Battery charge energy always less than discharge energy

[Phil M]

Morning all,

Over the 4-5 months my system has been installed, I'm pleased to say it's been working well, providing far more energy than we use. I have however noticed some odd numbers that I cannot reconcile. Every day that we've fully charged the battery, the battery always shows as exporting far more energy than it imports, for example, all this week the battery has been at about 80% at midnight, it then drops to about 60% over night then recharges to 100% in the morning, sits there at 100% until about 9pm the drops to 80% again by midnight. Each day however the "battery charge today (solar)" value is roughly 4kWh (makes sense as it's a 10kWh battery charging 40%) but the "battery discharge today (solar)" is around 8kWh. Typically discharge is 1.4 to 2 times greater than battery charge. Assuming it's not a magic battery producing power on it's own, can anybody explain this difference? All numbers are through Home Assistant, if that makes any difference.

Many thanks,

Phil

[WyndStryke]

I wonder if this might be SoC calibration.

How often do you typically cycle through the entire SoC range (i.e., 10% to 100% and back again?)

Does the battery SoC trace look normal? (No vertical jumps or drops?)

What do the 'total' figures look like? (So rather than the charging and discharging from midnight, the total charging and discharging since the system was commissioned?)

[Will]

The good news is that energy can't evaporate; it is being used or was never there to begin with.

It's a known thing that solar only charging can cause the battery to lose calibration so a couple low amp cycles from full to empty to full again should fix it.

The problem is one of the downsides to LFP cells is that the voltage doesn't change much between empty and full so the bms has to count the watts in and out to guess what the SoC is based on it's measurements, voltage and temperature.

I cover this condition in this video

I show you how to adjust the charging amps here:

[Phil M]

The graphs all look normal with no sudden jumps. We typically cycle from 100 to 60% and back but have had the odd cycle from 100-10-100%, but not in the recent good weather. The totals for the system since installation are 479kWh import and 888kWh export, so nearly double out than in. I could understand the battery % being out if it hadn't cycled but I'd assumed the kWh in and out would be fairly accurately measured.

Phil

[WyndStryke]

I think the advice is to try to hit 10% at least once per month.


I always have a slight discrepancy between charging and discharging, which I think is due to a silent 60W trickle charge during the 'feed-in first' mode when PV is active. The 60W doesn't seem to be picked up by the stats. But I don't think that is what you are seeing.

> are 479kWh import and 888kWh export,

Have you added PV into that? i.e., - (export + household load) = (import + PV + losses)

[Will]

The bms runs off the batteries and the charge/discharge isn't lossless so there will be a gap between the charge/discharge of around 5-7% roughly. What are you seeing?

The graphs all look normal with no sudden jumps. We typically cycle from 100 to 60% and back but have had the odd cycle from 100-10-100%, but not in the recent good weather. The totals for the system since installation are 479kWh import and 888kWh export, so nearly double out than in. I could understand the battery % being out if it hadn't cycled but I'd assumed the kWh in and out would be fairly accurately measured.

Phil

[Phil M]

Thanks all for the replies.

A typical full set of stats this month is as follows:-

Grid Import - 0.1kWh
Grid Export - 33.5kWh
Solar Production - 47.8kWh
Bat Import - 4.6kWh
Bat Export - 7.4kWh
"Home" - 17.2kWh

Home seems to be calculated from Home = Grid import + Solar Production + bat export - Grid export - Bat Import
I know that when the battery is very low and charges from the grid from 10-15%, it doesn't register as importing to the battery but does register when it exports that energy, meaning I can explain the battery import < export in winter or low sun but I can't explain that in sunny weather when the battery doesn't deplete that low.

I also know that the typical grid import of 0.1-0.2kWh is more like 0.3-0.5kWh taken from my smart meter, which is believe is due to the CT clamp being unreliable at low current draw.

On a graph, the last 24 hours look like this:-

The midnight to now lines match close enough but you need to ignore the discharge before midnight. Could there be something funny about this perhaps. I also notice that from about 9am till 11am today, the battery was charging but also has a constant discharge of 128W, which seems odd.

Next sunny day I'll do a forced discharge to 10% and see if it changes anything.

Phil

[Will]

Are you using modbus or cloud integration for the data? If the cloud data it's not all that accurate vs what you get from modbus

[Phil M]

All data is through the modbus integration.

Phil