KERS (Kinetic Energy Recovery System)

[Frankman]

Does anybody know if the EUC uses the KERS (Kinetic Energy Recovery System) when braking or when going down hill?

[Keith]

2 minutes ago, Frankman said:

Does anybody knows if the EUC uses the KERS (Kinetic Energy Recovery System) when braking or when going down hill?

They all use KERS or more correctly regenerative braking - it is the only way they can be slowed down/stopped. Braking energy has to go somewhere and it isn't dissipated as friction heat. 

P.S. That is why downhill with a full battery is so challenging.

[Chriull]

And efficiency for regenerative breaking is presumably very bad - there is just not enough space to use appropriate supercapacitors, which could efficiently and fast enough "absorb" the generated energy. They would also increase the weight noticable. Also there are no converters to bring the generated voltage in a usable range... Could easily be, that driving downhill too slow costs energy from the battery - but i did not dig deep enough into this topic for real qualified statements.

The best "kers" with wheels could be to use the speed to go up inclines

[jer]

1 hour ago, Chriull said:

And efficiency for regenerative breaking is presumably very bad - there is just not enough space to use appropriate supercapacitors, which could efficiently and fast enough "absorb" the generated energy. They would also increase the weight noticable. Also there are no converters to bring the generated voltage in a usable range... Could easily be, that driving downhill too slow costs energy from the battery - but i did not dig deep enough into this topic for real qualified statements.

The best "kers" with wheels could be to use the speed to go up inclines

Not so sure that the efficiency for  regeneration  is particularly bad or worse than say an EV, everything just goes backwards.

With EV's my understanding is that about half the energy can be recovered by regen.

The mainboard has to do both things ie converting from dc to ac and vice versa.

Could be/may well be wrong but that's my understanding.

Jer

 

 

[Chriull]

Yes, Ac/dc and dc/ac conversion is done by the mainboard-what i meant was dc dc conversion (up and/or down), so the battery can (effectively) be charged.

My point of missing knowledge/understanding concerns the voltage levels. Imho the generated voltage of the wheel is more or less linear with the rpm. The battery voltage imho has to be higher than the generated voltage so the wheel can drive (acceleration,balancing). On the other side generated voltage would have to be higher than the battery voltage to charge the battery...

So for my understanding this are contradictory "states", so it should not work out - but in reality it seems to happen... ?

[jer]

33 minutes ago, Chriull said:

Yes, Ac/dc and dc/ac conversion is done by the mainboard-what i meant was dc dc conversion (up and/or down), so the battery can (effectively) be charged.

My point of missing knowledge/understanding concerns the voltage levels. Imho the generated voltage of the wheel is more or less linear with the rpm. The battery voltage imho has to be higher than the generated voltage so the wheel can drive (acceleration,balancing). On the other side generated voltage would have to be higher than the battery voltage to charge the battery...

So for my understanding this are contradictory "states", so it should not work out - but in reality it seems to happen... ?

I am out of my zone here.

In regen mode the voltage will build up in the motor winding until if flows.

So there is no dc/dc  conversion required

The transistors chop the current from single to 3 phase and they must be able to do the reverse.

Which is pretty darn clever

Jer

Someone who actually knows about  these things can come along and make it correct

[Chriull]

1 hour ago, jer said:

...

In regen mode the voltage will build up in the motor winding until if flows.

That could be a great input towards the solution - the motor generates a current. So depending on the load a voltage is defined.

But then i still dont understand the winding constant which i thought is a factor between rpm and generated voltage...

But one does not have to know everything

1 hour ago, jer said:

Someone who actually knows about  these things can come along and make it correct

+1

[Keith]

2 hours ago, Chriull said:

That could be a great input towards the solution - the motor generates a current. So depending on the load a voltage is defined.

But then i still dont understand the winding constant which i thought is a factor between rpm and generated voltage...

But one does not have to know everything

+1

You Are correct, taking Ohms Law into account the load does define the voltage. You are also correct that, into an open circuit, voltage will be proportional to RPM for a motor when acting as an alternator. That voltage, passed through the FET's to convert it back to d.c. Is fed to a battery (the load) which has very low internal resistance. Assuming good quality 25mOhm cells in 16s2p configuration then pack resistance will be around 0.2 Ohms. Since V=IR a current of 5 Amps would flow into the battery if the generated voltage was 1 volts above the current battery voltage, 10 Amps if it was 2 Volts above. In reality there will be losses in the electronics and in the motor coils as well.

In other words the low resistance of the battery will prevent a very high voltage being generated across the battery. What is not so clear to me is what stops the voltage from going above the maximum acceptable to the pack, my guess is that it relies upon the BMS to prevent more than 4.2 V/cell happening.

[Chriull]

52 minutes ago, Keith said:

...

In other words the low resistance of the battery will prevent a very high voltage being generated across the battery. What is not so clear to me is what stops the voltage from going above the maximum acceptable to the pack, my guess is that it relies upon the BMS to prevent more than 4.2 V/cell happening.

Imho thats the reason, why one should not go downhill with a full battery pack. The BMS will shut down as soon as the first cell reaches the overvoltage threshold. For this case a heatplate is missing to burn the excess energy ? Or a bms controlboard communication to start tiltbacks early enough to stop the rider...

[Chriull]

On 2.5.2016 at 2:46 PM, jer said:

...

Someone who actually knows about  these things can come along and make it correct

Happened!

@esajposted a great link in his great post:

Namely http://electronics.stackexchange.com/questions/56186/how-can-i-implement-regenerative-braking-of-a-dc-motor and there the "related link" http://electronics.stackexchange.com/questions/56170/how-can-suddenly-stopping-a-spinning-motor-cause-my-supply-voltage-to-shoot-up-w gives a nice overview/the answers how regenerative breaking works...

Should imho be a lecture to save for "our" firmware developers like @electric_vehicle_lover, once they have to consider braking - which will just happen the first time they testdrive their own firmware! Could save some fried motherboards...

[Mono]

On 5/2/2016 at 0:35 PM, Keith said:

They all use KERS or more correctly regenerative braking - it is the only way they can be slowed down/stopped. Braking energy has to go somewhere and it isn't dissipated as friction heat. 

P.S. That is why downhill with a full battery is so challenging.

I am pretty sure there are two modes of braking on my Gotway. Light braking which is regenerative, and strong braking, one could also call it power braking  which uses (a lot of) energy from the battery, I guess by trying to induce a backward momentum, usually successfully 

[esaj]

13 hours ago, Niko said:

I am pretty sure there are two modes of braking on my Gotway. Light braking which is regenerative, and strong braking, one could also call it power braking  which uses (a lot of) energy from the battery, I guess by trying to induce a backward momentum, usually successfully 

I was suspecting this earlier, but nowadays I think it's all just regenerative braking. The kinetic energy + using energy from battery trying to drive the motor backwards would seem to provide so much waste heat that something would have to burn. But of course, I don't know for sure.

[Mono]

1 minute ago, esaj said:

I was suspecting this earlier, but nowadays I think it's all just regenerative braking. The kinetic energy + using energy from battery trying to drive the motor backwards would seem to provide so much waste heat that something would have to burn. But of course, I don't know for sure.

I believe I monitored the current flow during braking some time ago to come to this conclusion, didn't you as well? IIRC the current was much higher than it would be allowed for charging the battery. 

If you are correct now, idling should be energetically cheap compared to accelerating. 

Here is a thought experiment: we roll slowly backwards and then lean strongly forward. You current theory suggests as long as the wheel spins backwards, the forward acceleration is done by regeneration, and only the moment the wheel switches to spinning forward the battery starts to drain.

[esaj]

3 minutes ago, Niko said:

I believe I monitored the current flow during braking some time ago to come to this conclusion, didn't you as well? IIRC the current was much higher than it would be allowed for charging the battery. 

Here's the same old graph I've posted many times, with me doing strong brakings on Euc Extreme's MCM2s. The current goes well into the negative territory during braking. Watch the speed line (blue) vs. the current (green). And yes, the current can be (at least for a while) much higher than what is recommended for charging the batteries. This one had 4 packs in parallel, so overall, that's around 10-12.5A per pack at spikes (or 5-6.25A, if the Gotway is reporting twice as high current as in reality, like some people say). Using too much current for a longer while to charge the packs would probably overheat them, but as the strong brakings only last a couple of seconds at best, I don't think it's enough to overheat the packs that fast.

 

3 minutes ago, Niko said:

If you are correct now, idling should be energetically cheap compared to accelerating. 

It's the coming to standstill and then starting to move again -point where the current shoots through the roof for a split second. Again, another graph I've posted multiple times, riding 20-25 degree uphill with the MCM2s. The first spike is getting going from standstill:

Takes about as much current as climbing the hill at the steepest points. That happens pretty much everytime you come to a stop (like changing direction during idling) and starting in the other direction again.

 

3 minutes ago, Niko said:

Here is a thought experiment: we roll slowly backwards and then lean strongly forward. You current theory suggests as long as the wheel spins backwards, the forward acceleration is done by regeneration, and only the moment the wheel switches to spinning forward the battery starts to drain.

The forward acceleration/backwards deceleration is the motor braking. Once you come to a stop, the above "spike" will happen, as the back-EMF is essentially 0V (thus the boost-converter effect of regenerative braking cannot push up the voltage), and the voltage difference between the packs and the motor will be at its greatest (lots of torque needed to get moving again, static vs. dynamic friction).

[Mono]

On 6/10/2016 at 2:35 PM, esaj said:

Nice graphs! A small technical remark: a vertical grid would help to analyze the graphs.

I think we can quite clearly see the change from regenerative braking to power braking in this graph (including the semi-bug in the controller software). At around 6.804 and 12.803 there is a kink in the power and current graph. More importantly the voltage changes from increasing to dropping while the current is still (increasingly) large negative. Voltage drop is indicative for large power drain. Why would the voltage drop if this were still a regenerative situation? Do we know any reason other than high load where we see the voltage visibly drop in short time?

[esaj]

12 minutes ago, Niko said:

Nice graphs! A small technical remark: a vertical grid would help to analyze the graphs.

I think we can quite clearly see the change from regenerative braking to power braking in this graph (including the semi-bug in the controller software). At around 6.804 and 12.803 there is a kink in the power and current graph. More importantly the voltage changes from increasing to dropping while the current is still large. Voltage drop is indicative for large power drain. Why would the voltage drop if this were still a regenerative situation? Do we know any reason other than high load where we see the voltage visibly drop in short time?

The sign of the current tells which way it is travelling (+ = from batteries to motor, - = from motor to batteries), so it's still in the regenerative-mode, even though the voltage starts to drop. As long as the current isn't positive, it's still charging the batteries, not drawing power from them. Why the voltage drops a bit even though the batteries are still charging, I really don't know.   And in general, it gets really complex trying to analyze situations like these, as there's so much going on between the batteries, mosfets, PWM and different motor coils & their magnetic fields   Actually the voltages and currents are fluctuating a lot (at the rates of several tens of kilohertz + smaller noise probably up to megahertz range?), but the graphs won't show it (probably the wheel reports some sort of averages, the voltage is measured before the high-side mosfets etc...)

[Mono]

23 minutes ago, esaj said:

The sign of the current tells which way it is travelling (+ = from batteries to motor, - = from motor to batteries), so it's still in the regenerative-mode, even though the voltage starts to drop. As long as the current isn't positive, it's still charging the batteries, not drawing power from them.

For the current shown in the Gotway app itself, this is not the case. The sign of the current changes if the direction of drive changes. The sign seems to indicate in which direction the torque goes, not whether the battery gets charged or drained. Didn't you use the same data stream?

[esaj]

34 minutes ago, Niko said:

For the current shown in the Gotway app itself, this is not the case. The sign of the current changes if the direction of drive changes. The sign seems to indicate in which direction the torque goes, not whether the battery gets charged or drained. Didn't you use the same data stream?

True that, but I'm still skeptic about the motor actually using battery power to brake... Where would all the energy go? When charging the batteries, a large portion of it is "soaked" up by the charging, but if both the motor and battery would be dishing out energy, it would have to burn off as heat somewhere. The mosfets would burn up, the wires would melt? But, I'm not an expert on these things, hopefully someday someone who really knows his/her way around 3-phase motors and their driving can explain it to us "mundanes"

[Mono]

On 6/10/2016 at 4:21 PM, esaj said:

True that, but I'm still skeptic about the motor actually using battery power to brake... Where would all the energy go?

I don't know, but most likely it will heat up motor and wires. The specific heat capacity of copper is 380 J/(kg K) = 380 Ws / (kg K) = 1000 Ws / (520g 5K). That is, power braking for 1s (about the period we see in the graph) with 1000W will heat up 520g copper by 5 degrees. This doesn't look like being inconsistent with what I know of the build of the wheel or its behavior. It's also clear that it is not possible to drain a battery charge of, say, 100Wh = 360000 Ws like this without overheat. In other words, minutes of continuous non-regenerative power braking are not an option (hard to test though). 

I can still-stand on the wheel and push against a wall leaning forward creating a 20+A current (i.e. 1200W or so) for a second or two. In this case no mechanical energy is generated or converted, as everything stands still. I can hardly see that then the current flows first out of the battery and then goes back in and I can do this forever without draining the battery. So the heat must go somewhere. If I would have much of a doubt, I could try this for longer, to see where the battery charge status goes, but currently I don't have really any doubt it will go straight down, so the energy does go somewhere.

EDIT: another observation hinting that power breaking is non-regenerative on the MCM2: breaking gets (much) softer when battery is low.