Due to the recent events in the German grid, the man made outages in Berlin and Reutingen, and the sabotage events trying to short cirquit the high voltage land lines made me do trial runs for grid loss scenarios.
My system is a 3 phase Multiplus II 48/5000 with Enphase iQ7A solar generator and Pylontech 3000c batteries. I am able to take each phase individually off grid.
Switching of any of the 3 phases puts the system in grid loss mode which is expected and works well. For my “off grid test weekend” I switched off all 3 phases and everything was fine. During the day the battery was charged and the Multipluses throttled down the solar generator with frequency increase when the battery was getting full. (I automatically switch of the solar generator if the SoC is 95% but that’s another story).
I live in a part of the village which was built in the seventies with underground power lines, now 50+ years old. Every second or third winter there is a single phase failure and for the repair, the power company has to temporarily switch off power. This are the outages which are actually happened and ESS covered them all.
So I was also “simulating” singe phase outages and then I discovered the following:
As long as L1 is off, Grid Loss works as expected, even if L2 and/or L3 is present.
As soon as L1 is on, L2 and or L3 is off, the solar generator is unable to operate..
Looking for the cause I was monitoring the Multiplus output frequency and discovered that about every 25 seconds the frequency changes to 51,5 Hz and then drops back to 50,0 Hz repeatedly. This prevents the solar generator operating (forced switch off frequency) which is bad if not discovered.
It seems to me that the system tries to adjust Multiplus output frequency and phase to the grid frequency and phase (PLL) as it is the case when all 3 phases of the grid are back, and the system switches back to grid mode.
Is that new finding, a known problem or a wrong configuration on my system?
It is usual with 3 phase equipment to install an under-voltage trip, and / or missing phase detector, such that if one or more phases go undervoltage, the complete system is disconnected. This will take care of the problem.
Taking the whole house from the grid would take care of that for sure - but that would be a manual task, at least for me.
And partial grid failures like OP describes do occur: a few years back, a farmer did mow some meadow, which had not-yet-connected electrical wires sticking out (as preparation, if those lots are ever to be built upon) - result: Phase 2 was missing. Didn’t help we were on vacation and so all I could do was watch my UPS getting depleted from afar - no Victron system yet)…
Yes, they do Switch as a group, all 3 MPII do an impressive klonk simultaneously (6 relays do make noise)
However I think the Multiplus @ L1 has a leading role for the group which is important for the resynchronization with the grid before closing the relays when all phases of the grid are back.
How am I supposed to understand this? Does that mean that if, for example, L2 or L3 fails, L1 will also be disconnected?
That shouldn’t be necessary in a three-phase Victron system. As soon as the configured grid parameters are exceeded (voltage, frequency), all three MultiPlus2 units must switch off.
For that to work smooth you would also need automatic phase selector to choose any from available phases to power this phase detector and relay as this is an NO relay maning that powerless it is closed. When no power on any phase this is ok, but if there is power, but not on all phases in order to provide stable voltage supply for this circuit you should also have an automatic phase selector like PF-431. Then it would energize the relay (disconnecting all phases from multiplus input) only in case there is voltage, but not good enough (like one phase missing or voltage outside spec)
But main question here is - WHY do we need something like this?!? I mean, Victron should be able to detect that grid voltage on any phase does not meet spec and switch to inverter mode. That is one of its major funkctions. And when in inverter mode it should HOLD STABLE 50Hz at least untill battery full… Regardless of voltage presence on L1 or L2. As long as not all 3 phases are in spec it should not even try to sync with grid.
The phenomen observed by @power4me looks like either a fault or design error and it should be analyzed by Victron staff
This device will switch de relay open if there is an over or under voltage or a fail of either phase.
If only one phase fails it opens, if 2 phases fail it opens, if 3 phases fail it opens, if 3 phases have a blackout only then it stays closed, this as this device gets its power supply from either one phase…
Exactly what you want to protect your inverters.
Be aware, the transformer (power supply to the PCBs) in your Victron inverter is made for 260V, above 270V they might fail (see other topics), also even with the internal relay off, there is also a measuring input of input 1, so also there things can fail if overvoltage.
Which it does, but still its internal trafo and the measuring line are connected to the ac input…for that you need such devices.
Europe has already for years an unstable net, millions of such devices are since installed.
Imagine you just have a net connection…your washing machine switches off due to high voltage, no, neither your telly, etc, therefore these devices are installed before this net enters your house…so, also before your inverters, and, all, not only Victron.
As mentioned above I have individual 50 A circuit breakers in front of each MPII AC input as suggested by Victron. Therefore it is easy to simulate all possible combinations of grid loss.
Can someone with a similar setup and the means to measure the frequency on one of the MPIIs AC output, try to confirm my findings?
Combinations L1 off L2 off L3 off / L1 off L2 on L3 off / L1 off L2 off L3 on / L1 off L2 on L3 on show:
for a couple of seconds, (non excacly measured because of the slow update rate of my meter)
Interestingly the 49,963 Hz was exactly the frequency of the grid before switching one of the phases of and after connecting back to the grid. It seems that if L1 is present the Multiplus-II tries to track the grid frequency even if the system is “off grid”. Why it periodically switches to the forced switch off frequency of 51,5 Hz is not understood.
In my opinion this tracking should only happen when all 3 grid phases are present and not before.
It would be really helpful if someone could confirm this behavior.
In the described scenario, there were no load dumps. With L1 present it happens periodically like a clockwork for as long a I tested. The frequency shift itself wouldn’t bother me if it weren’t the 51.1Hz which forces, by regulations, the switch off of the solar generator.
Without L1 presence I see the 50.001 Hz until throttling PV power due to high SoC everything works as expected.
I think it is a problem of the synchronization algorithm acting as if the grid is back even when not all phases are present
Yes, that’s the rock solid solution, covering also other aspects. Downside, at least for me is:
I do not have the space in my meter cabinets
It’s expensive
My electrician an me built the system as per Victron ESS recommendation. There is no mentioning of PNM31 or the like in front of the 50A circuit breaker. I also do have an SLS and Flahstrap before the grid meter as required by the grid company and VDE.
I see no technical nor, regulatory reason that the MPIIs in 3 phase ESS mode could not handle that issue properly. There is no reason to treat L1 other than L2 or L3.
Shutting down when a single phase is lost is all well and good, but wouldn’t it also be possible simply to continue using the phases that are still available and use the inverter only for the missing phase?
It shouldn’t really be that much of an issue. But I suppose that would then go well beyond the scope of this topic.
Normally, it doesn’t take that long to resolve a problem. But if it did, it would be nice if the battery didn’t run down so quickly. In addition, the MP2 operating in inverter mode could then be supported via the 48 V bus by the other two until the phase came back.
This has been asked and answered several times previously.
The ability to do this requires ESS not to be installed, as that wants to bridge earth and neutral on grid loss, something that is problematic when some units are present/have grid.
I fully agree on that! I would expect from Victron to wither connect with grid (if all 3 phases are in spec) or to work in island mode with perfect 50Hz as long as there is possibility to charge battery from ac coupled PV regardless of situation on grid side. Cannot really understand why there is such issue. Cannot also simulate this scenario as my current setup does not alow that.
