Be that as it may it begs the question WHY???
I'm not certain that I've "got" it, but let me see - I'm sure Chris will correct me if I'm way off beam here!
He has a substantial number of batteries. ok, that's a given.
He has a 24V system. That's historical fact.
He has established that connecting cells in parallel at the lowest level practical for the cells/batteries in use helps.
Sooo... break the "substantial" number of batteries into two equal groups.
Call them bank1 and bank2.
Connect all bank1 cells in parallel.
Connect all bank2 cells in parallel.
He already had a "battery bus" positive and negative rail.
He's now introduced a third rail. Lets call it "midpoint".
Bank1 connects from "battery bus (negative)" to "midpoint".
Bank2 connects from "midpoint" to "battery bus (positive)".
We now still have the full 24V as before, but we also have a nice central reference.
Adding two identical voltmeters allows measuring bank1 and comparing to bank2.
Since the two banks are in series, any current into or out of one must be identical to the current into our out of the other, and since the two banks are identical numbers of "identical" and same age batteries, they should stay at an equal SoC.
If one cell in one battery in one bank should fail - open or short - the capacity of that entire bank will be reduced - not by a lot, but by a measurable amount. And given the constant and substantial load the banks get in a day, should result in a detectable difference fairly quickly (if anyone is there to watch).
I was thinking that the meters don't really give enough precision, and that I'd like to see a centre-zero meter and a simple opamp driver to show imbalance, but thats a lot of extra complexity compared to what chris has built - and heck, at the end of the season we should have a good idea if it's working as he hopes!
So, Chris - how far wide of the mark am I?