Energy consumption and regen on mountain rides?

[Aneta]

A scientific question for those who've done big mountain rides such as Mt. Wilson (1350m of elevation gain): what's your estimate for watt-hours a) consumed and b) regenerated per 1 meter of elevation gain per 1 kilo of total weight?

Since wheel batteries at "0%" are actually not at 0 SOC (they still have voltage of about 3.3V), for more accurate estimate subtract 10% from rated capacity of you battery, for example, for 16X that would effectively be 1554 - 1554*0.1 = 1400Wh.

[The Fat Unicyclist]

I was trying a local (exploratory) ride that had an elevation gain of only ~300m... But on the descent, the headwind was significant enough that we had 0 regeneration going downhill and actually used more power...

[MrRobot]

Is just going downhill enough to regen or do you need to be braking too?

[mrelwood]

12 minutes ago, MrRobot said:

Is just going downhill enough to regen or do you need to be braking too?

The full term is ”regenerative braking”. You need to be braking, otherwise where would the wheel get the energy to recharge the battery? Image letting go of a volleyball, if the ball would go faster than you, you are braking. If the ball would go slower, you are accelerating, or otherwise using energy.

The battery level might recover a bit even if no regen is happening. Riding up a hill is a burden for the battery no matter what the speed, wind etc. Riding even a shallow downhill is much easier for the wheel, and if you are riding at the same speed as the volleyball, you are only using very little energy from the battery, for balancing only. After the uphill the battery would certainly recover a bit even without any regen taking place.

[Planemo]

Is there a point at which regen doesnt actually happen? I mean, under really heavy braking, does the motor actually use power to brake? Because theres only so much it can try to put back into the battery?

As you can guess I am a bit clueless how eucs deal with braking 

[Aneta]

3 hours ago, MrRobot said:

Is just going downhill enough to regen or do you need to be braking too?

It depends on slope and speed. If slope is slight and you're moving at enough speed to generate significant aerodynamic drag, the gravitational "thrust" won't be enough to overcome the resistance, so the motor must work in "drive" mode as usual. If slope is steep and g-thrust is higher than resistance, you are essentually continuously braking, and the motor will be in regen mode.

It all can easily be seen by observing the current voltage relative to "resting" voltage. Voltage goes up = you are regenerating/braking.

[Seba]

27 minutes ago, Planemo said:

Is there a point at which regen doesnt actually happen? I mean, under really heavy braking, does the motor actually use power to brake? Because theres only so much it can try to put back into the battery?

As you can guess I am a bit clueless how eucs deal with braking 

1. At top of the hill start your EUC app, whatever it is. You just need voltage readout. Wait a while to let the voltage stabilize.
2. Start riding downhill, but keep speed at about 15-20 km/h.
3. You will soon realize that voltage goes up noticeably higher than voltage you've seen before ride.

This is because when going downhill, your engine acts as a generator. It will convert mechanical energy from braking to electric power that charge the battery. Electric engine is reversible machine. It will start rotating when you power it up with electricity. Or vice versa - when you start rotating it, it will produce electricity. Our EUCs work in similar way to this - https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricity

[Aneta]

27 minutes ago, Planemo said:

Is there a point at which regen doesnt actually happen? I mean, under really heavy braking, does the motor actually use power to brake? Because theres only so much it can try to put back into the battery?

As you can guess I am a bit clueless how eucs deal with braking 

I don't think so, although some e-scooters, like Xiomi m365, do use battery power to brake hard. I believe this is done to save the motor from overheating. Hard regen braking is equivalent to essentially shorting the phase wires, which means that no current can go out of the motor and into the battery, but the current generated by back EMF at high speeds in the shorted windings will be huge, which can quickly overheat them. Shorting phase wires is the hardest braking possible, simply because it generates currents impossible to achieve by actively using the battery.

[Aneta]

Regen braking (with extreme case of shorted phase wires, in which 0% is actually regen'ed, 100% of current is circulating within the motor) is essentially generating Eddy current in motor's windings by moving magnets. Anyone can do this experiment, e.g.

(note that when he used copper coil, there was no slowdown - I think it's because he forgot to short the coil on itself!) Or disconnect phase wires from the controller on your wheel and short them - you'll be surprised by the stiff resistance of the motor to rotation by hand.

 

 

[Chriull]

13 minutes ago, Aneta said:

I don't think so, although some e-scooters, like Xiomi m365, do use battery power to brake hard. I believe this is done to save the motor from overheating. Hard regen braking is equivalent to essentially shorting the phase wires, which means that no current can go out of the motor and into the battery, but the current generated by back EMF at high speeds in the shorted windings will be huge, which can quickly overheat them. Shorting phase wires is the hardest braking possible, simply because it generates currents impossible to achieve by actively using the battery.

I had once a wheellog log showing braking (dropping speed) while the battery voltage was also dropping. So it seems that EUCs can use the battery for braking, too. Or just they "older" ones did (seldomly).

Regenerative braking is a "soft" braking technique (current wise). The mosfets are controlled so that the back emf generated from the motor gets a bit higher as the battery voltage and by this a controlled current is flowing into the battery.

The "middle" variant would be shortening the motor (coils) - this is imho not implemented with EUCs. In such a case a huge current could flow. Of course by PWMing this can be fine tuned.

The "strongest" variant can be braking actively using the battery. By this a negative torque is created in the motor, so there is a current higher than the shortcircuit current possible. Since this is achieved by changing the BLDC motor commutation, this can be very finely tuned (like the forward acceleration) until full backward "thrust" meaning more or less changing the polarity of the battery in regard to the generated back emf. And by this here currents higher as the short circuit currents can flow.

In each of this cases the current flowing is also flowing through the coils - it's a closed circuit! So there is no saving the motor from overheating - the only possibility to achieve this is to brake less...

Regarding regenerative braking - i once read a master thesis describing a self build trike with BLDC. They measured efficiency for regenerative braking somewhere around 15%

Afair for real live values @Marty Backe once reported his battery charge states going up and down the mountain?

[Aneta]

47 minutes ago, Chriull said:

I had once a wheellog log showing braking (dropping speed) while the battery voltage was also dropping. So it seems that EUCs can use the battery for braking, too. Or just they "older" ones did (seldomly).

Regenerative braking is a "soft" braking technique (current wise). The mosfets are controlled so that the back emf generated from the motor gets a bit higher as the battery voltage and by this a controlled current is flowing into the battery.

The "middle" variant would be shortening the motor (coils) - this is imho not implemented with EUCs. In such a case a huge current could flow. Of course by PWMing this can be fine tuned.

The "strongest" variant can be braking actively using the battery. By this a negative torque is created in the motor, so there is a current higher than the shortcircuit current possible. Since this is achieved by changing the BLDC motor commutation, this can be very finely tuned (like the forward acceleration) until full backward "thrust" meaning more or less changing the polarity of the battery in regard to the generated back emf. And by this here currents higher as the short circuit currents can flow.

In each of this cases the current flowing is also flowing through the coils - it's a closed circuit! So there is no saving the motor from overheating - the only possibility to achieve this is to brake less...

Regarding regenerative braking - i once read a master thesis describing a self build trike with BLDC. They measured efficiency for regenerative braking somewhere around 15%

Afair for real live values @Marty Backe once reported his battery charge states going up and down the mountain?

My knowledge of how BLDC motors work is poor, I'm just learning. But is there anything wrong with this calculation and conclusion:

From https://www.ebikes.ca/tools/simulator.html, if we choose "Custom motor", we see some default value for "phase-to-phase" winding resistance, 0.12Ohm. I guess they chose some typical value for ebike motors. Further, in "Custom controller" we see the default resistance of MOSFETs and lead wires 0.03Ohm. Total 0.15Ohm. Suppose, our EUC is 84V and no-load speed is 70kph. We're riding at 50kph, then back EMF will be (50/70)*84 = 60V. If we short phase wires, the phase current will be 60/0.15 = 400A. For active braking with battery, the controller should be able to produce more than 400A of phase current to beat the braking power of just shorting the phases. Are any controllers in EUCs capable of this? Anyone ever see 400A phase current in Wheellog?

15% from ebike experiment seems to be too low, EUCs should be more efficient as braking on downhill is continuous and optimal. I recently did a ride with more than 1km of elevation gain and IIRC at least 40% returned back into the battery, but I didn't make exact measurements.

[Chriull]

1 hour ago, Aneta said:

But is there anything wrong with this calculation and conclusion:

No. Just full breaking is never happening as it would blow the battery, the wires, the controller and the motor... Most probalbe with the controller as the weakest link.

While breaking like with driving the wheel still have to balance the rider, so it happens finely tuned.

[Marty Backe]

11 hours ago, MrRobot said:

Is just going downhill enough to regen or do you need to be braking too?

If you are going downhill, whether actively braking or not, you are recharging. The wheel wasn't providing some level of braking, you would soon be like a runaway truck going down a mountain 

From my extensive experience mountain riding, you get regenerative charging always, but you get more charging if you are going steeper and slower.

[Marty Backe]

I have experienced impressive recharging (25% recharge level) when descending about 4000-feet down a mountain, over a period of a couple of hours. It's actually faster than what you would get with a fast charger!

[Blueblade]

1 hour ago, Marty Backe said:

If you are going downhill, whether actively braking or not, you are recharging. The wheel wasn't providing some level of braking, you would soon be like a runaway truck going down a mountain 

From my extensive experience mountain riding, you get regenerative charging always, but you get more charging if you are going steeper and slower.

I have often wondered about this, (is it better to ride dowhill fast or slow) but never analyzed the wheellog data with enough detail to determine it for myself.

I had assumed that it would be better to ride downhill at a moderately fast speed-(the faster you spin an unpowered motor, the more power it generates), but obviously, the wind resistance becomes a factor at some point. Also the time spent descending is a large factor of course-- the longer it takes the better.

[Aneta]

30 minutes ago, Blueblade said:

I have often wondered about this, (is it better to ride dowhill fast or slow) but never analyzed the wheellog data with enough detail to determine it for myself.

I had assumed that it would be better to ride downhill at a moderately fast speed-(the faster you spin an unpowered motor, the more power it generates), but obviously, the wind resistance becomes a factor at some point. Also the time spent descending is a large factor of course-- the longer it takes the better.

Riding at fast speed downhill reduces amount of energy available for regeneration, since potential energy is wasted on aerodynamic drag. For example, suppose 100kg total weight goes down 6% grade, the g-thrust (component of weight along the road) is about 6 kilos; now if the rider goes as fast as to produce 3 kilos of air drag, this essentially reduces the amount of potential energy for regeneration by half - e.g. riding down a 1000m mountain would regen the same amount as if riding only 500m. The other 500m is lost to Global Warming!

Edited November 15, 2019 by Aneta

[Mono]

6 hours ago, Aneta said:

Riding at fast speed downhill reduces amount of energy available for regeneration

As the saying goes, you can't have the cake and eat it too. Energy one spends on moving one cannot use for regen, obviously.

[Blueblade]

It does make sense, and @Marty's "field testing" seems to confirm it.

I do wonder what it would look like, on graph(s) say riding down a fixed distance on a fixed incline, at various speeds, to see how dramatic the difference is.

I would try some wheellog experiments of this nature if i had the time. (I know just the place to do it too).  Maybe someone esle here, who is a bit geeky like myself, already has... <cough> @Chriull

[Planemo]

On ‎11‎/‎14‎/‎2019 at 8:38 PM, Seba said:

1. At top of the hill start your EUC app, whatever it is. You just need voltage readout. Wait a while to let the voltage stabilize.
2. Start riding downhill, but keep speed at about 15-20 km/h.
3. You will soon realize that voltage goes up noticeably higher than voltage you've seen before ride.

This is because when going downhill, your engine acts as a generator. It will convert mechanical energy from braking to electric power that charge the battery. Electric engine is reversible machine. It will start rotating when you power it up with electricity. Or vice versa - when you start rotating it, it will produce electricity. Our EUCs work in similar way to this - https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricity

Thanks, I understand the above now and I can see how, under light braking, it would work. But say you are doing 40kmh on the same hill and brake as quickly as you dare. Does the same regen process still apply? Or, because you are now asking for a lot more braking, does the motherboard actually *use* power (albeit kind of reverse-fed into the motor) to slow it down?

I guess it would help if we knew how many amps are generated when braking as hard as possible from high speeds. Could all that be fed back into the battery? Or does it just regen only so much and the rest gets fed into the heatsink as a total loss? Or does it reverse feed it into the motor as above?

[Mono]

2 hours ago, Planemo said:

gets fed into the heatsink as a total loss?

The motor seems to provide the perfect heatsink for that purpose.

[Seba]

7 hours ago, Mono said:

The motor seems to provide the perfect heatsink for that purpose.

It may seem so, but there is a one problem. In fact, braking by simply shorting motor windings will soon lead to motor failure. This is because motor windings have very low thermal capacity. Winding-to-core isolation and wire-to-wire isolation limits power dissipation capability due to poor thermal conduction. This is why in non-regenerative, real word braking scenarios external resistors are used (usually bulky and with forced cooling). This however is far from being energy efficient, so now regenerative methods are favoured (by feeding power back to the grid or battery). Of course for short periods of time it's possible to use motor directly as a braking resistor, but this is risky.

Owners of KS-18L, XL and KS-16X may take a look at their CSV logs created with EUC World app (this also applies to older, euc.world WheelLog versions). There are additional fields wh_discharge and wh_recharge (names are self-explanatory). They are not precise values but are very useful to make some research how much energy is drawn from the battery and what portion and when is regenerated back to the battery. In EUC World there is also motor temperature (for compatibility purposes it's placed in cpu_temp column. It can be used to track power losses inside the motor.

I get 167 Wh during discharge and 13,8 Wh during recharge (regeneration) during my daily, 9 km commuting with speed of about 25 km/h, 8-10°C ambient, non-aggressive riding style and with moderate hills on my way. But from statistics I get regenerative load peaking at 50 % or even more of normal load. With new EUC World that I'm testing now I also receive occasional "overvoltage" alarms during hard braking what means that system voltage in my KS-18XL goes above 85 V. And I don't speak about fully charged battery. This means that during hard braking regenerative current may go quite high - during rather normal braking I constantly gets regenerative currents of 10 A and more.

[Mono]

5 hours ago, Seba said:

motor windings have very low thermal capacity. Winding-to-core isolation and wire-to-wire isolation limits power dissipation capability due to poor thermal conduction. This is why in non-regenerative, real word braking scenarios external resistors are used (usually bulky and with forced cooling).

I indeed estimated the thermal capacity of the motor some time ago and came to the conclusion that it should work for a few emergency brakes. This is indeed what I observe in practice: braking gets squishy after a few attempts. That is BTW not different from regular cars AFAIK, where brakes under heavy use will overheat as well.

[Werner]

There's additionally one thing to think about:

Some wheels may do not , as my Z10 does not, any recharging if the battery temperature is to high. 

For example, if the temperature of the Z10  battery is above 42°C it will not be charged. Not with the standard charger and not by regeneration. 

What's about your wheels? Is there any equal problem known already? 

[Struck]

On 11/16/2019 at 9:17 AM, Seba said:

Owners of KS-18L, XL and KS-16X may take a look at their CSV logs created with EUC World app [...] There are additional fields wh_discharge and wh_recharge 

Hi Seba, why is it that users of older wheels do not have that? What I have now is four columns with a bunch of zeroes!
Are you receiving more data from the newer models? afaik you can compute that with the available date (which will be very wrong, since the sampling rate is not enough to have anything accurate)

Best,

[Seba]

5 minutes ago, Struck said:

Hi Seba, why is it that users of older wheels do not have that? What I have now is four columns with a bunch of zeroes!
Are you receiving more data from the newer models? afaik you can compute that with the available date (which will be very wrong, since the sampling rate is not enough to have anything accurate)

Best,

It's an experimental feature I introduced long time ago to analyze if this could be useful for better battery charge monitoring. As I own KS-18L and XL, this feature was added only to these wheels. Later I expanded it for KS-16X. In general wheel must report negative current values for regenerative braking and many wheels doesn't. For example KS-14 and KS-16 report only absolute current values, so there is no way to determine if current flows out of the battery (during "normal" ride) or is flowing into it (when braking or riding steep downhill).