Wind Farm Optimization with WRF and artificial intelligence

The worldwide trend towards renewable energy is becoming more important for the world and the living beings in the ecosystem with the increasing surface temperature as a result of the emission of greenhouse gases due to the use of fossil fuels and the melting of glaciers. The share of wind energy in the energy market has been increasing over the years and becoming more widespread. At least one year of wind data is required for wind power plant installations. There are some economic and geographical obstacles to obtaining these data by measurement. In sea and ocean installations, providing accurate data and planning the initial installation is critical for transmission and production efficiency. Short-term and rapid prediction of wind components allows preventing possible problems and planning maintenance and repairs. Today, weather forecast models in which atmospheric conditions are represented by physics equations and wind data can be predicted reliably with artificial intelligence. The aim of this study is to run the WRF numerical prediction model with appropriate variables and to ensure that appropriate wind speed and power potential estimates are estimated accurately and quickly with the help of artificial intelligence. In this study, a hybrid model was created using the Global Forecast System (GFS) and hourly measurement data with a resolution of 0.25 degrees, weather forecast model data for a one-year period, and a multivariate recursive neural network. Our predictions, made with atmospheric parameters and historical data with the Weather Research and Forecast Model (WRF) model in Turkey, were improved with the Long-Short Term Memory (LSTM) model and a prediction success of 0.04 average absolute error (MAE) was achieved. With these short and long-term forecasts, lower-cost projects can be developed in the Turkish energy market and possible future investments, and Turkey's wind energy potential can be comprehensively handled and used with more efficient wind energy production.