Jump to content

Chen

Members
  • Posts

    47
  • Joined

  • Last visited

Everything posted by Chen

  1. What are the testing conditions for Arev? thanks!
  2. Dear Prof.André Mermoud: Recently, I have utilized PVsyst 8.0.19 for simulating the IV curves of BC modules under module-level and string-level shading scenarios. Given some inconsistencies identified in the string-level results, I wish to arrange a discussion with you on this matter. I increased the Arev factor of BC module to over 1000, and first simulated cell shading at the module level with 1 cell shaded. The results showed a 35% power reduction. When I used the string simulation tool to replicate the same number of shaded cells—i.e., a string configuration with one module per string and 1 cell shaded—the results displayed a significant discrepancy. Does it a bug? Does the Arev factor of BC modules fail to take effect at the string level? thanks! Best regards!
  3. what does the C0 C1 C2..C5 mean? thanks!
  4. Dear PVsyst team: For the Imp/Vmp contributions 80/20%, is there any relevant literature to support the value of this parameter? Thank you!
  5. Ok,I wonder if our software can be considered this.
  6. Dear sir: For the bi-facial module system , when calculating the cable loss of strings, should the current on the rear side of the modules be considered in our software? thanks!
  7. Chen

    tilt problem

    OK, u are welcome
  8. Dear sir: I found that the inclination angles of the two are not the same. Is it a bug?THANKS
  9. Dear sir: Can you take a look at this problem? Why is the backside mismatch loss calculated separately from the front side, rather than being added to the front side like irradiation. thanks!
  10. When I completely shade the front side of the module and allowed the back side to receive light, I found that the loss due to weak light was relatively high.
  11. OK,Thank you for your detailed answer.
  12. ok,I agree with your viewpoint, but do you think this calculation is reasonable? if the fornt side of module receive irradiance of 1000 W/m², and the rear side of module receive irradiance of 200 W/m², front side:(f × 22- f × 22) / (f × 22) = 0% (relative) rear side:(f × 21.34 - f × 22) / (f × 22) = -3% (relative)
  13. I understand. What I mean is, if the irradiation intensity on the back of the module is relatively low while that on the front is relatively high, so does it imply that the low-light performance of the rear side is worse?
  14. Because the irradiance of the front side or the rear side of PV module will be different, in this case, there will be a difference in conversion efficiency between the front and back sides,with respect to 1000 W/m².
  15. Dear sir: Has the software taken into account the low-light performance on the rearside of bifacial module? thanks
  16. Dear pvsyst team: If I have an irregular trackers system, with some different pitches and GCR, what is the difference between the reference trackers and automatic mode ( deselecting the reference pair of trackers)?what is the automatic mode? thanks!
  17. OK,thanks.
  18. Will the temperature increase and power decrease when local solar cells of photovoltaic modules experience hot spots be considered in the power generation simulation?thanks
  19. OK,thanks!
  20. Dear PVsyst team: I saw an article in the literature that mentioned a comparison between the settings of UC and UV and actual simulation tests. Can we see if there is still a need to optimize UC for bifacial modules? thanks de Oliveira A K V, Braga M, Naspolini H F, et al. Validation of Thermal Models for Bifacial Photovoltaic Systems under Various Albedo Conditions[J]. ESS Open Archive eprints, 2024, 63: 06323163. https://onlinelibrary.wiley.com/doi/abs/10.1002/pip.3892 CTG_PIP_Temperature.pdf
×
×
  • Create New...