I have a problem in one of my installations. It is a 3-phase system with 3 Multiplus II and Cerbo GX. There is ESS running. The batteries are 3 Pylontech US3000, connected via a Lynx Distributor.
After running 1 1/2 years without problems, it now happened several times that the 220V output shut down for only a few seconds. The customer can not tell, if it was just one phase or all 3, because it was so short. In have different errors in the VRM alarm-protocol and on the GX. The GX device shows most times VE.Bus error#1 and just one time VE.Bus error#3. On the VRM alarm-protocol only the error#3 appears, but always 3 times with just 1 or 2 seconds time difference. The error#1 on the GX device is about one hour later.
I have no idea how I can localise the problem. It is very rare, the first time in August 24 and now 3 times in December. And the shut down is extremly short. I checked the DC fuses, but they are OK and I also checked the communcation connections. Thats all short original Victron cables.
I made some more investigations and found out, that all 3 phases shut down at the same time. At that time also the VE.Bus went to off and the DC current was quite high. I am not sure what causes what and why. Today it happened 3 times and yesterday one time. Does anybody has an idea?
What does the battery DCL look like at the point of shut down.
Something is send the signal to shut off.
My first suspect will be the battery bank shutting the system down
A sudden high discharge suggests wiring or load or switchgear problem. Possibly in the load side (logging will be missing it).
Yes, the DC has a sudden high discharge at that point. I do not believe that wiring is causing the problem, because the system run for more then a year without any problems and no changes. I attach a picture of the system. AC-Load could be a problem, but you can not see a specific situation at that point. Only two times the phase 2 had a higher peak at that point, but still not very high. What do you mean with “switchgear problem”?
I also attach a new screenshot with battery current and voltage included.
DCL and CVL are both completely stable over the whole time. Only CCL went sometimes down from 111 A to 103 A for about an hour, but not at the time of the event.
I will check, if there is switch to feedback to the grid. I did not do the electric part, that was done by the electrician, who made the house installation. We feed into the grid and I thought the smartmeter would handle that. What kind of switch could be there and how could that trigger the discharge?
I was trying to work out where the discharge was going to since it isn’t your loads on the output. But grid is also a load. ESS sometimes ‘tests’ the grid to see if it can frequency shift. It can cause a reverse power flow under some circumstances.
It looks a bit like the battery voltage (or possibly a cell voltage) drops below minimum. I obviously don’t have the same resolution available as you would in the VRM. So this is a guess.
With your battery bank wired as it is balance may also be an issue. (So a weak bank.)
Yes, I am learning much more now about diagnose. It is the first time that one of my 8 installations has a more complex problem.
The battery was not charged to 100% since end of november. We do not have much sun at the moment. But the problem appeared for the first in August, when the battery was fully charged every day.
Ha ha. It is the one that will make you an expert we call that one the ‘school fees’ install here.
It may be worth charging them. What is the deviation between high and low cells?
What grid code do you use? have you programmed LOM detection?
You’re running 15kVA of system off 6kVA of battery, ones that are known not to age well when undersized.
If I had a system like this behaving unreliably, that’s where I’d start.
The pack is 10.5kWh, which is capacity and has little to do with performance.
They are limited to 37A per battery, that is the DCL that is set, so 111A. That is the normal continuous operation range.
So at 52V = 5772VA.
Relying on “peak” performance for that system is not a great idea. Low frequency inverters need to “lean” on the battery for transients, which you can often get away with when everything works fine, but when grid has an issue, the pack takes a beating.
The minimum numbers of batteries required is reasonably well documented on the victron site, though uses the lesser powered US2000’s, but you can do the maths from there.
Oh, thank you for the explaination. That is very helpfull.
The system was never at this limit. The last 30 days the maximum load was 1500W. In last summer there was a maximum of 5000 W for e few times, when a waterpump run. At that time the error appeared for the first time and then again only in December. Now the system is running fine for two weeks.
That makes it so difficult to find the reason. The best reason for me at the moment is a weak grid. We are in the country side and this system is at a farmer who is at the edge of the village.
I think that this coupled with the banks inability to deal with inrush because of its DCL ability may have something to do with it.
(You can see it in the voltage dive when current is drawn.) Though i don’t think its the whole picture because log distance is 60seconds and alot can happen in that time.
The system seems to self recover so to me this seems the most logical reason.
Maybe LOM detection and weak ac enabled will help and closing up grid voltage tolerance wil help so it rejects earlier.
Ultimately though I think you need a more robust DC bank. The batteries will have started to age now and the issue will only get worse.
we call that one the ‘school fees’ install here.