Production prediction in solar power plants using deep learning

As industrial activities increase day by day, the population continuously grows, and the use of technological devices becomes more widespread, there is a consistent increase in energy consumption. Due to the impending depletion of fossil fuels and their negative effects on the environment, the demand for renewable energy sources is steadily increasing. This study has been conducted with the aim of contributing to the balance between the production and consumption of electrical energy in light of this increasing demand. The study examines four popular deep learning models used in production forecasts of solar power plants: Long Short-Term Memory (LSTM), Multilayer Perceptrons (MLP), Recurrent Neural Networks (RNN), and Gated Recurrent Units (GRU). The results indicate that the RNN and GRU models demonstrate higher prediction accuracy compared to the others.