3-Phase MultiPlus-II ESS instability

DIY
, , , , ,
1

Hi All,
I am experiencing severe system instability with the 3-phase ESS setup.

Setup:

  • 3x MultiPlus-II 48/5000/70-50
  • Cerbo GX
  • Pylontech US5000
  • Victron VM-3P75CT (Grid meter)
  • 2x 1-phase 5 kW PV inverters + 3-phase PV meter

Ever since a recent grid outage and utility intervention by Fluvius, the system goes wild whenever the MultiPlus units are active with ESS enabled:
-Wild power swings: The system enters a heavy feedback loop. MultiPlus units rapidly swing between max charge (+130 A) and max discharge (-130 A) every few seconds.
-Incorrect meter readings: Grid readings bounce wildly between +10 kW and -10 kW. In Venus OS, AC Loads drops to 0 W because it detects a massive fake grid export (-9.7 kW) while the house is actually drawing power.

So far, the main grid meter was replaced with a new Victron VM-3P75CT. With the MultiPlus units off, all three phases read correctly with correct signs. When the MultiPlus units are turned OFF, all grid, PV, and house consumption readings immediately stabilize and show 100% accurate, real-time values.
Since the PV setup consists of two single-phase units on a 3-phase meter (producing energy only on L2 and L3), I suspect a phase mapping / rotation mismatch between the grid meter, PV meter, and MultiPlus AC-In terminals.

Questions:

  1. Could a phase mapping mismatch on an external PV meter cause the ESS algorithm to trigger feedback loops of this scale (±10 kW)?
  2. What specific parameters or diagnostics in Cerbo GX or VEConfigure should be inspected to locate the exact cause of this control loop oscillation?
  3. Could an internal phase-angle or voltage sampling calibration on AC-In have been corrupted by the grid surge?

Thanks for any help or advice!

2

I would start with checking all wiring and phase sequence, also check if the current clamps are placed right and tight (put a tyrap on them)

I see a lot of problems with those ct clamps, slightly loose and measuring will be incorrect

do you know what the grid operator changed?

3

Hi @DuivertNL
Thanks a lot for the suggestions.

Looks like the problem is solved. Everything works stably for two hours now and the wild power spikes are completely gone.

To give you the full context: we replaced the grid meter, double-checked all wiring, and completely reconfigured the Multiplusses, but none of that fixed it at first.
Then we realized the local network connection was extremly poor. It turned out the network switch was damaged (probably by the voltage spikes). Once we replaced the switch, the connection between the grid meter and Cerbo stabilized instantly. Since then, no power jumps at all and the grid is holding the setpoint.

As for the grid operator, they said they didn’t change any settings on their end.

Thanks again for your help.

4

Nice to hear its solved!