when a LiFePO4 cell reaches 90% SOC under charge, progressively more of the charge current is going to building a surface charge. (it represents about 1-2AH when i charge to 3.9vpc on my lfp40s).
over a long period of time (more than 10 hours in my experience) this surface charge is lost to the internal shunt resistance as heat.
the internal shunt resistance losses are continuous, and of course greater when the voltage on the cell is higher (such as when surface charge is present)
The internal shunt resistance varies a fair bit between cells, of course you already know all this, this is just for some background.
now, to a degree shallow charging will work for a time.
however cells with the lower shunt resistance will tend to a higher SOC relative to the rest of the cells in the pack.
Once the cells with the lower shunt resistance are also the highest SOC cells of the pack, this method of balancing will no longer work, and will become counter-productive, of course as a temporary measure, not so bad.
ive found the tendancy for imbalance with lithium is much less than that of lead acid. single cell chargers are far less efficient, and won't do a *better* job than a resistive cell balancer setup. ( switch mode power supplies converting 240vac to 3-4vdc arent particularly efficient )
of course im always looking for cheaper to implement approaches to battery management. my BMS cost as much as my pack did (for my scooter anyway)
Lithium plating happens usually during overcharge, when the voltage across any part of the cell exceeds a certain voltage (4.3v i think? i forget, ill have to look it up again, but outside of a cells operating range).
i have yet to actually destroy a cell in this way, so i dont really know what one would do.
during the actual overcharge ive heard cell heating may be a symptom.
you will notice higher internal series resistance and lower capacity after such an event is what i suspect.
I am not saying shallow charging is a true bms but will bring and keep healthy cells within 5% charge tollerance. Of course this only works with Lithiums because of all their tendecies to slowly build a surface charge while other further discharged cells continue to charge up at a normal progression rate. Again it is not a bms just an acceptable solution compared to charge and burn........
Also though a shallow charge does not have a capability to take wildly imballanced cells and bring them close. This is the reason I suggest a break in cycle with the algorythm previously suggested.