Single-phase dual inverters on a 2-phase system

[outofmyshed]

[*]I've recently become involved with my local village hall. About 2 years ago they took advantage of a grant and had a hybrid system installed but haven't had much of a clue about how to operate it so, as I've got experience with my Fox system at home, I've volunteered to help. However, the way the system is operating doesn't make sense to me, but I don't understand enough about how 2 phases would interact - my assumption was that they would be completely separate from each other from an inverter POV.

The setup seems quite unusual. There is a 3 phase grid supply and meter, from which two phases are being used to supply two separate single-phase distribution boards. The various circuits (lighting, heating, sockets) are distributed across each phase. Each phase has its own H1-3.7-E (Gen1) with an 17.2kWh ECS2900 battery stack and 2.8kWp of panels on each connected to one string. CT clamps look correctly placed (but could be wired with incorrect polarity, as they have been extended with Cat5e - haven’t physically checked that yet), and CT2s are not used. There is no other useful energy monitoring going on.

Cut-out, meter and CTs:

The problem I've observed is that on both inverters, when the batteries are full (as they nearly always are now) LoadPower is always negative, Feed-In Power looks too high (each inverter only has 2.8kWp of panels) and is the inverse of MeterPower, and OutputPower (and PVPower, which has the same value) always = Feed-in Power + LoadPower. Based on the Feed-in stats, each inverter is exporting a lot more power than it ought to be.

Inverter 1:

Inverter2:

Intuitively this looks wrong to me, based on the experience with my own single-phase H1. Does it look like the exported power across the phases is being summed together in some way, as the numbers look too high for the string sizes? Why is LoadPower on each phase negative - my understanding is that it should not be. Does this all point to the CT clamp polarity being incorrect, and/or some interaction between the phases?

[outofmyshed]

An update: I checked the CT clamps and restarted both inverters. CT clamps were correct, and both inverters work as expected when reacting to load on each separate phase - when the batteries attached to each are under Max SoC (100%).

The issue seems to be when the batteries reach 100% and each H1 begins to export, the Load and Feed-in Power metrics look incorrect, which is skewing the overall portal data. I've screenshotted the v2 web portal charts as it makes it clearer:

Inverter 1:

Inverter 2:

On each, it's at the point that the battery reaches 100% that Load Power goes negative (and is therefore within the 'Supply' range) and the Feed-in Power increases, and it looks like each inverter is exporting considerably more power to the grid than excess PV than it has available. Clearly this can't be the case, so it looks like a possible measurement/cloud reporting problem?

The H1 firmware versions on each are (I believe, quite old):
Master : 1.48
Slave : 1.02
Manager : 1.45

[Dave Foster]

Sorry for the slow response, there is nothing wrong with the setup as you have described, as long as there is proper separation on each phase the H1 inverters will manage their own phase in isolation from the other phases.

The CT clamp for each inverter must be fitted to the phase the inverter is on and there must be no physical connection between phases - normally on a single phase system the inverter CT clamp would be fitted on the live feed of the smart meter but in a 3 phase system it should be fitted to the L1(2/3) where it is split to feed each phase.

I've looked at the latter 2 traces for inverter 1 / 2 but they are exactly the same as if its the same device, certainly the trace suggests that export is being incorrectly measured as load which as i've said implies CT location (they appear to be in the correct places so have the CT's been swapped between inverters ?, and also check that the CT's are enabled in 'Setting, Feature, Meter/CT' since this was a known problem in earlier firmware)

You should be able to see all of the stats locally at the inverters in the menu but i've never known the cloud to confuse information like that.

The fact that you are seeing negative load when the battery is full suggests that export is being incorrectly counted as load which can only really happen if the CT clamps are fitted in the wrong location or each inverter is seeing the others output i.e. this is definitely a metering problem.

Also just to double check you shouldn’t have a second CT clamp fitted which is common on systems that have multiple inverters, on a split phase system the inverters mustn’t be able to see each other and so only a single CT clamp should be fitted where the phase is split.

The firmware versions are very old but still reliable, so you should be fine running them.