2021.7 Forecast.solar: what do these Power sensors indicate

[Mariusthvdb (Marius)]

after a day or 2 with this new integration, I am still a bit puzzled with the Estimated power sensors. They should indicate the actual usage of that moment, Power production sensors indicate a momentary ‘current’,.
As such the unit is ok as is the unit of the ‘aggregate’ Energy sensors in kWh.

However, when this is my momentary state:

what does it mean that the next 24 hours my estimated power is 1,530 W? This seems to be impossible prediction. Power is momentary, and not for the next hour, or day?
Even more surprisingly is that the estimated power for the next hour is more than that of the whole next day, being 2,697 W…

of course, the next 12 Hours sensor simply is wrong, with the Peak time to be in 2 hours

EDIT:

just found this: component/forecast_solar: change the naming of the power production values by RubenKelevra · Pull Request #52789 · home-assistant/core · GitHub

they’re point in time sensors (as I wrote above momentary) and indicate the ‘moment’ at the next 1/12/24 hours…

fixing the issue of the next 12 hours…

[klaasnicolaas (Klaas Schoute)]

We may have to do something about the entity names, it’s just difficult to come up with a correct name that is immediately clear to everyone For example, we now have total energy production today and tomorrow, but what should the entity name be if we do this 7 days ahead.

Perhaps extra clarity in the docs will suffice. But the power values are always moments in time and the peak time then indicates the moment when this value will be highest. energy production this hour and next hour are calculations.

[Mariusthvdb (Marius)]

hi Klaas!

I believe the Pr suggestion to use ‘in’ fits the bill perfectly, immediately identifying this to be point in time sensors.

have a look, seems much more to the point:

[klaasnicolaas (Klaas Schoute)]

Now I see this, yes this is much better!

With regard to estimated energy production per day, how could we best call it after tomorrow?

[Mariusthvdb (Marius)]

you mean you want that for every day in the following week? Will this be cumulative in that case, as in ‘from now till the day after to morrow’ or, will they indicate the estimated energy production of that exact day?

if the latter you could do this:

[klaasnicolaas (Klaas Schoute)]

It will show the estimated total energy production that day.

[BruceH5200 (Bruce Hartley)]

I’ve installed the latest beta. 2021.08b5.
The esimtaed power production now is showing as 0.223W when it should be 223W or 0.223kW
It seems to have been ok on 2021.07 ?

Is anyone else seeing this ?

[Mariusthvdb (Marius)]

Well I am confused too.

If the kWh data are correct using the . To indicate the decimal then the W are off. Or the other way around if course …

[BruceH5200 (Bruce Hartley)]

I’ve created a pull request to fix.

ForecastSolar - power production now w not k w

[Mariusthvdb (Marius)]

yeah, I guess we should either see 3 077 W (3 thousand seventy seven) Watt or (what has changed in the underlying) 3.077 kW.
personally I dont like the kW because it is easily mistaken for other units in the electricity domain. Id vote for bringing back the full number and keep using W.

[BruceH5200 (Bruce Hartley)]

yes, that’s what I’ve done. I’ve removed the /1000 so it reports Watts correctly.

[klaasnicolaas (Klaas Schoute)]

The PR of Bruce has been merged, so that problem is solved

[BruceH5200 (Bruce Hartley)]

I’m confused by my data, Previously I was just using the daily forecasts which seemed approx ok. ( I have 4 arrays in different directions, so I am sum 4 integrations).

I’ve started looking at the raw data from the API to try and make sense of it and even that is confusing.

for one direction I get…

"watt_hours": {
            "2021-08-04 05:34:00": 0,
            "2021-08-04 05:40:00": 0,
            "2021-08-04 05:45:00": 2,
            "2021-08-04 06:00:00": 8,
            "2021-08-04 06:30:00": 38,
            "2021-08-04 07:00:00": 90,
            "2021-08-04 07:30:00": 168,
            "2021-08-04 08:00:00": 277,
            "2021-08-04 08:30:00": 416,
            "2021-08-04 09:00:00": 568,
            "2021-08-04 09:30:00": 746,
            "2021-08-04 10:00:00": 939,
            "2021-08-04 10:30:00": 1149,
            "2021-08-04 11:00:00": 1374,
            "2021-08-04 11:30:00": 1603,
            "2021-08-04 12:00:00": 1802,
            "2021-08-04 12:30:00": 2011,
            "2021-08-04 13:00:00": 2227,
            "2021-08-04 13:30:00": 2445,
            "2021-08-04 14:00:00": 2656,
            "2021-08-04 14:30:00": 2856,
            "2021-08-04 15:00:00": 3160,
            "2021-08-04 15:30:00": 3414,
            "2021-08-04 16:00:00": 3616,
            "2021-08-04 16:30:00": 3763,
            "2021-08-04 17:00:00": 3869,
            "2021-08-04 17:30:00": 3955,
            "2021-08-04 18:00:00": 4021,
            "2021-08-04 18:30:00": 4072,
            "2021-08-04 19:00:00": 4109,
            "2021-08-04 19:30:00": 4134,
            "2021-08-04 20:16:00": 4139,
            "2021-08-04 21:01:00": 4139,

But then below that I get…

        "watt_hours_day": {
            "2021-08-04": 4139,
            "2021-08-05": 4555,
            "2021-08-06": 4882,
            "2021-08-07": 3893

Why are the watt hours day figures so low? are they showing watt hours total for the day in which case they are very low? or are they showing something else?

This is for…
https://api.forecast.solar/<API KEY>/estimate/<lat>/<lon>/43/-54/1.6

[BruceH5200 (Bruce Hartley)]

Sorry, I was being blind.
The figures throughout the day are obviously cumulative, so the daily ones are the end of each day.
I’ll now try and figure out why my hourly forecasts look wrong.

[blowk1 (Blowk)]

I can figure it out.
When calling for example https://api.forecast.solar/estimate/51.054340/3.717424/33/160/4.760, the results are different than the values in Hassio

Does anyone know why the values from Hassio are higher then the values from the API?

Update (8/11/21):
My mistake. The azimuth is different in Home Assistant compared to API. Changing it to the right azimuth in the API corrected the problem.

[Forecast.Solar (Knut Kohl)]

@BruceH5200

Why are the watt hours day figures so low? are they showing watt hours total for the day in which case they are very low? or are they showing something else?

watt_hours is summed up watt hours over the day and watt_hours_day is the estimate watt hours for the day.

Ther last watt_hours for the day should always match the watt_hous_day for the day.

 "2021-08-04 21:01:00": 4139,
...
"2021-08-04": 4139,