Short-Term Forecasting of Wind Power Plant Generation Based on Machine Learning Models

The current excessive exploitation of fossil energy sources has led to skyrocketing prices and significant environmental challenges, including greenhouse gas emissions and climate change. In response, there is an increasing focus on renewable energy sources as viable alternatives to fossil fuels. Among these alternatives, wind energy is particularly noteworthy due to its unlimited availability, cost-effectiveness, and minimal environmental impact. Two critical factors that influence wind energy generation are wind speed and direction, both of which tend to be inherently unstable. To address this, machine learning algorithms have been developed for forecasting wind energy in the contemporary technological landscape. This forecasting process leverages historical data to predict future outcomes. This paper will specifically examine machine learning techniques - Support Vector Regression (SVR), Linear Regression (LR), and Random Forest Regression (RFR) - to predict wind power using weather data as input. The results from these prediction models, developed with a 4-day forecast at 1-day intervals, will be compared against actual results from preceding periods. All methodologies and visualizations will be implemented using Python and Visual Studio Code software.