Trying to understand MultiPlus II phase shift behaviour

[Dean_Commodore (Dean Perry)]

I’m going down a rabbit hole trying to understand the frequency-shifting behaviour of Victron systems regarding the management of PV inverters - ESS in this particular case, but the PV assistant case is interesting too.

This particular site is an ESS system located on the fringes of the grid distribution, so the grid voltage can be all over the place, from below 230 volts to 258 volts or so (the distributor’s stated “service level” is 230±10%). Not a lot to be done about that problem, other than paperwork and ask for transformer taps to be adjusted (…)

The grid is often out of tolerance with respect to the grid code on the MultiPlus II, and it disconnects pretty frequently. When connected to the grid, the Fronius will also complain periodically (Error 102) before the disconnection.

The system has 2x 8kVA MultiPlus II (L1,L2) and 1x 5kW Fronius on AC Out (L2). There is 1x RS450/100. About 50/50 split on the solar between those. 6x Pylontech US5000. MultiPlus II are on firmware 2621556, the Fronius on fro36120 (…), Cerbo on 3.67

In the pictures below,

• the Pylontech battery system SoC is around 20%
• Pylontech CCL is 480A, so willing to take on more than 20kW of charge.
• The AC loads are continuous, and about matched with best-case solar production (at around 8-11kW).

I’m pointing that out because I don’t understand what’s going on with the frequency shifting - it should not want to throttle the Fronius for any reason that I understand.

In this first pic of my Node-RED dashboard, you can see that I’m trying to force the MP-II to ignore the grid (/Ac/Control/IgnoreAcIn1) - but this appears to make no difference for the behaviour in question.

The two “Okay to use AC” indicators are the result of a function that checks if the grid is not between 225 and 250 volts.

The Grid Voltage values are taken from an EM24.

So to begin with, the frequency here is good (49.9Hz), and the Fronius is in one of its startup states:

This next pic is where the behaviour in question happens - short time after the above, but the MultiPlus-II frequency has shifted to 52.9Hz briefly like maybe for three seconds. But it is enough to upset the Fronius, which flicks back to trying to start up again.

These next two are the same situation, but without me trying to force the AC In to be ignored - I’ll note that after I changed that setting the MP-II briefly tried to connect and then decided not to do that (2nd pic):

The MultiPlus II flicks to 53Hz briefly and then back to 50Hz - not stuck on 53Hz (not that problem). Plenty of load, plenty of headroom on battery voltage, lots of places for the power to go. It only does that when the input is in the “ignored” state.

When the grid is actually isolated (with a breaker, not “ignored”) that situation changes. The VRM logging resolution doesn’t quite capture the periodic “flicking” behaviour, but the graph below is pretty obvious. When in this state, fully disconnected, the Fronius will quite happily ramp up production:

For completeness, I will mention that the frequency “flicking” and the MultiPlus II relay test “clickety clack” appears to happen roughly in the same period. This behaviour does not occur when the grid voltage is more reasonable, and the system is quite happy in that case.

So I suspect that it’s something to do with the grid test on the MultiPlus II, but I’m not sure what is going on.

Thanks for reading so far (!), let me know your thoughts.

[lxonline (LX)]

The frequency shifting is based on the dc voltage.
And for some reason where the battery is in the flat part of the curve is where it ramps up and stays stuck up even though soc is clearly dropping. (Or at least that us my current theory). It could be because of the support on the dc bus of the dc coupled solar?

I have one system (with a solar 50/50 split between dc coupled and ac coupled-it is not a fronius) that occasionally seems to stick at throttled which does drive me crazy as you say there is clearly a need. (Solved by the fudge fix -making a brief connection then disconnecting.)

When the inverter briefly tries to connect it on its own resynchronized with grid frequency then disconnected and carries on at that frequency.
As an FYI we don’t have a grid code so its not that.

The other fix i have seen tried by other users on the forum is adjust the throttle frequency in ess assistant so it throttles but not to no production.

[BjoernK (Björn Kellermann)]

I have 5 PV inverters on ACout and 1 MPPT.
When I disconnect the grid, frequency rises to shutdown frequency before reducing to 50Hz.
If the battery is full, frequency is reduce to match below current load. So the AC-PV serves a part of the load and the MPPT does does fine controlled adjustment.

I noticed one PV inverter is not working. After checking the grid code of the micro grid inverter I noticed the resume frequency is missing the decimal sign. It returns at 502 Hz and not at 50.2 Hz. As I cannot change the grid code of that device, this single one is not working during grid failure.

As your grid seems to be very unstable or out of spec, why not disconnecting from the grid and recharge the batteries with a telco PSU?

Than your micro grid will be in spec and your devices/home has no risk of overvoltage.