[RE-wrenches] Open Loop vs Closed Loop

[Zeke Yewdall]

For the definitions of open loop vs closed loop according to how we use
those terms for lithium batteries:  the definitions you gave are according
to engineering control systems theory.  But these terms have kind of been
adopted by the solar industry without really applying them correctly in my
opinion.

Open loop batteries there is no communications between the inverter (or
charge controllers) and the battery.  The inverter/cc makes its decisions
based on voltage and current -- just like we always did on lead acid
batteries.  It might be more advanced, having a battery current monitor or
an AH counting monitor on the battery, like some of the victron systems and
I think Trimetric had a charge controller that worked with their meter, so
adjust charging based not just on the voltage and current at the charge
controller, but also voltage and current at the actual battery terminals or
calculated SOC.  But there is still no communication with a battery BMS.

I would still say that there is feedback in this setup.  The inverter/cc is
continuously adjusting the current based on what the voltage is doing.  If
the voltage tries to go up above absorb setting, it reduces the current to
keep it from doing so.  If a load turns on and voltage drops when it's in
float, it increases current to keep it at float voltage.  So it's not
really open loop in the strict sense of an open loop control.  An open loop
charge control would be one that puts in 20 amps for 4 hours on a timer,
and pays no attention to battery voltage, for example.

What we call closed loop lithium battery communication is really taking the
control away from the inverter/cc and giving it to the battery.  Instead of
the inverter/cc watching the voltage and deciding what to do... how much
current to put into the battery.. the battery BMS sends a set of rules to
the inverter saying how much it wants.   It tells the inverter -- give me
as much power as you can.  or, limit your power to 5000 watts now, or I'm
full, only do enough power to handle the load.

I'm not sure exactly who makes some of the decisions... for example does
the battery tell the inverter to turn off for low battery when the BMS sees
below a certain SOC, or does it just tell the inverter what the SOC is, and
you program the inverter to turn off below that SOC.  Effectively it
doesn't matter, as when we are programming it as a user, we don't actually
know which of the devices is making the decision, but we just program the
result of the decision... i.e. turn off when SOC drops to 10%.

There is still another level of control the BMS has, whether in open or
closed loop.  If the inverter does not behave... either pushing the battery
voltage out of acceptable ranges in open loop, or not following the
commands of the BMS in closed loop ... then the BMS will turn off the
charging and discharging contactor and turn off.  It's protecting itself...
sometimes with disastrous consequences for the inverter and rest of the
system.  I once had a battery turn off the contactor due to overvoltage
from a 6kW DC generator that was not throttling back charging properly.
The DC generator spiked to 75 volts due to sudden loss of load, which fried
everything in the system that couldn't take 75 volts.  Amazingly, the
radian inverter survived it.  My theory of operation is that I never want
the battery BMS to feel like it needs to open the main contactor to protect
itself...  that often means a on site physical restart.  Worst case can be
a black start situation where you have to manually bypass the BMS, or
destroyed equipment from charging sources spiking without a battery to
smooth out the voltage.  So, in open loop I always try to program such that
the BMS never gets unhappy... stop charging before it hits high voltage.
Stop discharging before it hits low voltage.  Etc.    When in closed loop,
you hope that the battery having control over the inverter prevents
situations like this, but they can still happen, especially on the low
end.  SOC indicated by BMS's can drift from reality, just as SOC from AH
counting meters can over time, if they don't reach reset conditions
regularly.  And many battery BMS's are blind to very small currents... less
than 1 amp for most BMS's in the 100+ amp size, and less than 0.2 amps for
lower amperage ones.  This means that a very low discharge from an inverter
tare load but no AC loads on it, or DC LED lighting, can slowly discharge a
lithium battery without the BMS counting down on the SOC.   After a few
weeks, the actual SOC may be 30%, while it still reports 100%.  This means
you can start discharging it and have a voltage dropout while SOC is still
quite high.  At this point, the battery BMS will usually cut off due to low
cell voltage, and suddenly recalibrate itself to 0%... when it may have
said 60% just a minute before.  So, even with closed loop comms, you can
still have issues if you don't reach a true 100% every week or two I find,
due to SOC drift.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20250905/14a59a68/attachment.htm>