I was confused at first, but I'm seeing this too.

If it helps, these steps are how I made sense of it:
1. Have a network which is stable (no changes to power sources or demand) with just one large flywheel, which is charging
2. Note the speed at which the large flywheel charges
3. Build several small flywheels in a row (e.g. 4 of them)
4. Large flywheel's charge speed drops to about 20% (as expected, I think)
5. Check on small flywheels and wait until they're all fully charged
6. With no more energy going into the small flywheels, the large flywheel's charge speed is... 20%. (inducing weeping and gnashing of teeth)
7. Demolish all of the small flywheels simultaneously and see that the large flywheel jumps back to full speed charging

if you'll excuse me I need to go lament the fact that I stacked horizontal medium flywheels as supports for my large flywheels

ough... so, I did some more testing, and if this is a bug, it's vanilla, and the flywheels will need their own solution.

If you have multiple fully-charged gravity batteries and one gravity battery that is still charging, that last one will charge much faster if you disconnect the full ones from the network.

It's vanilla behavior that all batteries consume (and, if full, waste) an equal amount of the excess energy. You'd never know it when using a bunch of gravity batteries at the same altitude, but flywheel sizes introduce some new complexity.

Hey, sorry only just seen this

Yeah the charge rate is implementing vanilla logic - in fact, all the code that handles power storage (or the 'Battery' class as a whole) is all vanilla code, though some of it is unused at present.

Wow...

I was loosing my mind figuring out why I don't manage to store enough energy during wet season...

What an annoying bug to still exist. uuughh.
At least I know whats wrong now.