how to chose temperature, dc, etc. models?
42 views
Skip to first unread message
Kamila M
Jan 2, 2023, 10:47:08 AM1/2/23
to pvlib-python
Good day, everyone!
I'm trying to simulate a real pv system using the pvlib library.
Are there any rules or speculations what dc/ac/aoi/spectral/temperature model is better for a particular case if I have system description which is suitable almost for each model?
I compared pv output calculated with different temperature models and it was almost the same, but aoi models, for instance, had greater impact on the pv output.
Also, there can be different combination of temperature, aoi, ... models.
Maybe there are also some rules how to chose model, if it' possible to use only faiman temperature model.
Thank you for your response!
Kamila
Marvin Valentin
Jan 2, 2023, 7:43:32 PM1/2/23
to Kamila M, pvlib-python
HI Kamila.
I am using aoi. You can use one source for simplicity so long as you get a clear idea about the process.
cheers,
marvin
MARVIN T. VALENTIN
--
You received this message because you are subscribed to the Google Groups "pvlib-python" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pvlib-python...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pvlib-python/962b8f3c-14a0-43f3-9d44-dd07e4612350n%40googlegroups.com.
Kamila M
Jan 5, 2023, 10:39:49 AM1/5/23
to pvlib-python
Hi, thank you for your response!
Maybe I stated the question not clear enough.
Talking abot aoi calculations, what kind of aoi model ('physical', 'ashrae', 'sapm', 'martin_ruiz', 'no_loss') is better for a particular case? Or maybe one of them is suitable for all cases.
The same concern about temperature models ('sapm', 'pvsyst', 'faiman', 'fuentes', 'noct_sam'), dc madels (sapm', 'desoto', 'cec', 'pvsyst', 'pvwatts') and so on.
вторник, 3 января 2023 г. в 01:43:32 UTC+1, marvin....@upwr.edu.pl:
Kenneth J. Sauer
Jan 7, 2023, 8:49:40 PM1/7/23
to Kamila M, pvlib-python
Hi Kamila,
I'll try providing an answer and Cc'ing the group:
I think to some extent it's a matter of opinion and use case. For
example, are you trying to have comparable simulations to NREL's
System Advisor Model (SAM) software? Or to PVsyst? PVWatts? And, what
module technology are you modeling? A standard PV module? A BIPV
module? A Tesla roof tile? e.g., The IAM model you choose may
"reflect" the particular IAM curve of one technology better than
another. Such (hypothetical) questions & answers might influence your
choice of models.
Arguably, PVsyst is the most "bankable" solar PV energy simulation
algorithm today, so my quick opinion is to start by attempting to
align with/match that (as many modelers commonly do). However, if you
don't have a PVsyst license, then you might want to try matching the
algorithm (i.e., sequence of models) of whatever other software you're
running instead for comparison, e.g., the free & open-source NREL's
SAM, in order to get a working & well-calibrated/well-tested PVLIB
algorithm up-and-running.
For PVsyst comparisons:
IAM model: Use the physical IAM model (based on Fresnel equations &
Snell's Law) for standard PV modules with solar glass. ASHRAE is also
an option in PVsyst which may "reflect" certain window-like BIPV
glasses better (or other less common use cases), however, this is
uncommon. (In my personal opinion, the ASHRAE model is mostly obsolete
for solar PV.) IAM characterization testing is needed to determine the
optimal model choice and fitted model parameters of said model.
(Again, for simply setting up your PVLIB algorithm, just choose
whichever model aligns with the software you're comparing to and apply
the same model parameters in both your PVLIB algorithm and the other
software used for algorithm benchmarking aka crosschecking.)
Thermal model: Use PVsyst which is based on Faiman. If you have wind
speed data, Uv should not be equal to 0.
Any other Qs? Feel free to ask. Or you could try posting your question
to the PVLIB Stack Overflow webpage:
https://stackoverflow.com/questions/tagged/pvlib/
In summary, for getting a PVLIB algorithm up-and-running, choose the
sequence of models and set of user-inputted model parameters that
matches to the software you're comparing to. Alternatively, for
precise modeling of a specific PV technology (or set of technologies,
i.e., system), you may need to do some testing (or review/evaluate
testing data) of said technology in order to determine the best (most
appropriate) model and best fit coefficients of said model. I hope
that makes sense & helps answer your question.
Last but not least, the PVPMC website itself has a nice overview of
the various models, if you haven't seen it already. e.g., Click on
"Modeling Steps":
https://pvpmc.sandia.gov/
--Best,
Kenneth Sauer
> To view this discussion on the web visit https://groups.google.com/d/msgid/pvlib-python/925fcb08-e911-4ed6-9612-b77a8d8f9739n%40googlegroups.com.
I'll try providing an answer and Cc'ing the group:
I think to some extent it's a matter of opinion and use case. For
example, are you trying to have comparable simulations to NREL's
System Advisor Model (SAM) software? Or to PVsyst? PVWatts? And, what
module technology are you modeling? A standard PV module? A BIPV
module? A Tesla roof tile? e.g., The IAM model you choose may
"reflect" the particular IAM curve of one technology better than
another. Such (hypothetical) questions & answers might influence your
choice of models.
Arguably, PVsyst is the most "bankable" solar PV energy simulation
algorithm today, so my quick opinion is to start by attempting to
align with/match that (as many modelers commonly do). However, if you
don't have a PVsyst license, then you might want to try matching the
algorithm (i.e., sequence of models) of whatever other software you're
running instead for comparison, e.g., the free & open-source NREL's
SAM, in order to get a working & well-calibrated/well-tested PVLIB
algorithm up-and-running.
For PVsyst comparisons:
IAM model: Use the physical IAM model (based on Fresnel equations &
Snell's Law) for standard PV modules with solar glass. ASHRAE is also
an option in PVsyst which may "reflect" certain window-like BIPV
glasses better (or other less common use cases), however, this is
uncommon. (In my personal opinion, the ASHRAE model is mostly obsolete
for solar PV.) IAM characterization testing is needed to determine the
optimal model choice and fitted model parameters of said model.
(Again, for simply setting up your PVLIB algorithm, just choose
whichever model aligns with the software you're comparing to and apply
the same model parameters in both your PVLIB algorithm and the other
software used for algorithm benchmarking aka crosschecking.)
Thermal model: Use PVsyst which is based on Faiman. If you have wind
speed data, Uv should not be equal to 0.
Any other Qs? Feel free to ask. Or you could try posting your question
to the PVLIB Stack Overflow webpage:
https://stackoverflow.com/questions/tagged/pvlib/
In summary, for getting a PVLIB algorithm up-and-running, choose the
sequence of models and set of user-inputted model parameters that
matches to the software you're comparing to. Alternatively, for
precise modeling of a specific PV technology (or set of technologies,
i.e., system), you may need to do some testing (or review/evaluate
testing data) of said technology in order to determine the best (most
appropriate) model and best fit coefficients of said model. I hope
that makes sense & helps answer your question.
Last but not least, the PVPMC website itself has a nice overview of
the various models, if you haven't seen it already. e.g., Click on
"Modeling Steps":
https://pvpmc.sandia.gov/
--Best,
Kenneth Sauer
Reply all
Reply to author
Forward
0 new messages