Ocean data classification using unsupervised machine learning: Planning for hybrid wave-wind offshore energy devices

Wave-wind hybrid energy systems have become a recent focus due to the benefits they provide in terms of power reliability and long-term cost reduction for offshore power production. The efficiency of these systems partially depend on the wave and wind properties of the installation site. In this work, an unsupervised machine learning technique was used to classify the regions in the United States based on data provided by the National Data Buoy Center (NDBC). The aim was to create an initial assessment tool to improve decision-making in the design process for wave-wind hybrid systems with high adaptability within each region based on the recorded data. The data for all the weather stations was collected from 2010 to 2019. Three features of wave height, wave period and wind speed were used to cluster the regions during 2019, using the 72 stations with available data. The classification procedure was then repeated for the available data for five-year (2015-2019) and ten-year (2010-2019) periods. For the clustering purpose, the wave and wind properties for each year were considered as features. The characteristics of each cluster were discussed and the results were compared with the outcomes for the classifications based on one-year data. Further, detailed statistical information regarding each cluster were provided.

Automated summary

Wave-wind hybrid energy systems have gained attention for their benefits in power reliability and cost reduction for offshore power production. This study classified regions in the United States based on wave and wind properties using unsupervised machine learning, aiming to improve decision-making in system design with high adaptability. The characteristics of each cluster were discussed and compared with outcomes based on one-year data, providing detailed statistical information for each cluster.