Guidance on PVlib model overprediction during sunny conditions

[Rana Badar Ashraf]

Dear all,

I am developing a pvlib-based PV power forecasting model for a rooftop PV installation. The physical modelling chain is configured as follows:

• Hay–Davies transposition model for plane-of-array irradiance
• SAPM temperature model for cell-temperature estimation
• De Soto single-diode model for DC power calculation
• PVWatts inverter model for DC-to-AC conversion
• 
  
• The model consistently overpredicts PV output during clear and high-irradiance days. I would appreciate your guidance on the possible causes of this sunny-day overprediction and its solution.  
• 
  
• Best regards,
• Muhammad Badar Ashraf Rana

[cwh...@sandia.gov]

Hi Muhammad,

We'll need to know a bit more information:

1. How large of a difference do you see? Even with data perfectly representing the system and environment, the models have error, on the order of a few percent at best.

2. What time of day do you see overpredictions? Throughout the day, only at early/late hours, or some other pattern?

3. Where did you get the parameters for the module and inverter? How confident are you that these parameters represent the system that is built?

4. How confident are you that the actual system is operating free from losses? A pvlib-based model typically represents a system in an ideal state, i.e., without shading, disconnected strings, misaligned racking, etc.

5. What is your source of irradiance and temperature data?

Regards,

Cliff