Determining Meteorological Parameters Influencing Photovoltaic Solar Energy Generation in Quezon City Using Machine Learning Algorithms

Abstract

One challenge in adapting to energy generation using solar photovoltaic (PV) modules is its variability with changing weather conditions. In this study, we aim to determine the effect of meteorological parameters that have the most effect on the variability of solar energy generation (SEG). Our study is conducted in Quezon City, part of the National Capital Region, Philippines. The maximum temperature, relative humidity, man temperature, and cloud opacity have the most effect on the variability of the SEG among the eight meteorological parameters that we consider in our study based on the principal component regressor (PCR) and random forest regressor (RFR) machine learning algorithms. The PCR model explains 55.5 and 49.2% variability in SEG of the training and test sets, respectively. On the other hand, the RFR model explains a 77.1% variation of the SEG in the training and 52.7% in the test set. Furthermore, the two models provided comparable predictions of SEG.