Using machine learning to derive spatial wave data: A case study for a marine energy site

• DOI: 10.1016/j.envsoft.2021.105066
• Publication Date: 2021-05-05T12:49:28Z
• Authors: Jiaxin Chen; Ajit C. Pillai; Lars Johanning; Ian Ashton
• Keywords: Buoy; Wave model; Significant wave height; Wind wave model; Wave Power; Sea trial

Ocean waves are widely estimated using physics-based computational models, which predict how energy is transferred from the wind, dissipated, and spatially across ocean. Machine learning methods offer an opportunity to these data with significantly reduced input power. This paper describes a novel surrogate model developed random forest method, replicates spatial nearshore wave by Simulating WAves Nearshore (SWAN) numerical model. By incorporating in-situ buoy observations, outputs were found match observations at test location more closely than corresponding SWAN Furthermore, required time factor of 100. methodology can provide accurate in situations where power transmission limited, such as autonomous marine vehicles or during coastal offshore operations remote areas. represents significant supplementary service existing models.