Predicting the Energy Production of a Rooftop Pv Plant by Using Differential Evolution Algorithm

Abstract

In this study, a simple and plain closed-form mathematical expression has been obtained to precisely estimate the monthly production of a rooftop photovoltaic (PV) plant installed in Adana, Turkey. The proposed model is developed by utilizing the Differential Evolution (DE) optimization algorithm based on the PV plant's 5-year (August 2013 -July 2018) real measurement data. The PV plant is a grid-connected rooftop solar PV system located at Kıvanç Textile in Adana, Turkey. The PV system is equipped with an online monitoring system that provides real time data. The study shows the actual energy production is 730 MWh/year on average for 5 years. In order to test the robustness and precision of the present model, it has been compared with the long-term real measurement data of the PV system. The key benefit of the model is giving a convincing prediction of the future production of the PV panel in a simple way. It also does not require any further information other than time. Average percentage error was reached as small as 7.4% for 5-year data.