I am trying to achieve two things in Radiance for study:
to convert the global horizontal irradiance to direct and diffuse irradiance
to convert the direct-normal-irradiance and diffuse-horizontal-irradiance (two types: one is full spectral and one is visible spectral)
Based on the two moves above, I’ve got two questions:
What command is more accurate for converting the data from GHI to DNI and DHI? Gendaylit or Gen_reindl (from daysim) ?
Can full spectral and visible spectral both be OK for Gendaylit input? It seems that there are scripts in Gendaylit to convert the the luminance values into radiance integrated over the visible range of the spectrum. Does it mean that the input has to be in full spectral range?
For 1) I would recommend to check literature about these models, there are plenty of studies, e.g. one of them from TU-Vienna: http://www.ibpsa.org/proceedings/BS2011/P_1322.pdf.
However, these models cannot make miracles and all of them show a high uncertainty. If ever available, measured data from both direct and diffuse/global is preferable.
In gendaylit, there is the Erbs model included, which is (or was) widely used for PV applications and showed a slightly better performance than the Reindl model for mid-European climates. However, I think from theory the Reindl model can consider also air humidity and temperature, which might end up with better results for humid climates (but this is just an hypothesis). But I’m not sure if this feature (temp+humidity) is included in gen_reindl. Basically both models (Erbs, Reindl) have a similar level of uncertainties.
For 2) : These are the input option for gendaylit (from manual page):
-W direct-normal-irradiance (W/m^2), diffuse-horizontal-irradiance (W/m^2)
-G direct-horizontal-irradiance (W/m^2), diffuse-horizontal-irradiance (W/m^2)
-L direct-normal-illuminance (lm/m^2), diffuse-horizontal-illuminance (lm/m^2)
Irradiance in gendaylit as input is always full spectrum and no spectral no weighting applied, whereas illuminance is using per definition v(lambda) weighting (and is therefore considering only the visible range). As described in the man page, for -W and -G a luminance efficacy model from Perez is applied (in addition to the Perez sky luminance distribution model).
Many thanks for your quick reply! I have some more questions based on the above.
For gendaylit, is there a way that I can convert a list of data at the same time.? I have already known that gen_reindl can covert a .txt file with a list of GHI.
If it is full spectrum, what range can it be considered as full-spectrum? Yesterday, I’ve tagged you in another topic, the SPN1 mentioned has a Spectral Response of 400-2700 nm. Can this be a full spectrum range?
For GHI decomposition into DNI and DHI you can use pvlib (a library in Python). Full spectrum, or broadband, at least for solar PV applications is from 300 to 2800 nm. This wavelength range is based on the spectral response of typical pyranometers.
You should input the irradiance for full solar spectrum in gendaylit. The default output is radiance integrated over the visible band unless you use the option -O1.
I have some more questions based on the above.