Capacitors and regenerative braking

[sophiecentaur]

Without some serious experimenting and measuring, you cannot tell just what the arrangement is achieving. Without thinking too deeply about it, I would worry about any long term harm that the capacitors could be doing to the batteries. Lions need to be charged right.

 

[essenmein]

The V/C characteristic of a Capacitor is a constant slope (the Capacitance. For a Lion battery, the drop in volts from near full to dying is pretty much a horizontal line. From full to usably empty, the battery volts only drop by around 10%. I can't see the similarity there.
I could imagine that, if weight is not a problem, with the battery fully charged, carrying some Capacitance could give you some advantage with acceleration for a short burst but, with the best will in the world, you can only get 10% of the charge in the Capacitor. But I don't see a problem with including some smart switching in a design that's obviously going to be a bit pricey in any case. (I'm only arguing against the simple idea of parallel connection.)
One point in favour of the SuperCapacitor is that the actual Energy capacity is all at the top of its voltage range ( Energy = (CV²/2) so you don't have to nearly discharge it. Any installation with SuperCs in it is going to have some fancy switching control in any case if the user is going to have any idea what the accelerator pedal will do for him.

Clearly a Li ion is not actually a capacitor, but their dynamic impedance is not that different during short current pulses where the V is assumed to be more or less constant, after all both are used as "short term" voltage sources. Now I'm comparing large electrolytics to Li ions, not something like a film or ceramic where you can get v low ESL.

Also its quite reasonable to model a Li ion as a capacitor with the right esr/esl and the capacitance value being determined by the energy stored in the battery between Vfull and Vempy.

Paralleling capacitors to a battery for short bursts would make sense if the battery source impedance is "slow" vs your application or you have very large inrush currents that might damage them (eg the starter batteries in the links above), but you would have to measure how much the volt dip is during said pulse at the inverter terminals to determine if this has any value. Keep in mind that any motor drive is already going to have some pretty high quality capacitors on its internal DC bus to deliver the xkHz PWM current and keep that HF current local to the inverter. Basically for transient current its an impedance matching game and the capacitors would have to approach the source impedance of the batteries for them to have any worth while effect.

To me the OP is not pushing his batteries (rated 150A and only pulling 100A) and unless the voltage dip is significant at the inverter terminals a capacitor would likely have minimal effect for transients and would need to be huge to provide noticeable energy storage increase.

IMO determine if there is even a problem with the set up as is by measuring before adding more stuff to it, if its a hobby and you just want to have the caps then fine but in any sort of product development there'd have to be a pretty compelling case to add them to warrant the cost.

 

[essenmein]

Without some serious experimenting and measuring, you cannot tell just what the arrangement is achieving. Without thinking too deeply about it, I would worry about any long term harm that the capacitors could be doing to the batteries. Lions need to be charged right.

No problem with charging the li ion with parallel caps, battery management systems monitor cell voltage and a parallel cap won't affect the end voltage level, just the time taken to get there.

 

[sophiecentaur]

No problem with charging the li ion with parallel caps, battery management systems monitor cell voltage and a parallel cap won't affect the end voltage level, just the time taken to get there.

But where would be the advantage in using a Capacitor compared with just a bigger battery bank? Do you not want to get some benefit from the fact that excess motor volts can be used to put charge into the capacitor without invoking a regulator to protect the battery? The QV slope of the capacitor will severely limit its usefulness unless you can manage the Energy. Connecting a battery straight across it prevent the capacitor volts from going high enough for it to store excess energy.
This may be less harmful than I thought but it still doesn't allow the capacitor to be used to advantage.
I'm assuming that there will be some cost-benefit consideration when the value of the capacitor is chosen. It's true that a big enough capacitor could store loads of energy at battery volts but the 'square law' relationship means that, say 10% higher volts on it than the battery volts would mean 21% extra stored energy in it. I would have though it would be worth going for
The 'equivalent capacity' of a battery is enormous, based on Q/V, if the Volts don't change much.
"just the time taken to get there". If they are in parallel, what will make the volts increase faster?

There is nothing wrong with contemplating a hybrid system if you are prepared to make full use of the two storage media. One thing that would be useful would be to charge the battery from the C when the system is not being used. That would involve a boost circuit - to augment the buck circuit for when the C is highly charged.Unless you are prepared for some testing, you would have no actual quantitative idea about whether or not your modification is doing what it could to should.

 

[essenmein]

But where would be the advantage in using a Capacitor compared with just a bigger battery bank?

IMO there isn't one unless you have a very specific problem to solve.

 

[RedBaron]

Clearly a Li ion is not actually a capacitor, but their dynamic impedance is not that different during short current pulses where the V is assumed to be more or less constant, after all both are used as "short term" voltage sources. Now I'm comparing large electrolytics to Li ions, not something like a film or ceramic where you can get v low ESL.

Also its quite reasonable to model a Li ion as a capacitor with the right esr/esl and the capacitance value being determined by the energy stored in the battery between Vfull and Vempy.

Paralleling capacitors to a battery for short bursts would make sense if the battery source impedance is "slow" vs your application or you have very large inrush currents that might damage them (eg the starter batteries in the links above), but you would have to measure how much the volt dip is during said pulse at the inverter terminals to determine if this has any value. Keep in mind that any motor drive is already going to have some pretty high quality capacitors on its internal DC bus to deliver the xkHz PWM current and keep that HF current local to the inverter. Basically for transient current its an impedance matching game and the capacitors would have to approach the source impedance of the batteries for them to have any worth while effect.

To me the OP is not pushing his batteries (rated 150A and only pulling 100A) and unless the voltage dip is significant at the inverter terminals a capacitor would likely have minimal effect for transients and would need to be huge to provide noticeable energy storage increase.

IMO determine if there is even a problem with the set up as is by measuring before adding more stuff to it, if its a hobby and you just want to have the caps then fine but in any sort of product development there'd have to be a pretty compelling case to add them to warrant the cost.

This made me understand so much better what's going on and what to look for. Thank you.