Dear All
I am working on parameter estimation of solar PV cells and modules with metaheuristic optimisation algorithms. With a 26 pairs of I-V data and also with a 46 pairs of I-V dataset from a published paper, RMSE is low and parameters are successfully found. However, with a 800 pairs of I-V points of a PV module, RMSE is around 0.28 that seems too high. Why this happens? It is due to the large size of used dataset? or I must apply any change when estimating parameters of PV modules rather than PV cells? And do you have any dataset of experimental I-V points of PV cells? I would like to use them as the case studies of my future paper.
Thank you
Kind Regards
Ahmad
Dear Ahmad,
You excite an important point which is the scattering of the I-V characteristics and consequently the performance parameters of the cells and modules such as the open circuit voltage, the short circuit current and maximum fill factor and the conversion efficiency. It is so that these parameters can be defined for the solar cells with the same technology and fabrication steps. It is so that the performance parameters scatter from cell to cell and from run to run. Manufacturer spend a lot of efforts not only to get reliable products but also the make their best to produce solar cells with close tolerances because these tolerances has adverse effects on the module and array performance. For example, the mismatch between the solar cells performance parameters leads to loss in the conversion efficiency. Looking at the short circuit current
the smallest short circuit current of cells connected in series will be inherited to the the whole string.
From my point of view you can make statistical analysis that are interested to give certain target performance parameters. That is you can make tolerance analysis for the cells of the production line.
Trying to widen the tolerance analysis for example single crystal solar cells with certain rated pea power from different production lines will give wider tolerances.
You can get data from the fabricators, they have the the greatest data bases or you get the data from the the big research institutes like SERI in USA.
Wish you success
Dear Azadeh
Thank you for your answer. Can you please refer me to the webpage including those V-I data?
Dear Ahmad
Before I say my idea, I should mention that I do not know the optimisation method you are using. However, I understood that you use different IV curves of modules and solar cells published in the papers! You did not mention whether the results are indoor or outdoor tests as Anton mentioned. You did not mentioned exactly what kind of parameters you are looking for as well.
You should note that:
- Each PV module has it's own solar cells which of course has different electrical characteristics! Even the same technologies have different designs which leads to different electrical and optical characteristics in module level
- You always have different loss and gain mechanism in PV modules. You can have different optical gains through cell metallization, backsheet or etc. as an example, if the cells are multi-Si, you have lower efficiency compared to Mono-Si. However, in module level, multi-Si cells have have higher cell to module power ratio due to reflections inside the module.
- There are always mismatch between cells and therefore, the final Isc of the module will be defined through the worst cell in the string.
- The Rs of cell to module changes due to interconnection losses. It means that of Rs in cell level is your mettalization but in module level is metallization+interconnection+junction box
- If the results from outdoor measurement, then the story is completely different !!! you have to consider a lot of other environmental impacts.
I am not pretty sure whether this helps you or not but you have not mentioned about the categories that you chose those IV curves.
Kind regards,
Hamed
Dear All
Can you please help me to find experimental I-V pairs data of PV cells such as multi- crystalline (S75 and S115), mono-crystalline (SM55 and SQ150PC) and thin film (ST36 and ST40)?
In a paper, I found that these data are available at this webpage, however, I could find nothing there.
http://www.solarcellsales.com/techinfo/technical_docs.cfm
Azadeh
Thanks for your help, however, in none of those PDF files, experimental I-V papers do not exist.
Could you please specify what exactly you need? I mean what kind of parameters? Only Isc and Voc or you want more information about recombination and diffusion effects?
The question is somehow vague in this case.
If you are looking for Solar cells from Shell! (which is not a well known solar cell producer), you can write them directly.
I have set of experimental data for solarbatteries. And I have set of SPICE-model parameters for Crystalline Silicon PV-battery.
The I(V)-curve for small Amorphous Silicon PV-battery is very differ from I(V)-curve for big Crystalline Silicon PV-battery. So simplest model with Photo-Current Source and Shunt Diode (classic Spice-model) is not suitable for Amorphous Silicon PV-battery.
https://www.researchgate.net/post/Can_someone_suggest_the_suitable_SPICE-model_for_Amorphous_Silicon_PV-battery
there are many reported work in parameter identification the highest accuracy can reach up to RMSE=9.9e-4
https://www.researchgate.net/publication/299343083_the_role_of_articial_intelligence_in_photo-voltaic_systems_design_and_control_A_review
Research the role of arti�cial intelligence in photo-voltaic systems ...
- Badges
- Science topic