Wind Power Forecasting Using Support Vector Machine Model in RStudio

Abstract

Wind energy has gained a lot of importance in few decades, as it is the cleanest form of renewable energy and available at free of cost. Wind power generation is increasing rapidly due to concerns about global warming and financial incentives from government. With the increased percentage of wind power in power grid, wind power forecasting has become essential to system operator for electric power scheduling and power reserve allocation. Wind power producers can get benefit of wind power prediction while bidding in the electricity market. This paper presents the wind power forecasting model using Support Vector Machine technique. In this model, wind speed, wind direction, air temperature, air pressure and air density are taken as input parameters to build accurate and efficient model. New statistical computing software called RStudio is used to develop prediction model. The model is trained and tested by using data of Dutch Hill Wind Farm available on website of National Renewable Energy Laboratory (NREL). According to results, the model performs significantly better than linear SVM and other regression models of SVM. The model is used to implement one hour ahead wind power forecasting and the results are validated with dataset.