[BUG] Net Energy Production Increasing Despite No Grid Export

[federicotravaini]

Describe the bug
After several weeks of testing, I believe there is an issue with the "Lifetime Net Energy Production" value. As I understand it, Net Energy Production refers to the amount of energy sent back to the grid. In my case, this should be close to zero, since I'm not yet allowed to export energy to the grid (yay Italy, such a great country).
However, this value keeps increasing, as if I'm sending a lot of energy back (10/12 kWh a day).
Could there be a mistake in how this entity is calculated?

Account type

• Installer
• DIY / DHZ
• Home Owner (This integration is not (fully) functional with a home owner account!)

Envoy

• Metered with CTs installed

• Metered without CTs

• Standard

• FW version: D8.2.4382

• Amount of micro inverters connected: x

Home Assistant

• Version: 2025.4.1

[vincentwolsink]

Hi @federicotravaini, this integration does not calculate any values, it just fetches from the Envoy and presents it as a sensor.

Just to double check, if you could post the output of https://envoy.local/ivp/meters and https://envoy.local/ivp/meters/readings here (don't forget to login to the Envoy first).

[federicotravaini]

Here you go but I cannot find "net production" data

readings.json

meters.json

[vincentwolsink]

Net production is "actEnergyRcvd": 849257.110 on line 76 in the supplied readings.json.

[federicotravaini]

I’m analyzing the energy data from a three-phase Enphase system and noticed that the exported/imported energy values reported in the JSON files (especially in actEnergyDlvd and actEnergyRcvd) are significantly different from those shown in the Enphase app.

After some investigation, I believe this discrepancy is due to how three-phase systems often work: to minimize net energy exchange with the grid, the inverter balances the phases. As a result, what is recorded as “exported energy” in the JSON files can sometimes be internal phase balancing that doesn’t actually represent net export to the grid.

In other words, the JSON data includes as “exported energy” some of the power that is simply used to cancel out consumption on other phases, not true export. This makes the data misleading or at least hard to interpret without knowing how exactly Enphase aggregates or filters these readings.

I’ve tried to reverse-engineer this behavior by summing per-phase data and comparing it to aggregated values, but with the current structure of the JSON data, it seems not possible to accurately calculate true import/export values as seen in the app.

I’m attaching:
• the raw JSON files retrieved at that moment (from the “readings” and “meters” endpoints),
• and a screenshot of the lifetime status of my system taken at the exact same moment

readings.json
meters.json

[catsmanac]

In 3 phase installations with batteries, the battery will export to the grid on one phase the amount it lacks on another phase. This other phase pulls the missing amount from the grid, as if it is using the grid as a ‘transport’ between phases. The Envoy 3 phase net-consumption CTs will report the amounts on both phases, resulting in too high export on one and too high import on the other.

[federicotravaini]

Thanks for the clarification. However, the main issue is that the Enphase app correctly shows the real energy exported/imported, excluding flows due to phase balancing.

The raw JSON includes these internal phase flows, which leads to overestimated import/export values.

Is there a way to access the same corrected data shown in the app, which excludes artificial transfers between phases?

[catsmanac]

The Envoy provides in /production the net-consumption segment. In there, the changes of whLifetime up/down may provide you with consumption/production values. You'll have use some calculations, sum negative changes to one entity and positive changes to another one. That may give a corrected import/export,

[vincentwolsink]

@federicotravaini There might be a way to do calculations in the integration, but I don't see a way to get to the numbers in the screenshot using the numbers from the json data. Also, Enphase stores historical data in their platform. The counters on the Envoy itself seem to reset in some rare occasions. So they might not match at all.