plus reduced tyre wear, plus a host of other gains I haven't forseen I'm sure.

Anyway, somewhat more on point...

So they will apply KERS as they shift through the gears which isn't too bad as that would be, what a 4,000rpm gap? The flywheel will taper off from distributing useful amounts of power at what, the 20-40,000rpm range? (I would imagine Batteries have the same problem as they approach a potentially damaging state of charge, like too low) so the gearing difference is slightly less horrendous than it could have been... Still plentty enough to put a team off I should imagine.

The more I think about it, for KERS to realise it's full potential, you really need to disconnect the engine from wheels, in the traditional sense. It will require some sort of mechanical coupling, but one shared with the KER's rather than used by it (especially as the Power storage of KERS increases). This would be different to the current approach of simply adding power at the crank, or gearbox.
The best way I can think to describe it would be:
KERS and the engine adding power to seperate sides of A limited slip differential. the output of the diff would then be passed to the existing diff system where it is split between left and right wheel.
Each side could be sepearatly geared to suit it's own 'power-band'.
This would raise the possibility that engine speed would not be directly related to wheel speed (and really push the case for the use of C.V.T.s...) I think.

[4dsux]

As far as I can tell, Williams is using an electromechanical system, where the batteries are replaced by a flywheel. My guess is there are two motor/generators. One connected to the drive, one to the flywheel. Recovered energy immediately powers the motor/generator connected to the flywheel. Stored energy can be released by switching the functions of the motor and generator.

An interesting thing about flywheels is they could have a significant amount of gyroscopic precession. Teams could orient it in such a way that as the car rotates through a turn, the precession pushes the inside tires down and the outside tires up.

[Roland Ehnström]

An interesting thing about flywheels is they could have a significant amount of gyroscopic precession. Teams could orient it in such a way that as the car rotates through a turn, the precession pushes the inside tires down and the outside tires up.

Very interesting! But I wonder how big this effct would be, perhaps it is neglectable?

[xpensive]

This I believe has been mentioned before, but nonetheless most interesting.
Not by a long-shot my area of expertise, but I guess that you would have to orient the flywheel shaft vertically to get the desired gyroscopic effect to help against roll around corners as well as with anti-dive and squat, correct? Mounted longitudinally, it would just make the car difficult to steer, no?

[Ciro Pabón]

Neglectable my behind. I think we realized that (at this forum) well before it became "common knowledge".

I copy/paste what I wrote 9 pages ago, in this thread, when the subject was discussed:

All right, I'm back. Let's do one of our frequent back-of-the-envelope checkings. This is going to be boring because I'll write while I calculate.

First, how much is the energy we need to rotate a car, let's say, 10 degrees in 0.5 seconds? I suppose this is enough for "attitude correction".

For this estimation, let's also assume the car is an uniform prism that weighs 600 kilos and has, I don't know, 2 m of width and 3.5 meters in length. So:

angular velocity: (10 deg * Pi rad) / (180 deg * 0.5 sec) = 0.35 rad/seg

moment of inertia: 1/12 * 600 kg * [(2m)^2 + (3.5m)^2)] = 813 kg-m2

rotational energy: 1/2 * 813 * 0.35^2 kg-m2/s2 = 50 Joules

Well, I'm a little rusty, and the energy is pretty low (a mistake on my part?): it's the first time I've calculated this in years... I'm sure somebody will check. (NOTE: nobody did)

Now, we have a flywheel that weighs 5 kilos (I assume) and a radius of, I don't know, 10 cm. How fast it has to rotate to provide the same 50 joules of rotational energy? I assume no losses here, we could factor in them later.

If the flywheel is an ideal thin cylindrical shell, then:

moment of inertia of flywheel: 5 kg * 0.1^2 m2 = 0.05 kg-m2

rotational energy of flywheel: 1/2 * 0.05 kg-m2 * w^2 = 50 kg-m2/s2

So, the angular velocity needed to get the same 50 joules is:

w = (2 * 50 / 0.05)^ 0.5 rad/seg = 44 rad/seg
w = 44 rad/seg * (1 revolution / 2 * Pi rad) * (60 sec/1 minute) = 420 rpm

A change in rotational speed of 420 rpm in the flywheel is enough to move the car 10 degrees.

My largest "mistake", which I made on purpose, comes from the fact that I assumed a car with no friction with the ground. If the calculations are OK, they are true only for a car "on ice".

First conclusion: for "cars on ice", the change in flywheel rotation is pretty low, as is the energy needed. If we're talking of "damping" the movements of the car on the suspension, that is, without moving laterally the wheels, only to "help" the shock adsorbers the same way Renault mass damper worked, the idea is not only feasible, but also practical.

So, it's practical to use a flywheel to correct attitude for aerodynamic lateral sensibility (if you assume displacements of the same magnitude of the suspension travel).

Yes, yes, I know I did not took in account the energy spent on compressing the springs, but probably (not sure here) the spring is helping the flywheel, not opposing it.

So, at first sight we need a new regulation (like the Renault damper ruling) or some teams could use this trick (if I'm not totally wrong for reasons I still don't comprehend).

Well, you guys, I'd would be extremely grateful if someone (Miguel, CMSMJ1, where are you?) could check this first attempt. I'm sure this post is not going to be one of the most read here (it's boring as hell), but it was fun to write.

Finally, I have the hunch that 50 joules is too low. Last time I calculated by hand second order inertia was in 1994, I have been using tables or computer programs since then... Any ideas?

If the previous post it's true (and I havent rechecked my figures, but, you lazy bunch haven't either! ), then you can "move around" the car a lot with the KERS flywheel, unless there is some obscure regulation that prevents it. I'd say that you can even overcome friction.

Does this mean "Goodbye, under/oversteer"? I'd love to get a post with numbers or specific regulations, instead of opinions...

[4dsux]

I don't think that's the same topic... I was suggesting that the speed of the flywheel doesn't have to change at all to achieve an anti-roll effect on the chassis around a corner. The act of steering the car and charged flywheel around a corner creates a force perpendicular, which would help keep the car level in the corners. It's a pretty cool effect you can test with a rollerblade wheel in your hand. Try to tilt it and it will try to tilt in the perpendicular direction. Once I figure out how to calculate all this out I'll post it.

[Ciro Pabón]

I think, on the contrary, that it is the same idea. If you change the rotational speed of the KERS, the object rotates. If you don't, it resists rotation, because the angular momentum is constant (in a closed system).

Of course, an F1 car is not a closed system: you can use the friction of wheels with asphalt to impart angular momentum (as in the previous calculations).

The principle you mention is used in rotational dampers. Here you have the article in this site about the J-damper, which works by the same principle you mention, 4dsux:

http://www.f1technical.net/features/10586

And this is another damper, at Taipei 101, working during the earthquake of May 12, 2008 (it has nothing to do with the thread, but I think is cool: I've never seen people worried about a damper working instead of worrying about the earthquake... ).

[youtube]http://www.youtube.com/watch?v=NYSgd1XSZXc[/youtube]

The equations are:

Angular momentum = Moment of inertia x rotational speed

Kinetic energy = Moment of inertia x (rotational speed)^2 / 2

The derivative of the angular moment is the torque.

You simply calculate the angular momentum (I gave the equation for moment of inertia of a rotating hollow cylinder in the previous thread). Any change in the vector orientation of the angular momentum has to come from the friction of wheels.

Finally, welcome, 4dsux, I hope you stay around, good point.

[4dsux]

Ok I think I have some rough numbers. I'm horrible for stupid mistakes, so I hope its right End result was a fully charged flywheel will provide 34N of extra tire load to each inside tire, and lift each outside tire by the same amount. So... not much, but noticeable.

400kJ capacity
flywheel made of a steal disk, density 7.85g/cm^3 (a guess)
flywheel radius, r= 0.1m
flywheel thickness = 0.03cm
chassis rotation in a corner = w_chassis= 0.6 rad/s
track width = 1.5m

flywheel volume = 9.42x10-4 m^3
flywheel mass = 7400g

Inertia of disk = 1/2mr^2
I= 1/2*7.4*0.1^2 = 0.037kgm^2

Energy of rotation = 1/2Iw^2 , I is inertia, w is angular velocity
E= 400000J = 1/2*0.037^2*w^2
-> w = 4650rad/s = 44400rpm (the flywheel's speed fully charged)

precession torque = I*w*w_chassis
0.037*4650*0.6 = 103Nm

load difference between two sides of car= 103*(1.5/2) = 137N

divide by 2 tires per side, and for 2 sides, = 34N load added to each inside tire and removed from each outside tire

[xpensive]

With all due respect Ciro, I believe that you for once might have got your wires crossed.
What you have done is just to compare the theoretical kinetic energy of a rolling chassis with a flywheel. Interesting, but perhaps not really what we are looking for here, besides, I fail to understand how a change in flywheel speed would rotate the car?
In my (modest as always) opinion, 4dsux on the other hand makes perfect sense, it's the resulting precession torque from the flywheel to counter roll that is interesting. Brilliant in its simplicity. =D>

[Moanlower]

It's been estimated that you'll gain approx. 0.2 seconds/lap with a proper functioning KERS system. Now I have a simple question. How much will they lose per lap with the extra 30kg if it's not functioning for whatever reason ?

[Ciro Pabón]

With all due respect Ciro, I believe that you for once might have got your wires crossed.
What you have done is just to compare the theoretical kinetic energy of a rolling chassis with a flywheel. Interesting, but perhaps not really what we are looking for here, besides, I fail to understand how a change in flywheel speed would rotate the car?
In my (modest as always) opinion, 4dsux on the other hand makes perfect sense, it's the resulting precession torque from the flywheel to counter roll that is interesting. Brilliant in its simplicity. =D>

Well, I think I have my wires crossed all the time.

Would you be so kind to explain why angular moment is not conserved? AFAIK, there are two ways to turn the car around: using a momentum flywheel or a gyroscopic flywheel.

The first effect (momentum flywheel) turns the car in the opposite sense of the accelerating flywheel (I'm assuming the flywheel axis is vertical). The second efect turns the car when the axis of rotation displaces, like when the car goes down a hill, again, thinking of a vertical axis. In both cases there is a change in angular momentum vector (size of the vector in the momentum flywheel, direction of the vector in the gyroscopic one). Am I wrong?

BTW, I haven't thought that in vertical curves, like Eau Rouge, a single flywheel would "throw" the car out of the racing line (or change weight distribution, depending on the orientation of the flywheel) by gyroscopic "stabilization". I guess that's the reason why all teams use counterrotating flywheels.

It's been estimated that you'll gain approx. 0.2 seconds/lap with a proper functioning KERS system. Now I have a simple question. How much will they lose per lap with the extra 30kg if it's not functioning for whatever reason ?

Well, they don't lose time by weight, the cars all weight the same. All cars weight 50 kg less than the max (without ballast). They lose time because they lose ballast, but I have no idea how much: you need to know a lot about car dynamics to evaluate that (or to run a car with ballast instead of KERS, but I guess nobody is publishing the figures).

Finally, all the figures presented are speculations to show what's the effect that a single rotor KERS could have, let's make that clear. The simulations carried out by FIA shows that there is no effect in lateral dynamics.

If someone is interested, here there is the Power Point presentation on the subject:

http://www.icatconf.org/sunum/Sunumlar% ... Kovacs.ppt

And this is the relevant slide:

[xpensive]

Well, there you go I guess, we are talking about two different things, change in momentum (by angular accelleration) and pure gyroscopics, where I believe that 4dsux was discussing the latter. An accellerating shaft of any kind will always cause a counter-acting torque, like watching a top-fuel dragster twisting, isn't that the same thing?

[ISLAMATRON]

It's been estimated that you'll gain approx. 0.2 seconds/lap with a proper functioning KERS system. Now I have a simple question. How much will they lose per lap with the extra 30kg if it's not functioning for whatever reason ?

There is no extra 30 kg of weight... whether the KERS is installed or not, or even if it is installed but not functioning all the cars will race at the minimum weight of 605kg.
F1 car = 605 kg
F1 + KERS = 605 kg

All that is being done is that 30kg or so of ballast that was formerly in the car is being replaced with a KERS sytem which provides 80hp for 6.66 secs while using no extra fuel.

But yes the KERS sytem does slightly raise the CoG, and how much that would increase lap time is yet unknown to us. But seeing that the teams anticipate a decrease in lap times of a couple tenths with a functioning KERS, it could be somewhere around that number.

[tk421]

There is no extra 30 kg of weight...
F1 car = 605 kg
F1 + KERS = 605 kg

I concur. It's 30kg of ballast just where you want it for improved balance/handling/dynamics, or

a KERS sytem which provides 80hp for 6.66 secs while using no extra fuel.

Well said, ISLA =D>

[alexbarwell]

My take is that the gyroscopic effect entails the force to redirect the rotational inertia of the rotating body. Rotating about the body's rotational axis would be extremely negligible given that the casing is the only part truly being acted on, the rest is already spinning at a far greater rate and is separated/isolated by the bearings.
If the gyro mass spins on the x-axis, any re-orientation of that in the x-axis will be virtually nil, but the rotary momentum of the body will be reactive if rotated in the y or z-axis. Only with the flywheel rotating in the horizontal plane will there not be the gyroscopic effect to overcome that would otherwise be a resistance to turn in. Maybe if the flywheel was encased in a gimballed housing, or as mentioned by some - 2 counter-rotating flywheels to zero the effect.