Unit Three Kartini, Ulin Nikmatul Choiroh
Using a new hybrid technique, this research study provides a 1-hour prediction for solar power output from a power producing facility. When it comes to improving solar power generation operations, meteorological data plays an important role. The Decomposition and Deep-Neural Network approaches are used to create this hybrid model. This new hybrid model incorporates meteorological data. As an input to the Deep Learning Neural Network model, the first hybrid model uses decomposition. It is the mean absolute percentage error of the Decomposition-Deep Neural Network) model that serves as a statistical indication of statistical error. Decomposition Deep NN-based predictions are compared to observed data, decomposition technique, and simulation results, which show that the approach provided in this research can determine hourly solar power generation with acceptable accuracy. © 2021 IEEE.
Universitas Negeri Surabaya, Dept. Electrical Engineering, Surabaya, Indonesia
