International Journal of Software Engineering for Smart Device
| Volume 12, No. 4, 2018, pp 19-24 | ||
| Abstract |
Short-Term Photovoltaic Power Generation Forecasting by Input-Output Structure of Weather Forecast Using Deep Learning
|
|
Abstract
In this study, hourly meteorological data and power generation data were used to predict sunshine, solar radiation, and solar power generation by the input-output variables of four models that are dependent on the weather forecast. The results show better predictions of data structures using weather forecasts than typical data structures. Meanwhile, the recurrent neural network (RNN) and the long short-term memory (LSTM), which are suitable for time-series data structures, performed better than the dynamic neural network (DNN). Model 4 provided the best results for the estimation of the sunshine and solar radiation.