question

The rule only applies to the PV inverter, it is not relevant to the DC chargers.

Your pack does need to be able to cope with sinking power in the event of an issue.

You mean US3000C?

At 8 batteries you are on the edge of the sizing for 15kVA when grid attached, for off grid (ie grid failure) your’re a bit under. Ultimately the batteries will start moaning if they aren’t happy.

Alternatively, just limit the feed-in a bit.

First: the topic opener @dognose writes that he has a Fronius Symo 12.5 (= 12.5 kVA) with 8000 watts of solar. I interpret this as 8 kWp solar panels on the 12.5 kVA PV inverter. The smaller value of 8 kW then applies for the factor 1 rule.

I then calculate as follows:

The batteries are discharged to a minimum of 48 V, for example. If in such a case the grid fails during maximum PV production, the battery must be able to absorb up to 8000W / 48V = approx. 170 A for a short time. But only for a very short time, as the power of the Fronius is then reduced after a few seconds by increasing the frequency (MG50).

If the batteries are designed in such a way that they can always absorb at least 170 A for a few seconds, everything should be fine in terms of battery capacity.

What I really don't like about our considerations so far is that we don't charge the battery to 100% so that we still have enough buffer in the battery in the event of lost grid until the Fronius power is reduced by increasing the frequency. Firstly, this wastes battery capacity and secondly, the battery is no longer balanced.

I have therefore thought again about how the battery can still be fully charged despite being too small according to the Factor 1 rule:

As already mentioned, I have a combination of a 14.3 kWh battery (Gobel Power) and a Fronius 10 kW inverter.

We have already discussed this for batteries that are not full: Factor 1 rule would actually prescribe a battery with at least 10 * 3.2 = 32 kWh for Fronius 10 kW. However, if the battery is never discharged below 48 V, a smaller battery is sufficient as long as the BMS reports CCL of at least 10000 W / 48 V = 208 A. Or if you have a 200 A BMS, it is sufficient to permanently limit the Fronius to 48 * 200 = 9.6 kW.

However, this 200 A current surge may only last a few seconds at the time of the grid failure until the Fronius is regulated down to the charging current limit set in the Victron DVCC by increasing the frequency.

However, grid failure at full battery status is still a cause for concern when the BMS reports a lower CCL to the Victron. I have now come up with the following idea for this:

According to the Factor 1 rule, the Fronius may officially only deliver 14.3 / 3.2 = 4.5 kW with a 14.3 kWh battery.

My plan would now be to have the Fronius produce 10 kW at full power as long as grid is available and BMS reports CCL >= 200 A.

However, as soon as the BMS reports a CCL < 200 A when the battery is getting full, I will limit the Fronius output to 4.5 kW. It then officially complies with the factor 1 rule. It should be possible to specify the power limitation of the Fronius via Modbus TCP, which could be implemented in Node-RED.

As soon as the BMS reports CCL >= 200 A again due to an empty battery, I increase the power limitation on the Fronius back to the full 10 kW via Modbus.

So far my current plan...

In my opinion, there is still enough free "space" in the battery left to take the load of 8kW for let's say 3 seconds, if the battery is not at 3,65V/cell.

We all know, that at the steep part of the charging curve (approx. above 3,45V/cell) the capacity gain is only very small.

But nevertheless, a peak of 8kW (that is 151,5A @ 52,8V) at a sudden grid failure, would only provide a capacity increase of about 0,126Ah within 3 seconds.

This is only >0,1% of the whole stack capacity, so should be no problem at all if the Fronius can be reduced by frequency shift within a few seconds.

So if the battery is not always loaded up to a max voltage/cell, I see no problem there.

The point is, that you define the 100% SoC state.

And as long this 100% SoC point is not defined at the absolute max. rating for the cell voltage at 3,65V/cell, then the battery is always physically able to take some energy without harm.

The question is now, what the BMS will make at this high SoC level and if the MP2 can still drop the energy into the battery if the CCL is set to 0A?

So, today I ran into the situation of an accidential grid-down while there was around 4000 Watt Feedin and the Battery was at 100% SoC.

I'm still here, system didn't explode :P

The system almost immediately turned off the MPPTs and the Fronius, and battery charge spiked to about 3000W despite it's SoC was 100%, just to switch over to provide the 2500 Watt consumption a few blinks later.

AC-Coupled PV remained turned off and the MPPTs where slowly brought up to meet current consumption, normalizing battery discharge down to 50ish watts.