Artificial neural network-based output power prediction of grid-connected semitransparent photovoltaic system

Abstract

The solar photovoltaic system is an emerging renewable energy resource. The performance of the solar photovoltaic system is predicted based on the historical experimental dataset. In this work, the real-time prediction models are developed for the output power prediction of the STPV system. The performance of the semitransparent photovoltaic system is predicted for the Kovilpatti region where the climatic condition is hot and humid. The short-term power is predicted for the hourly, daily, and weekly average are considered. The feature selected for the prediction of the output power of the STPV system comprises of the solar radiation, ambient temperature, and wind velocity of the Kovilpatti region. The result reveals that the output power prediction of the hourly, daily, and weekly power have the very high value of the correlation coefficient of R. The final model produced accurate forecasts, with a Root mean square (RMSE) of 0.25 in ELMAN and 0.30 in FFN and 0.426 in GRN. These features of the training algorithm indicate that the model is not dependent on the model’s position or configuration in the simulation.