Machine learning approach for optimal determination of wave parameter relationships

Abstract

Wave parameter relationships have long been determined using methods that give non-standard and often inaccurate results. With increased commercial activity in the marine sector, the importance of accurate wave parameter relationship determination has become increasingly apparent. The outputs of many numerical models and buoy datasets do not include all requisite wave parameters, and a typical approach is to use a constant conversion factor or relationship based on defined spectra such as the Bretschneider or the joint North Sea wave observation project (JONSWAP) spectrum to determine these parameters. Given that relationships between wave parameters vary significantly over both hourly and seasonal and annual timescales, the currently employed methods are lacking, as subtleties are missed by the simpler approach. This paper addresses the determination of wave parameter relationships using a machine learning (ML)-based model, identifying and selecting the optimal method for the conversion of wave parameters (T e , T 01) in coastal Irish Waters. This approach is then validated at two sites on the West coast of Ireland. The aim is to highlight the utility of ML in approximating the relationship between wave parameters; using both buoy and modelled data, and mapping the predicted outcomes for a wave energy converter based on a variety of ML and measure correlate predict approaches.

Artificial Intelligence and Machine Learning in Energy Conversion and Management

“…It was employed to identify and opt for the most suitable approach for transforming wave parameters in coastal Irish Waters. This endeavor culminated in the attainment of a remarkably low RMSE value, specifically amounting to 0.17 [163]. Also, a multivariate polynomial regression model is implemented to produce a light-dependent resistor-based solar tracker that can learn and improve its action from daily interaction with the environment, which achieved a 0.0806 RMSE score [164].…”