Good evening, after tornado-like storms, a grid backup generator near me is making a racket, supplying power at 51.7 Hz and turning my attention towards my own energy generation: forced island operation. My “backup power” MP2/5 kW passed the 51.7 Hz through and consequently switched off the PV inverter. The three MP2/5 kW units operating in three-phase mode automatically switched my system to island operation. Is this different response an edge case?
The storm has shown me that I need to work on a good solution for properly curtailing solar power! The core of the issue is probably that I need to find a solution for obtaining accurate SOC data. So far, there is no BMS data available for the Cerbo or the MP2s (only third-party BMSs—JK, which sends the cell data to my phone via Bluetooth).
It would have been nice if frequency control worked as well as ESS operation, which has worked well for a long time—but that might have made me complacent, since enabling grid feed-in in my setup (with no exact assignment of the current cell voltage to the switch-over to feed-in) should be solved by next winter at the latest. In the current situation, I am working with 19S battery packs and am content with a maximum charging voltage of 64 V.
Does anyone have a workable approach that would mean I do not have to rebuild everything? I like the JK BMSs—with active balancing—for 20S. Are there new ones available now with RS485 and power FETs?
For general information: after 20 years of feeding power into the grid, I started experimenting with one phase two years ago and then added the three-phase setup—I really ought to have more time.
I look forward to any replies or questions that can help me move forward! Many thanks! Best wishes,
Matthias
Hi Dirk, in my “Assistant” variant, I’m being asked about the nominal frequencies of the PV inverters. Where (on which page) is this option normally requested?
The MultiPlus SOC value should be used for now — one MultiPlus was in a temperature pre-alarm state (flashing LED), but the frequency was fixed at 50.0 Hz.
Maybe I should look into the SmartShunt; that would be a good starting point.
I’m afraid that, although a SmartShunt is a good monitor, its data may not necessarily be usable by the “ESS Assistant” (otherwise, perhaps, the device would be included among the “BMS variants”)??
So, is the option shown actually enabled? You still haven’t answered that clearly.
Perhaps you could also give us a bit more information about your system and its settings.
In my opinion, more than 17S with LiFePO4 cells isn’t realistically possible with Victron, because it simply can’t be charged to a high enough voltage.
Good morning Matthias, I’m very pleased that you’ve decided to speak up here! First, regarding the borderline nature of 19S. In my opinion, my LiFePO4 batteries (I’ve had various sizes for years — 150Ah, 200Ah and 280Ah) start to diverge noticeably in their state of charge from around 3.4V; up to that point they remain synchronised to within a few mV. I can observe this very clearly via Bluetooth while charging. Since I’m currently self-sufficient apart from January and February, my “gut feeling” tells me to spare the battery when, for example, charging to 90% is sufficient. I also think that charging efficiency becomes significantly worse at the “high” charging voltages. I’ve never heard that LiFePO4 batteries need to be fully charged for a long service life.
Regarding the option shown: Wasn’t my answer correct? If my 10kW Symo is connected to the AC out of the MP2, don’t I have to select “Yes” in the assistant?
The noise from the diesel generator 150m away is expected to continue until Monday or Tuesday — forced island operation. Today, in the dark, I installed vacuum high-voltage relays in two strings; yesterday I tested whether they still worked after being stored for many years. So, for the time being, I’m in a good position when I’m at home. I’m sure you’re also familiar with the arc formation that occurs at around 800V DC and 10A of solar current with a primitive disconnection? Matthias, what do you think about using the SmartShunt as measurement information for ESS in my setup (not just for curtailment)? If it provides data, that would be useful at the moment. As for additional resistances: the three MP units are initially connected in parallel using 50mm² cable, and three battery banks are positioned next to one another using 35mm² cable. Would the measuring resistor (smart meter) have to be placed in the charge/discharge current of the “master”?
It would be much more elegant if the MP units could measure the battery voltage precisely via their existing voltage-sensing probes. I experimented with this two years ago, but if I remember correctly, the available options aren’t being properly utilised.
How can I easily read out my firmware versions? About a year ago, I brought everything up to a consistent, up-to-date version: Cerbo, MP2.
Thanks for any suggestions and input! Matthias
I forgot to mention the EM24 (hopefully I’ve got the name right), which measures the three phases really well and thus provides an important foundation for the ESS.
Dirk — I’m familiar with the 1:1 rule. It’s also the real problem facing the energy industry: energy consumption while nature supplies power “unpredictably”.
Dear Holger, what exactly is meant by balancing here? To me, the term refers to influencing the charging of cells connected in series, which have different charging voltages due to their differing characteristics (internal resistances), by accessing the individual cells so that each cell has as nearly the same charging voltage as possible.
The “emergency power” MP — my uninterruptible power supply in the past — is a partial system on one phase with around 5 kW peak, currently using the indestructible IG 60 from Fronius, which has lasted for 20 years (they shut down when they detect frequencies above 51 Hz).
The other part consists of three MP2/5 kW units, which can be put under strain by around 10 kW peak (the PV inverter is a 10 kW Symo) when, for example, both e Go lifes become fully charged and there are no other loads. Today was particularly exciting because of the constantly changing weather, and I was away. Nevertheless, the two Cerbos are now showing 85 and 94 per cent, which I only managed to achieve by switching off the string — as described above.
Balancing only takes place when the battery is almost fully charged… it doesn’t happen during the intermediate phase… all the cell voltages are almost the same then… simply because of the high charging/discharging current.
And it’s only from around 3.42–3.44 V per cell that the current drops and balancing takes place… it’s also important because this is when the SOC is set.
Yes, “YES” must be selected there, and my question was what you have selected there (and for the other settings).
Of course, you can do whatever you want, but there have already been a few attempts here with 18S LiFePO4, Li-ion or sodium-ion batteries. In all cases, you are operating at the upper limit of the Victron devices, and charging then no longer works properly.
You should fully charge the cells regularly, bringing the cell voltages up to 3.45–3.50 V and allowing them to balance there.
If you don’t do this, the cells will continue to drift further apart, but you won’t notice until you discharge the battery really deeply.
My recommendation for LiFePO4 is therefore a maximum of 17S, which keeps you well within the range that the Victron components can charge.
Thanks, Holger and Matthias! My JK BMSs balance the cells across the entire voltage curve when there is a difference of more than 10 mV between the highest and lowest cell voltage. The information about “ageing” is new to me. Unfortunately, I only look into this subject sporadically, as I have so many other things on my plate. But I’m not complaining! I’d be very happy to work with you to find the cause of the “failure to regulate”. There isn’t much you can configure for this in the ESS Assistant — I was just looking at the VR configuration of the single-phase system section, which is interesting in relation to our question. The response of the power reduction is certainly the BMS variant: here I have to [select] “system uses LiFePO4 with other BMS” — and is regulation carried out using SOC or voltage values provided by the MP2??
Matthias, if that were the case, why hasn’t a use for the MP2’s “measurement input” been implemented during firmware development, so that all voltage drops — even those from a SmartShunt setup — would be taken into account?? I don’t remember exactly, but I was really pleased at first that this measurement input is available as hardware.
Are you aware that, with my “BMS selection”, the power reduction should work??
Hello, now that I have a 50 Hz mains supply again, I can report on a peculiarity that may have something to do with the “forced” island operation (where regulating via a higher inverter frequency did not work): At the moment, my three-phase system is still charging, and the yellow bulk-charge LED is lit on all three MP2s. I never saw this state in island mode! Only phase 2 ever indicated the bulk state, didn’t it???
