Real production versus forecast production - anyone got data?

_ptashek_

Does anyone here use forecast data from Forecast.Solar (https://forecast.solar/) or other PV production forecasting system? I wonder how accurate this data is for Ireland, compared to actual production.

Why I ask is because I plan to add PV to my house some time next year, and have enough roof space to install 21.5kWp in 430W panels. But it's all SE (103deg), NW (283deg) and NE (13deg) aspects, so not exactly prime locations. Been using forecast data for the past couple weeks to see what production might look like from this setup. Today, based on 15 minute intervals, it forecasted 29.2kWh (11.1 NW, 14.8 SE, 3.3 NE), between 8am and 5pm. At least it's clear the NE aspect would be a waste of money.

bullit_dodger

You probably want to plug in your details into.....

JRC Photovoltaic Geographical Information System (PVGIS) - European Commission (europa.eu)

and see what you get.

Note: 21.5Kwp is a humungous system. Your well into a thing called NC7 terrority with respects to ESB. Three phase most likely. While there are a few on here with 10+kWp systems, they are rare. Most people unless you have an enormous consumption would have 6-9Kwp and do pretty well.

TerraSolis

https://www.boards.ie/discussion/2058331883/real-production-versus-forecast-production-anyone-got-data

NREL's PVWATTS is fairly accurate for this and very straight forward to use.

https://pvwatts.nrel.gov/

SharkMX

https://www.boards.ie/discussion/comment/121632815#Comment_121632815

That one is underestimating our actual generation for the last 2 years by about 25%

ECO_Mental

Here is my data for my sytem for the last 2 and bit years, in this period I have beat PVGIS by nearly 10%

TerraSolis

https://www.boards.ie/discussion/comment/121633095#Comment_121633095

Yeah the default parameters for the losses are too high and if your closest weather station suffers form some sort of bias it's skewed further. Numbers I use are as below:

I assessed PVGIS vs some other sources like renewables.ninja & feeding ERA-5 weather reanalysis data into PVLIB python for a study a couple of year back and found it rather inaccurate.

Mr Q

I used SMA Sunny Design for my estimate.

It was within 1% for 2023 based on a full year. Didn't check the monthly figures separately.

_ptashek_

Thanks everyone for the details!