Prediction of Renewable Energy Generation Using Machine Learning Methods
Due to the limitation of natural resources where conventional energy can be obtained, other energies sources must be found without such limits. For this reason, we are trying to cut down the usage of conventional energy by using renewable energy sources. In this thesis, we are going to look into different types of energy sources and compare them. Then we are going to focus on the hybrid system, its components, and its benefits. In order to have a reliable hybrid system each component has to be reliable and deterministic, but unfortunately, renewable energy sources such as wind energy and solar energy alone are not always both reliable and deterministic. One way to solve this problem is by integrating a backup system so energy output could always be steady as needed. By using solar energy during the day, we can help minimize the emissions of carbon dioxide produced by fossil fuels and slow down the depletion of fossil fuels. For this reason, there must be a reliable forecasting that can help predict the power output at a certain time. We will present several methods to predict power output such as ARIMA, ARMA, and artificial neural networks (ANN). Ultimately the goal will be to minimize the cost of electricity for the average consumer while at the same time helping reduce the use of conventional energy.