Redefining fatigue predictions: a multi-sea state HPC framework for FOWTs

Offshore wind is essential in the global transition to Net Zero carbon emission goals. As the industry pushes into deeper waters, fixed offshore wind solutions are no longer viable, increasing the reliance on floating alternatives. During their operational lifespan of at least 25 years, floating wind turbines are exposed to stochastic winds, waves, currents and the non-linear coupled loads, making fatigue assessment critical in their design and maintenance planning. The industry standard approach is to group similar conditions together into bins, each with a corresponding probability of occurrence based on historical data. However, by assuming all bin members are equivalent, this binning approach results in a loss of information, leading to inaccuracies. Here we propose a more detailed approach, called Numerical Prototype approach, where every individual sea state is considered, produces in turn fatigue estimates expected to be closer to reality since less information is lost due to binning. This paper studies the UMaine VolturnUS-S semi-submersible platform with the IEA 15 MW turbine for with a modified tower for an Atlantic site on the west of France. For the turbine tower, the Numerical Prototype results indicate lower cumulative fatigue estimations by 24 % for the principal direction of fatigue than those calculated using the classical binning method, while for the mooring line fairleads fatigue estimations are up to 14 % lower. These findings suggest that a more discretised calculation and more detailed representation of met-ocean loads lead to lower fatigue predictions, revealing the conservative nature of existing industrial methods. The binning methods currently used in industry result in conservative designs with increased material use and increased costs of floating wind turbines. The present results indicate for the first time a detailed methodology for fatigue estimation that allows optimised designs to reduce structural weight with consequent savings in both installation and material costs. Although the proposed methodology is computationally expensive, the potential savings offer significant benefits for project developers.