Prediction of Custom-Built Bi-Facial PV Panel Output Including Weather Parameters

A procedure for predicting the power output of a custom-made bi-facial PV panel which includes weather parameters like temperature, pressure and humidity is proposed in this article. Weather data as well as panel output were recorded from an experimental bench setup on the rooftop. The data was recorded for a period of three months in the winter season. The bi-facial PV panel was built by using two mono-facial panels hooked back to back. Three different expectation functions for arriving at the panel output were considered. The correlation between the recorded weather parameters and PV panel output were evaluated. It was found that multi-variable non-linear models are required to analyze the expectation functions and to predict output power. The recorded data from the test bench was used to train the proposed expectation functions. The developed algorithm showed excellent match between recorded and predicted outputs. The R-squared and variance properties were very good.