I am trying to estimate PV output from daily GHI only.
Is there a way to convert the daily GHI to total radiation on a tilted surface ie solar panel? Or do I have to use a model to convert to hourly GHI values and then calculate the beam and diffuse values etc?
GHI data I am using is here. (http://www.bom.gov.au/jsp/ncc/cdio/wData/wdata?p_nccObsCode=193&p_display_type=dailyDataFile&p_stn_num=056037&p_startYear= )
Thanks
Sorry afaik there isn't a method in pvlib currently that will approximate daily poa directly from daily ghi. And there isn't an upsampling method to convert daily to hourly ghi either. Both of these features would be welcome additions, and I urge you to open issues for them.
I like your ideas, but without knowing the distribution of clearness index or clear sky index I think it would be difficult to do a faithful upsampling.
I think to approximate poa, you will need to decompose the hourly ghi.
You can follow this example here: https://pvlib-python.readthedocs.io/en/latest/auto_examples/plot_ghi_transposition.html#sphx-glr-auto-examples-plot-ghi-transposition-py
If you can get the distribution of hourly clearness indices, perhaps from a nearby SURFRAD site? Then maybe you can upsample your daily ghi to hourly and use one of the dni estimation methods, Erbs is good, to decompose your upsampled ghi to beam and diffuse components. Then transpose to poa as in the gallery example linked above.
