Integration Solar inverter huawei 2000L

XalaTheShepard · January 23, 2025, 10:14pm

Dear all,

Today my new EMMA got installed together with the car charger. The home battery is installed, but will be activated in June when the solar panels will be installed.
I have installed the Huawei Solar Integration and tried to connect.
Everything is hard wired, but alas when putting in the IP address of the EMMA, I was unable to establish a connection.

I noticed in the issues that there is an issue with HA Core 2025.1.x (pymodbus) . Could this be related or am I missing something?

Many thanks for helping me out.

Kind regards,

Robin

Adamski · January 24, 2025, 5:52am

Have you seen this?

Warning
Users that use the native modbus-integration alongside this integration should not upgrade to Home Assistant 2025.1, as it contains an upgrade to an incompatible version 3.7.4 of the pyModbus library. Significant work has already been done to upgrade this integration to support v3.8.0 of this library. We expect this to release alongside Home Assistant 2025.2, which is expected to also upgrade to that latest pyModbus version. This will solve the compatibility issue.

XalaTheShepard · January 24, 2025, 7:44am

@Adamski , I saw the issue you mentioned. It was the same I was refering to. So most likely, I will have to wait untill HA Core 2025.2.X . But now I know that it is not an error in my setup

XalaTheShepard · January 31, 2025, 3:18pm

I saw there was an update today of the Huawei Solar integration which should automatically detect the setup. However, after installing the update, my Huawei inverter and EMMA still could not be found nor added. Is this still related to the Modbus issue in HA Core 2025 .1.X? Many thanks for the response and support. Kind regards, Robin

fabio.deagostini · February 1, 2025, 8:25am

Hi John, I’m new in Ha… so please be patience
Please, I would like to implement the changing of TOU ( to charge during the night the batteries), all from HA.
I’ve seen your post with che sensor SET_TOU_PERIODS…
In my ha with fusionsolar and Huawei integration, it thereìs not.
Where Can I work to add it?
Thks
Fabio

Johan71 · February 6, 2025, 6:35am

Has anyone dared to update to 2025.2 and can confirm whether it works or not?

Warning
Users that use the native modbus-integration alongside this integration should not upgrade to Home Assistant 2025.1, as it contains an upgrade to an incompatible version 3.7.4 of the pyModbus library. Significant work has already been done to upgrade this integration to support v3.8.0 of this library. We expect this to release alongside Home Assistant 2025.2, which is expected to also upgrade to that latest pyModbus version. This will solve the compatibility issue.

Piter123 · February 6, 2025, 6:47am

I just installed core 2025.2 and it seems to be working properly.

MariuszWoszczynski · February 6, 2025, 3:24pm

Did you solve your problem of the result of the calculation giving negative solar energy at night?

template:
  - sensor:
    - name: "energy_pv_daily"
      unique_id: "PV Energy daily"
      state: >
        {{ (states('sensor.inverter_daily_yield') | float) - (states('sensor.battery_1_rozladowanie_dzienne') | float) + (states('sensor.battery_1_ladowania_dzienne') | float) }}
      device_class: energy
      state_class: total_increasing
      unit_of_measurement: 'kWh'

nickegbg123 · February 10, 2025, 1:47am

Do you know how to set TOU after high and low electricity prices?

mmbogdan · February 10, 2025, 7:57am

I’m on core 2.1 and I have a issue with the integration, it worked on 2.0 just fine

flrnwrzl · February 10, 2025, 1:25pm

hello could you please share purchased parts. is there also any power supply needed? and where is the SDongle?

colyro · February 10, 2025, 1:41pm

hi

any feedback about how the efficiency conversions is working?

i have a 5ktl-m1 and a voltage of ~250v/string (+10kw battery) and i cant make it work due to chaotic behaviour of the efficiency of inverter i assume. i think its related to the fact that i have battery and i cannot find the effciency point

how should look like the config based on the enclosed efficiency inverter graph?

…

Argogma · February 10, 2025, 4:09pm

I dont know since when but I think since 2024.12 my Luna is missing some sensors:
select.battery_capacity_control_mode
sensor.battery_fixed_charging_periods
sensor.battery_time_of_use_periods

everything is okay.
At night when the luna2000 is in sleep-mode there are also
sensor.battery_bus_voltage and sensor.battery_bus_current missing.

Is there something i am missing or does someone else has the same issue?

Piter123 · February 10, 2025, 7:34pm

I have everything updated and I don’t see any problems. Core2.1

cailloup · February 18, 2025, 9:11pm

A way to solve this is to recompute solar power from current and voltage
And then integrate the value to have kWh, for example with this add-on:

In configuration.yaml:

template:
  - sensor:
      - name: "Solar power"
        unit_of_measurement: "W"
        device_class: power
        state: >
              {{ ((states.sensor.inverter_pv_1_current.state | float) *
                (states.sensor.inverter_pv_1_voltage.state | float) +
                (states.sensor.inverter_pv_2_current.state | float) * 
                (states.sensor.inverter_pv_2_voltage.state | float) ) | round(0) }}
  - sensor:
      - name: "House power"
        unit_of_measurement: "W"
        device_class: power
        state: >
              {{ ((states.sensor.inverter_active_power.state | float) -
                (states.sensor.power_meter_active_power.state | float)) | round(0) }}
  - sensor:
      - name: "Solar daily energy"
        unit_of_measurement: "kWh"
        device_class: energy
        state_class: total_increasing
        state: >
              {{ ((states.sensor.inverter_daily_yield.state | float) +
                (states.sensor.battery_day_charge.state | float) -
                (states.sensor.battery_day_discharge.state | float)) | round(2) }}

solar power integral compared to calculated energy from “solar daily energy” in the yaml
Integral is a bit higher since it’s before the inverter efficiency

No negative value, no dirty graph when battery charge from grid, no midnight spike when reset is at 00:01

colyro · February 20, 2025, 7:26am

having the inverter input power with the integral behind, gives me quite a deviation - up to 2-3kw/day which multiplied by 365 means a lot.

i am struggling to fine tune the efficiency for input power based on sunny and cloudy conditions to reach as much s possble the lowest deviation.

the formulate with charge/discharge battery i think it doesnt work for me as I have 2 inverters and the one with the battery has On-Grid charging turned on, in order to charge from the other one surplus…

jav · February 26, 2025, 2:17pm

Dear Joerg, I also have a WLAN-FE dongle and two SUN2000-5KTL inverters. I’m trying your setup: I have the dongle connected to my LAN via cable. I also configured the inverters so that they connect to my LAN via Wi-Fi and enabled the O&M checkbox. After I did this, all 3 devices (the two inverters + the dongle) connected via Wi-Fi and were given DHCP addresses. I could ping all 3 devices and pull data directly from the inverters via Modbus TCP on port 6607. Additionally, the dongle is still pingable to its wired IP.

Everything GOOD so far!!

However, if I reboot the inverters (eg. by upgrading them), they fall back to just wired, and neither the inverters or the dongle will try to connect back to the Wi-Fi.

I thought maybe disconnecting the ethernet cable would force everything to fall back to Wi-Fi again, but this only appears to be the case for the Dongle, which is visible on its WiFi IP, but not the inverters.

So… Did you experience this issue or did anything special to have the inverters stick to Wi-Fi?

Thanks!

Roving-Ronin · February 27, 2025, 12:18am

@jav you don’t say if they are L0 / L1 or M1 inverters.

If you have the USB sDongle (that can do either WiFi or Ethernet connection) then you should be sticking that into the Primary inverter and using that to connect both to the network.

Primary inverter is connected to the Secondary inverter (and consumption meter) via RS485 wiring, so you end up with 1 IP for the sDongle that the communicates to the Primary inverter, that in turn controls and obtains data from the Secondary inverter.

In HA you then just install the WLCRS integration and enter the IP for the sDongle (then option installer user/password, that is required if you want to control actions of the inverters/batteries). With v1.5.x of the intergration you should not need to enter ID numbers for the meter / inverters, it should auto detect.

jav · February 27, 2025, 7:08pm

Hi @jbhobson, thanks for your answer! Inverters are SUN2000-5KTL-L1, sorry for omitting that detail. Also I should have clarified that I’m not interested in getting data from the SDongle because that’s both slow and unreliable as noted by @manio here, @heinemannj here and many others. The main issue appears to be that the FusionSolar backend outages cause the SDongle to reboot itself and stop responding to requests via Modbus TCP. So Instead, I’d like to pull data directly from the inverters.

What I’ve found so far is that you need to meet 3 conditions for this to work:

Connect the SDongle via both cable and Wi-Fi
Wi-FI needs to be set-up in DHCP mode
O&M must be checked.

If you don’t use a cable, only the SDongle will attempt to connect to the Wi-Fi, but not the inverters.

If you use a fixed IP address instead of DHCP, this setting appears to get synchronized and both inverters end up with the same IP.

And if you don’t check O&M, the inverter refuses connections on port 6607.

So you need to meet all of the above (wired SDongle + DHCP + O&M). But I’ve found this set-up quite fragile because it breaks if either the inverters or the router reboots. I might have to use an intermediate OpenWrt or similar hardware that connects to the inverters as clients and have their wlans routed over to my wired LAN, similar to what @andrix described here but I’d rather avoid the extra hardware.

colyro · March 1, 2025, 4:20pm

hi

I was wondering if the efficiency loss could be such worst in terms of ~50%.
I’ve created a riemann left integral on sensor.inverter_input (the raw DC energy produced by panels without any conversion) and I got produced for today 2.17kWh (DC) while according to fusionsolar portal i got only 0.64kWh (AC).

Is this conversion loss doable as the graphs of the inverter (SUN2000-5KTL-M1) shows a minimum of 93% for strings with <470 Volts?
Voltage of the string installed on this inverter was at ~250V.