How to tell SAM to load all the years downloaded instead each year at time?

I want to use all the files from 1998 to 2020 in my simulation, but I guess SAM only uses the selected one.
The files are download all in a separated folder, I thought SAM was using all, but when I click to see the data only shows one year at time.
Am I loosing something?
I'm asking because when I choose automatic download SAM downloads a file with many "random" years, some years are like 2020 or 2019 are not in the file...

1 - Image from the separated folder


2 - Image from the files


3 - Image from SAM

Hi Leandro,

SAM's performance models use a single weather file that contains one year of data. It is not possible to run a simulation over multiple years using a sequence of weather files.

Weather files may contain "typical meteorological year" (TMY) data or single-year data. TMY data represents the typical solar resource over a multi-year historical period so that a single file contains data from different years. You can read more about TMY here: nsrdb.nrel.gov/about/tmy.html .

When you use the Download all years for P50/P90 analysis option to download weather files from the NREL National Solar Radiation Database (NSRDB), SAM downloads a file for each year in the collection (1998 to 2020 as of now) and puts them in a separate folder to make it easier to run P50/P90 simulations. See the "P50/P90 Analysis" video here for a demonstration: sam.nrel.gov/weather-data/weather-data-videos.html . For a paper describing how P50/P90 simulations work, see the publication listed under "P50/P90 Analysis in SAM" here: sam.nrel.gov/weather-data/weather-data-publications.html .



Best regards,
Paul.

Thank you so much!

Does SAM TMY gets the best hour data point to represente that specific hour for all same hour available in different years or it gets the whole month hour data?

For example:
Data from 2010 to 2020:
I'm assuming for example, the data point 15/06/2015 at 02:00 p.m. represents better (better median) of all the others 15/06/20xx at 02:00 points...
Or SAM gets the whole month (06) to represent better all the others 06/20xx months?
  If yes:
   Because then maybe, that speciicf month has a day with a worse result (worse median), affecting the simulation.


Hi Leandro,

Each TMY file from the NREL NSRDB is constructed by joining together 12 months from different years so that all of the data in a given month comes from the same year: All of the January data comes from one year, the March data from a different year, etc. The months are chosen from the historical period to represent the "typical" month over the period. This ensures that the data for a given hour (solar resource, temperature, wind speed, etc.) is always from the same hour.

Best regards,
Paul.

For uncertainty calculation it can run time-series. It would be pretty useful to add this for 8760 performance modeling. .