3 phase ESS Multiplus inverters hunting in afternoon

Q&A and troubleshooting
, , , , ,
1

Hi Community,

I have started the process of installing a grid connected ESS system. This is my first dive into the Victron environment, and I have enjoyed all the learning I have done so far. At this point the system has been functioning quite well for months, and I am confident I have all the required equipment to finalise the installation. I am a qualified electrician and not new to renewable technology.

I have following equipment installed.
3 x 48/5000 multiplus-ii in a three-phase configuration (one is the GX device)
1 x 48/5000 multiplus-ii as a redundant supply and single-phase generator input connected to GX via USB with a 300A smart shunt for more visibility.
2 x 280Ah 48V Lipo4 batteries
1 x MPPT 450/200 with 11.4kW of panels connected
1 x VM-3P75CT for grid metering
1 x VM-3P75CT with 6.6kW of Enphase solar connected
1 x EVCS
Lynx distributors and shunt for battery monitoring with significant cable sizing 70mm and 95mm2 as required

95% of loads and the extra Enphase solar are on the output side of the inverters.
All systems appear to work flawlessly when grid supply is disabled.

The issue I am having is the inverters seem to audibly and electrically start hunting in the afternoons, swinging quite significantly from charging to discharging and back again. If I notice the system doing this, I will change the mode switch from on to inverting only and the hunting will immediately stop, waiting an hour or so I can turn the system back to on. Although the screenshots showing the load coming on and off over minutes, this must be an average as actual hunting is happening seconds on seconds off. I have checked all the settings and made sure all devices are running the most up-to-date firmware.

Has anyone experienced this or know where I can start to look for a solution to fix it? I would like to keep the grid connection for redundancy and solar back feed credits.

Thanks,
Mick.

Some screenshots from VRM for reference.

Screenshot_28-4-2025_175616_vrm.victronenergy.com
Screenshot_28-4-2025_175616_vrm.victronenergy.com1631×857 138 KB

Screenshot_28-4-2025_175525_vrm.victronenergy.com
Screenshot_28-4-2025_175525_vrm.victronenergy.com1630×859 143 KB

Screenshot_28-4-2025_175352_vrm.victronenergy.com
Screenshot_28-4-2025_175352_vrm.victronenergy.com1631×859 141 KB

2

Theoretically this can only happen if the feedback of the energymeters is too slow.

Or ofc the control loop is acting too fast.

With standard motor control units you’ll have dozens of parameters and a scope function to tweak the control loop…

3

I did consider communication errors between the GX device and the energy meter. The energy meters are connected via Ethernet to the same network switch as the GX device via pre-made cables. I have swapped and relocated the switch to make sure it wasn’t the issue. I then setup the meters on the VE.Can network to see if there was less traffic and the problem seemed to be persistent.

I haven’t come across any recommendations on which style of communication connection is better for an ESS system. But I believe both LAN and VE.Can to be significantly fast enough to respond to changing loads.

4

That depends how many parameters you are reading. If you read everything the meter can supply you’ll have nice load and latency.

You could check with a pc and wireshark.

Reduce the data to what’s really needed, it might get better.