"Hello community,
I'm experiencing a strange behavior with my inverter. On sunny days with high PV output, the PV power drops significantly when there's grid consumption. You can see this clearly in the screenshot. The yellow line represents consumption and the black line represents PV power. Is this just happening to me?"
Master: 2.11
Slave 1.03
Version_Arm 1.87
Greetings
H3-10.0-E Strange behaviour on load
-
Dave Foster
- Posts: 1869
- Joined: Thu Oct 13, 2022 7:21 pm
It's not unusual to see drops in peak output with clouds or shading effect - have you ruled that out?
It is possible your inverter is tripping when exporting because the grid voltage is going too high - do you see any alarms or error messages for that period ? - If you look at Foxesscloud in the Inverter Details, in the empty graph at the bottom you can select RVolt, SVolt, TVolt sensors and that will show you the voltage for the day - and highlight the peaks - does that look to be high when the PV output falls ?
It is possible your inverter is tripping when exporting because the grid voltage is going too high - do you see any alarms or error messages for that period ? - If you look at Foxesscloud in the Inverter Details, in the empty graph at the bottom you can select RVolt, SVolt, TVolt sensors and that will show you the voltage for the day - and highlight the peaks - does that look to be high when the PV output falls ?
"Thank you for your quick response.
In this case, I can rule out clouds.
I can intentionally trigger this behavior by turning on larger consumers (kettle, oven, ...).
The R/S/T voltages also look very constant on this day. As I mentioned, the behavior is reproducible and was also present in older firmware versions. What additionally puzzles me is that there was more than enough PV power available, and I fed a lot into the grid."
In this case, I can rule out clouds.
I can intentionally trigger this behavior by turning on larger consumers (kettle, oven, ...).
The R/S/T voltages also look very constant on this day. As I mentioned, the behavior is reproducible and was also present in older firmware versions. What additionally puzzles me is that there was more than enough PV power available, and I fed a lot into the grid."
- Attachments
-
-
Dave Foster
- Posts: 1869
- Joined: Thu Oct 13, 2022 7:21 pm
Is it possible this is a 3 phase balance problem?,
The inverter when in balanced mode will be providing it's PV output equally across all 3 phases - so for example with 9kwp of PV each phase will be supplying 3kw (to a maximum of 3.33kw per phase as a 10kw inverter) - however if a large load is turned on in the house the inverter will supply the first 3kw from PV and then either use batteries or grid depending on how the inverter is configured.
This seems a possible hypothesis - as from a work perspective the inverter was doing 8.2kw of PV, the load appears to be over 6.2kw - more than a single phase is capable of providing even with the battery and so you would draw some power from the grid.
The inverter is doing (8.2 + 6.2 - {3.33 from PV/Battery} - 2.87kw from grid}) = 8.79kw so within its 10kw output but I wonder if the large load is being met from the battery initially and as the PV is 8.2kwp with only 2 phases the maximum output is (3.33*2) = 6.66kw - so the PV would have to derate temporarily until the inverter had time to balance the load (note this would appear worse if the load was switching such as an induction hob).
Don't forget that the graphs you are looking at are based on 5 minute samples, whereas the PV derating would only occur when the high load is presented but it would then appear on the graphs for a minimum of 5 minutes - looking at the inverter screen may provide more real time data.
The inverter when in balanced mode will be providing it's PV output equally across all 3 phases - so for example with 9kwp of PV each phase will be supplying 3kw (to a maximum of 3.33kw per phase as a 10kw inverter) - however if a large load is turned on in the house the inverter will supply the first 3kw from PV and then either use batteries or grid depending on how the inverter is configured.
This seems a possible hypothesis - as from a work perspective the inverter was doing 8.2kw of PV, the load appears to be over 6.2kw - more than a single phase is capable of providing even with the battery and so you would draw some power from the grid.
The inverter is doing (8.2 + 6.2 - {3.33 from PV/Battery} - 2.87kw from grid}) = 8.79kw so within its 10kw output but I wonder if the large load is being met from the battery initially and as the PV is 8.2kwp with only 2 phases the maximum output is (3.33*2) = 6.66kw - so the PV would have to derate temporarily until the inverter had time to balance the load (note this would appear worse if the load was switching such as an induction hob).
Don't forget that the graphs you are looking at are based on 5 minute samples, whereas the PV derating would only occur when the high load is presented but it would then appear on the graphs for a minimum of 5 minutes - looking at the inverter screen may provide more real time data.
Hello Dave,
thank you for your detailed thoughts.
Actually, it could be some kind of "overload" in this case. However, the drop in PV power occurs even with smaller loads (~2kW). I've recorded a short video to demonstrate the behavior.
Here's the link:
I've been observing this behavior for a while now and wanted to find out if it's just happening to me.
thank you for your detailed thoughts.
Actually, it could be some kind of "overload" in this case. However, the drop in PV power occurs even with smaller loads (~2kW). I've recorded a short video to demonstrate the behavior.
Here's the link:
I've been observing this behavior for a while now and wanted to find out if it's just happening to me.
-
Dave Foster
- Posts: 1869
- Joined: Thu Oct 13, 2022 7:21 pm
Good to see you have home assistant the faster polling interval is much better to see the changes.
It does look like a blending/load problem with the demand causing the battery to be used which throttles back PV on that phase and so the main MPPT's derate to cope - what looks like a slight recovery once it has had time to balance ?
You're on the latest firmware so nothing to do there...
It does look like a blending/load problem with the demand causing the battery to be used which throttles back PV on that phase and so the main MPPT's derate to cope - what looks like a slight recovery once it has had time to balance ?
You're on the latest firmware so nothing to do there...
Today, I ran the kettle for two minutes (~2kW).
For exactly that time, the PV power from the PV2 string dropped.
I find it strange, as the inverter should have more than enough power.
However, I can now imagine what's happening. The inverter sees the load on one phase and wants to balance it in the grid, but due to the 3.5kW limit per phase, it can't distribute properly.
For example, PV power is 8kW
R - 0W
S - 2000W
T - 0W
Now it wants to distribute the 8kW and balance the load.
R - 2000W
S - 2000W + 2000W
T - 2000W
But since it can only do 3.5kW on phase S, it will have to throttle back... That's my idea.
Is there maybe a setting that allows asymmetric feed-in?"
For exactly that time, the PV power from the PV2 string dropped.
I find it strange, as the inverter should have more than enough power.
However, I can now imagine what's happening. The inverter sees the load on one phase and wants to balance it in the grid, but due to the 3.5kW limit per phase, it can't distribute properly.
For example, PV power is 8kW
R - 0W
S - 2000W
T - 0W
Now it wants to distribute the 8kW and balance the load.
R - 2000W
S - 2000W + 2000W
T - 2000W
But since it can only do 3.5kW on phase S, it will have to throttle back... That's my idea.
Is there maybe a setting that allows asymmetric feed-in?"
- Attachments
-