Why big cap bank in parallel with battery on electric vehicle?

[zenerbjt]

Supposing you have a 5kW-15kW  battery operated electric vehicle, or drone, or ship. Battery voltage is 14S or 24S lithium. (48V or 82V). Why would you ever put a very large electrolytic capacitor bank right by the output of the battery? (ie, in parallel with the battery)
I have seen this on a  product, and wonder why they do it?

[Doctorandus_P]

Those motors are PWM controlled.
When PWM is "on" it draws a few hundred Amps from a "reservoir", and motor winding current increases.
When PWM is "off" the current is circulated through the (inductive) motor windings, and slowly decays.

by controlling the on to off ratio, the average motor current and power is adjusted to what the motor has to deliver to the load.

Batteries do not like to switch between 0A and 200A at a rate of a few hundred Hz or kHz, and the capacitors smooth this out to a near DC average current, which does not destroy the batteries within a week.

Then there is also the crest factor of the PWM and wire resistance.
Power loss in wires (constant resistance) is the square of the current, multiplied by R.

A DC current of 20A over 10mOhm resistance is:
20*20 * 0.01 = 4W of power loss.

A DC current of 200A with a 10% duty cycle is:
200*200 *.01 * .1 = 40W of energy loss in the same wiring harness.

Yet another reason is EMC emissions.
The motor itself and the electronics are hermetically sealed and shielded in cast aluminimum.
Long wires to the batteries like to work as antennas and high di/dt current changes radiate lots of electrical noise. So apart from the capacitors, you'll also find inductors to further filter the spikes.

Then there is also self inductance of the wiring. This may become such a hindrance that you can't even suck 200Amps out of the wiring at high PWM frequencies without significant voltage drop over the wires themselves.

[digsys]

Excellent explanation - I'd also like to add ...
Eons ago when we were designing our first LiIon packs for our solar car, I installed a few test instruments (spec analyser, dso, imp analyser etc) in the vehicle, and ran it around our test track under various conditions. The idea was to find out "real world" answers.
First observation (without any filtering) - even with the most efficient motor controller / wheel motor - there was 40V+ of full range spectral noise sitting on the pack, and far worse under regen braking. No battery can put up with this type of "noise". They are usually only  good to ~ a few hundred Hz, depending on chemistry. As mentioned - it degrades the battery life.
Then we started adding capacitors - bulk low ESR, RF, ceramics etc etc. These definitely made a HUGE difference, BUT up to low 10s of 1,000s uF etc, even they overheated and often were destroyed. I have video somewhere of some spectacular results :-)
- and this is just on a car weighing 180Kg and running 98.5% efficiency .. things get way more serious on RW EVs. That's why the cap banks are quite large - it's not like just adding a filter cap and that'll do. CAP banks can be in order of many Farads.
Also - IF you have regen, batteries HATE regen, they can't take the inrush currents. On MANY EVs, that don't have cap banks, regen % of what they are capable of is serious restricted, based on what the battery can handle. A waste in a way, but better something than nothing.