From Wind to Wave energy resource: forecasting methods analysis

The need for a more sustainable future led humanity to investigate new renewable energy sources, such as waves. Just like wind, the greatest advantage of the waves over other alternative energy sources is that it is easily predictable through the study of winds. Adequate wave predictions are crucial for the proper mapping and design of the wave farms that will contribute more and more to renewable energy generation in the future. The thesis aims to provide a general overview of wave prediction methods and then to build, analyze and compare three different types: an empirical method, that is the Sverdrup-Munk-Bretschneider (SMB) method, a multiple regression method and an artificial intelligence method, specifically a machine learning one called Artificial Neural Network (ANN). The selected site is the Island of Pantelleria. In the first phase, the models will be built using data obtained from ERA5, that is the fifth generation ECMWF atmospheric reanalysis of the global climate. Subsequently, data acquired from a buoy located on the site of interest will be implemented and the results will be compared to the previous ones. The goal is to find the most suitable prediction method for the selected site.