PVWatts V1 - Operating Cell Temperature = Dry Bulb Temperature

Hello Paul,

I´m having some issues with the calculations in PVwattsv1, which I believe stem from the thermal model calculations.
I´m using my own poa values (user_poa), which, as you can see in Fig.1 (output values from a simulation run of pvwattsv1 into an excel sheet) are being read correctly and fed into the algorithm.

However, for what seem to be random hours of the day, the calculated Operating Cell Temperature (tcell), seems to not be correctly calculated, and returns a value barely different from the dry bulb temperature for that hour (as you can see in Fig 2., a print screen of the Weather file being used, with the dry bulb temperatures for the problematic hours highlighted)


I would like to know your take on this, I´ve looked over the Fuentes model calculations on the pdf available on your website (Fuentes M.K., A simplified Thermal Model for Flat-Plate Photovoltaic Arrays, 1987) but can´t seem to find a reason for this error.

Any help would be greatly appreciated,
Kindest regards,
Miguel D.

Dear Miguel,

I would expect the cell temperature to be close to the ambient temperature under some conditions, for example during a period of low irradiance and lower ambient temperature with higher wind speed.

In the portion of the weather file that you show in the image, the DNI and POA values look unrealistically high to me. Those should be in W/m2. I would expect the maximum DNI value to be less than 1,400 W/m2.

In the data you show, the "User POA" values range from 97 to 3471, and the "DNI" values range from 852 to 3028. Can you check your DNI and POA data?

Best regards,
Paul.

Dear Paul,

Thank you once again for your very swift reply. I was indeed using fictional (and unrelated) values for DNI and POA, derived from real user data but multiplied by factors of 4 or 5, because I wanted to test the results with very different inputs(and am still waiting on the real inputs to be passed on to me).

I assume that the errors I was getting stem from the fact that these values were totally unrealistic, since now that I´ve run the simulation with the original values I´m not getting any errors! Thank you!

However, this brings another doubt to my mind: I can see that changing the DNI input values still significantly affects the end results, even keeping user_poa constant. Would this be due to the AOI correction? I can see no other module that takes DNI as an input.

As always, ever grateful for your help,
Kindest regards,

Miguel D.

Dear Miguel,

Yes. When you use POA data as input to pvwattsv1, the model uses the DNI data from the weather file in the transmitted POA calculation as described in Section 7 of Dobos (2013) "PVWatts Version 1 Technical Reference" NREL/TP-6A20-60272.

Best regards,
Paul.