Daily Electrical Energy Forecasting in Rooftop Photovoltaic Systems: A Case Study

Rezi  Delfianti; Eki Rovianto; Catur Harsito; Johan Alfian  Pradana; Viola  Pongajow; Salaki Reynaldo  Joshua

Authors

Rezi Delfianti Faculty of Advanced Technology and Multidiscipline, Airlangga University
Eki Rovianto Mechanical Engineering of Vocational School, Universitas Sebelas Maret
Catur Harsito Mechanical Engineering of Vocational School, Universitas Sebelas Maret
Johan Alfian Pradana Industrial Engineering, Faculty of Industrial Technology, Adhi Tama Institute of Technology, Surabaya
Viola Pongajow School of Metallurgy, Central South University
Salaki Reynaldo Joshua Department of Electronics, Information and Communication Engineering, Kangwon National University

Keywords:

Forecasting, Photovoltaic, Renewable energy, Electrical

Abstract

The tropical climate provides room for optimizing solar energy. However, the ability in functionality is not optimal because the existing photovoltaic has not been investigated how much energy is produced. This research provides precise daily predictions for the sustainability of Photovoltaic energy. The resulting photovoltaic energy will undergo a predictive model by comparing NASA data with the predicted results from observational data. First, the mean square error (MSE) step is carried out with the smallest target value, determining the value of the correlation coefficient from training, validation and test data. The resulting prediction is at a validated daily energy of 9.46 in epoch 124. The prediction success is 99.98% for 4500 days with a standard of 2.2 x 105 kWh.