Time Zones to adjust for hourly weather data

4 posts / 0 new
Last post
mikeroberts
Time Zones to adjust for hourly weather data

Hi Paul,

I'm creating real year weather files to use for modelling PV systems at a number of sites in Australia. I'm using hourly satellite irradiance data from the Bureau of Meteorology, along with hourly temp and windspeed from BOM Automatic Weather Stations.

I'm using the SAM SDK to calculate DHI for the weather file, but am having some difficulty trying to use timezones to correct for the fact that the irradiance data is recorded at a number of minutes past the hour.

The Satellite data gives GHI and DNI, tagged with the hour (hh:00) but the actual time the reading was taken varies with latitude. As an example, I have a site where the readings are taken 49 minutes after the hour.

I am using the SAM SDK to calculate DHI by supplying the GHI and DNI data to the irradproc module, called from within python, with irrad_mode set to 1 and using the Perez sky model. As I have tilt and azimuth set to zero, the poa_skydiff returned should give me DHI.

The timezone here is UTC + 10.

My understanding is that if in the absence of minute data (or if the minute column is set to zero) SAM would assume the GHI and DNI readings I am sending are taken at 30 minutes past the hour. In order to correct this,I am adjusting the timezone tz, but am not sure what value to use. I have tried tz = 10.82 (49 minutes is 0.82 of an hour) and tz = 10.32 (19 minutes or 0.32 past the half-hour).

Both are giving strange results, in particular a bad fit between DHI and GHI on very cloudy days and a dip and spike in DHI towards the end of the day. In fact, for these days, I get more 'normal' looking results using tz = 9.3 or thereabouts, but I don't understand why. I suspect I'm missing something obvious.

I've attached a few files with different values for tz.

Is this approach of shifting the timezone the correct one? And if so, what value should I be using?

Thanks for your help.

Paul Gilman

Hi Mike,

Have you tried adding a minute column to the data? You could set its value for all hours to either 19 or 49 minutes. If you did that, you could keep the correct time zone, and then the irradiance data would be calculated using sun angles for the correct minute of the hour.

Best regards,
Paul.

mikeroberts

Thanks Paul.
Yes, using minutes set to 49 seems to work.
I'm not sure why i didn't try this myself instead of doing a workaround. (Doh!)
But, out of curiosity, I'm interested to know why the work around didn't work.... i.e. why is 00:49 with TZ=10, different to 00:00 with TZ=10.83 ??

Paul Gilman

Hi Mike,

I'd have to look more carefully at this to answer your question, but now that SAM is open source, I can point you to the relevant code. SAM's Detailed Photovoltaic model uses the sunpos() function in lib_irradproc.cpp to calculate the sun position. You can see how it uses the time zone in those equations by searching the following file for "tz" -- for example see Line 152:

https://github.com/NREL/ssc/blob/develop/shared/lib_irradproc.cpp

Best regards,
Paul.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer