Energy does not reset each day

[markruys]

Energy does not reset each day

[SmallKingfisher]

I had the same issue, but it got fixed in version 1.2.2.

[JWDebler]

sometimes it works, sometimes it doesn't

[TheHawkAU]

I've noticed this behaviour too, although in my case it's half of yesterdays total. I am assuming it's the average of the last reading from yesterday + 0 from the first reading of the day.

Which means my graph looks like this with a high total until the new total for today for the day exceeds that number.

Do you have to provide that running total as part of the upload or will PVOutput calculate it for you?

I've also noticed that I'm getting a report all the way to midnight, but not through the morning. If it was the streetlight I would of thought it would be consistent until sunrise? I do wonder if this is a false number being the last number reported at the end of the day incorrectly reading over and over.

[eNeRGy164]

As far as I remember, running total is calculated by PVOutput.

Do you use the --skip-offline switch? The GoodWe API keeps reporting the last measurements, even when the inverter is offline.

[TheHawkAU]

I hadn't before, but I have now, just rebooted to ensure everything is applied and we will see how things go over the next couple of days. For the moment it appears to have stopped logging.

[TheHawkAU]

Update:

26/08 - the last reading of the day was repeated until midnight, then the following day started at half the previous days energy total. Once this was exceeded things continued normally.

27/08 - Again started with hald the previous days energy total. That evening I added in the --skip-offline command and rebooted the machine and it stopped showing records into the evening once the inverter went offline.

28/08 - Looks like success! The first energy reading for the day is 0.1 and it goes up over the course of the day again stops reporting once the inverter goes offline in the evening.

29/08 (Today) - Back to reporting half of the previous days reading at the energy total).
https://pvoutput.org/intraday.jsp?id=47592&sid=61960&dt=20190829

So the switch stopped all the late-night entries but doesn't appear to have resolved the underlying energy issue (was it just the reboot that did that?). Will see again tomorrow.

I wonder if this is still something to do with averaging between the last entry and the first for the day? Maybe with skipping offline there isn't then a reset to 0 at midnight? Will have to look into the code a little more when I have some spare time.

edit:
I've just checked the --csv output files are they do show "Energy today (kWh)" and it does indeed start from zero each day (although side note; the csv output doesn't include any time (or date) stamps which would be a nice addition).

Following the code through it does look like it's passing the correct values, evidenced by the CSV output, it also looks like it is passing the daily kWh total with each update (and it's not calculated by PVO) through {eday_kwh}.

Which leads me to wonder if it's a PVOutput issue after all, although why it's always exactly half of the previous days total is odd since that doesn't appear to be a calculated field??

update 08/09/2019:
I thought I should update this. I went through it all again and realised that I was following the instructions a little too closely and had re-installed 1.2.1 and not 1.2.2. Got the latest release and saw all the problems go away and it appears to be working as it should now.

[markruys]

I've added --skip-offline to the Systemd service example in the README.md.