forecast.solar gives a to low estimate for Southern Norway
Trying to use forecast.solar to estimate what I can expect for the next 12/24/48 hours, I always find the actual output to be to higher than what forecast.solar estimates. The included image is fetched from Home Assistant with both the forecast.solar integration, and a rest api sensor as well. Not sure how well the colors are seen, but here is an explanation: Orange: today forecast using api, updated every 30 minutes. Blue: forecast for tomorrow, else same as above Red: the sensor for what is expected in 24 hours, fetched from the Home assistant API Purple: the same as above, but in 12 hours Yellow: actual output from the inverter green: What the API gives at the beginning of each hour.
There might be a small error in the api-call, that the azimuth is -18 in the call, but should be -25, but when I run the api manually with -18 or -25, there is not much difference.
For yesterday, the peak power was 5510watts, while forecast.solar estimated 4610.
Here is the url I'm using (I've left out the api-key): http://api.forecast.solar/estimate/58.35/8.59/22/-25/7
Not sure what the history-call should output, I get almost the same as the estimate, that is today, and the next few days. I would expect maybe today, and the last few days.
The total estimate (from forecast.solar) for the days, fetched from the end of each day: 2022-04-13: 28651 Wh 2022-04-14: 28907 Wh 2022-04-15: 28979 Wh 2022-04-16: 34750 Wh 2022-04-17: 34357 Wh 2022-04-18: 29541 Wh 2022-04-19: 33783 Wh
The real production for the same dates 2022-04-13: 7.1 kWh (the weather was not as the forecast ) 2022-04-14: 11.8 kWh (Same bad weather forecast) 2022-04-15: 44.2 kWh 2022-04-16: 43.3 kWh 2022-04-17: 43.2 kWh 2022-04-18: 43.9 kWh 2022-04-19: 42.0 kWh
When the wether is nice, I always get ~20% more than what forecast.solar says. Should I increase the kwp parameter of the call, and add max inverter output as well ?