Some existing works.

Very few ideas are truly new or unique.  Often they are re-application of work done elsewhere, or more often - ideas and concepts independently developed (it is, after all, a very big world out there).  There are two kind of key concepts for this line of BMS / CAN charging systems:

• Ability to communicate battery status and needs out to many charging sources.
• System enforced coordination of those multiple charging sources.

The benefits of these two include:  Elimination of individual sensing wires from each charging source to the battery (We have  6x sets of voltage sensing wires attached to our house battery, 5x temperature senders..).  And the other benefit is to allow charging resources to work together in a unified way; eliminating the not uncommon teeter-tottering back and force fight between two or more independent sources as they play 'King of the Mountain'.  It also opens up the potential for more intelligent deployment of resources, ala - let the Solar panels do the final finishing charges, as opposed to keeping the generator running under a light load...


Here are a couple of examples of prior work that touches on these concepts:

The 1st illustrates the idea of using the CAN bus to inform remote devices of the battery's voltage/current/temperature status: "Distributed Power Supply Control Using CAN-Bus":
    www.ixs04.aps.anl.gov/News/Conferences/1997/icalepcs/paper97/p155.pdf


And here is a high end Solar MPPT controller who advertises a value of "Don’t waste solar power: all chargers will always be in the same state.." and more:
 

http://www.victronenergy.com/blog/2013/11/15/synchronizing-multiple-mppt-15070-charge-controllers-2/


Both of these are existing examples of the key goals of this 'systems' project, and highlight why selecting a CAN protocol is a bit difficult.  Note that the Victron MPPT controller uses its internal VE.connect protocol extensions to enable the coordinator, not NMEA-2000 - that is just used to report out the aggregated results.