Three years ahead solar irradiance forecasting to quantify degradation influenced energy potentials from thin film (a-Si) photovoltaic system

Abstract

Random forests estimates (RFE) based machine learning algorithm is proposed for forecasting the three years ahead solar irradiance. Accordingly, the corresponding degradation rate (DR) influenced energy potentials are evaluated. Here, the DR of amorphous silicon (a-Si) PV system is estimated based on the simulated performance ratios with historical weather data. Prediction of energy potentials is helpful in decision making on improving the solar power project implementations in selected tropical savanna climates region of south India. Keywords: Thin film (a-Si) PV, Machine learning, Degradation rate, Solar forecasting, Year ahead PV power, Performance ratio