Data Mining and Modelling for Wave Power Applications Using Hybrid SOM-NG Algorithm

Abstract

Renewable energies are increasing their importance in the current society so technical research is increasingly focusing on all of them. An uncommon one is wave power which obtains energy from wave displacements instead of sea level as tidal power. In this work a one year long data set containing hourly measures done by two different buoys was studied. Variable selection was done using a hybrid Self-Organizing Map with a model based method and the same algorithm was used to create a more accurate model to estimate the wave power from atmospheric data. The goals are to demonstrate the adaptability of the algorithm and to increase the knowledge about wave power. © Springer-Verlag Berlin Heidelberg 2013.