Fronius Push

Discussion on the Fronius Push

Firstly - great work on the new forums, they’re head and shoulders above the old ones.

Secondly - is anyone else experiencing random oddness with their stats since upgrading to the latest inverter firmware (fro28090.upd)? My average power (https://pvoutput.org/intraday.jsp?id=54404&sid=49466&dt=20170429) seems to randomly dip throughout the day, and the PVOutput service seems to stop counting my power generation at a point during the day which means my total figure is inaccurate.

Thirdly - any chance of getting the Push service updated to accept voltage stats from the inverter? It seems like you currently need the Smart Meter to get this information; however the inverter also generates voltage stats but these don’t appear to be accepted by the push service.

The Fronius push message doesn’t contain voltage information so it isn’t possible to capture it.

The Fronius push message doesn’t contain voltage information so it isn’t possible to capture it.

Ah, bummer. Thanks anyway.

Regarding my second point - I’ve noticed that doing a manual Output Add (even just a single entry) via the Live Loader seems to re-run the average calculation and resolve the issue of the incorrect average values. It still seems to be happening (see https://pvoutput.org/intraday.jsp?id=54404&sid=49466&dt=20170501 and https://pvoutput.org/intraday.jsp?id=57337&sid=51519&dt=20170501 for examples).

The orange line average power a calculated value based on the energy values submitted.

Hi BB (& others)
Just a quick question please.

On WP, the suggested format for the Fronius inverter push service is this:
“Inverter upload file name =
/service/r2/froniuspost.jsp?sid=51***&key=******&v3=None”

From the PVOutput Help file,
“v3=none” is used for “Disable energy upload, i.e. calculate from power”

I’m wondering why the v3=none is included in the WP info? (Lots have followed this suggestion/recommendation)
Especially when I read this, above: “The orange line average power a calculated value based on the energy values submitted.”
I’m easily confused about this stuff.
I’ve not included that (v3=none) in my configuration. I’m assuming that will not harm the data generated? :slight_smile:

I have a similar question based on ‘Bugsy’ settings above.

I have Net settings on my PVOutput - from a Fronius inverter.

Yesterday…

  • Daily consumption shown on PVOutput is lower by about 0.25kWh from actual.
  • Daily imported energy was about 0.6kWh lower than actual.
  • Daily exported energy was about 0.4kWh lower than actual.

That is a ~5% annually lower reported data from the actual data logged by the inverter.

I think my questions are…

  1. Will dropping the v3=none push more accurate energy data and decrease the above ~5% annual errors, AND/OR
  2. Would Daily Net settings (not just Net) push more accurate results for the above 3 elements?
  3. OR is it a combination of all of the above?

Thanks

One suggested setting when using Fronius Net is adjusting the Net Method to Instantaneous under Advanced Settings

Both v3=None and Daily Net won’t affect the import/export calculation which are dependant on the power data reported in the push message from Fronius.

Thank you for your answer - I have spent the last few days reading your posts on Whirlpool to better understand my pushed data sets and what I need to sensibly ask you so I might be able to better control the inaccuracy of my ‘Power Used’ field in PVOutput…

I have used API calls and imported data into spreadsheets and compared both the pushed Fronius inverter and (especially) the meter data and now understand the data sets that are pushed and the incremented forms it sends that data to PVOutput for processing.

Most pushed data is accurately incremented so there is no issue if my inverter is sending accurate data - and it is.

I think the next question I need to ask is therefore how PVOutput handles the pushed data it receives regarding PowerReal_P_Sum - which is a variable in time and not logically totaled and incremented in the push data?

Example…
If I set my push data to once every minute - but have my PVOutput set to a 10 minute Status Interval in ‘Live Settings’ - PVOutput would have been sent 10 sets of data from my inverter before PVOutput publishes its next update.

In this example above, how is PVOutput handling those 10 sets of pushed data in relation to the PowerReal_P_Sum - which is a variable that is sent at its usage level at the exact time of the push - but could have been much higher or lower a second before or after the push?

Is PVOutput (in relation to the PowerReal_P_Sum data)…

  1. Averaging the 10 sets of data?
    …or
  2. Only using the most recent set of data - and ignoring the other 9?
    …or
  3. Something else?

It seems to me because PowerReal_P_Sum is apparently a variable we are on a wicket to nowhere in terms of getting ‘Power Used’ accuracy - and makes sense why I can’t get it under control - but hopefully you have an answer.

Thanks for your help.

Edit 15/5…
I am pushing data using a 1 minute interval for my meter (every 5 minutes for the inverter) and by using Requestbin I can see the pushes are sending reliably - but I am getting a number of 0W ‘Power Used’ results in PVOutput, which seems to suggest the PowerReal_P_Sum data is being dropped by the PVOutput server? That is somewhat weird as the remaining data fields are registering correctly in PVOutput - I have seen this before - seems to happen less if I use 5 minute push intervals - but the 5 minute interval does not improve my ‘Power Used’ accuracy?

The last data push data point will be used. The data from Fronius is used as-is and it is not changed or calculated.

The timestamp provided by Fronius is rounded to the nearest Status Interval, i.e. 5-minutes

0W power used means the Net Meter Value (export) is higher than the Gross generation value at the same timestamp. It is not possible to have export that is greater than gross generation which results in a negative gross consumption value. This value is reset to 0W.

OK thanks for your help - I now understand the parameters I am working to.

I’m also trying to make sense of this. My Energy Used figures are considerably higher on PVOutput than measured by all other sources (e.g. Solarweb and my distributor’s portal). It looks as though PVOutput is using PowerApparent_S_Sum rather that PowerReal_P_Sum as used on Solarweb.

Assuming I’m correct in the above, which of those two figures would represent what I’m really using, and if coming from the grid, what I would be billed for? As far as I can tell, my retailer has only ever used the figure for real power.

If Net data option is enabled with n=1 then PowerReal_P_Sum is used as net and combined with inverter data to calculate gross consumption.

If Net data is not enabled then PowerApparent_S_Sum is used as gross consumption.

Recent corrective action demonstrates we are able to get traction from Fronius regarding programming issues, so I have another question in order that a useful proposal might be put to Fronius regarding the way they might better handle their PowerReal_P_Sum pushed data-sets.

The only data that PVOutput displays exactly as logged by a Fronius data logger is the daily generated kWh - which makes total sense as the pushed data is precisely summed and incremented at the selected push interval.

Fronius pushes PowerReal_P_Sum Watt usage as a variable at the precise point in time of the selected push interval - usage is not summed and incremented usage - therefore PVOutput is supplied imprecise pushed usage data.

As a consequence PVOutput fields ‘Exported’, ‘Consumption’, ‘Import’ and ‘Net’ at best have ~3-4% accuracy from my observation - crap in from Fronius = crap out of PVOutput.

So…if we were to put pressure on Fronius to rectify this PowerReal_P_Sum pushed data issue, what does PVOutput require from Fronius’ regarding the PowerReal_P_Sum pushed data-set to get data accuracy? Thanks

We just need a push value for consumption that matches what is being sent to solarweb.

It can be a new parameter in the push as you can define any parameter to record using v4=<param>

I discovered this thread just now which is pretty much about the same issue as my comments today in the WP thread at https://forums.whirlpool.net.au/forum-replies.cfm?t=2633039#r6.

We will be looking at some enhancements that hopefully should address these push issues. This won’t require any changes on the Fronius side, but will be looking at some volunteer systems to try out the new setting.

1 Like

Happy to assist - let me know

On the meter push configuration replace n=1 with auto=1

This setting relies on the first push of the day to mark as the starting point - if the change over is done midday then the total energy consumption figures may appear lower than expected.

I’m keen to be a guinea pig on this too.

Is all that’s required to become a ‘test system’ to set the auto=1?