FLOATBench: Benchmark for Floating Offshore Wind Turbine Fatigue
FLOATBench serves as a public benchmark dataset aimed at forecasting fatigue damage in floating offshore wind turbine (FOWT) towers. It comprises 582,120 fatigue-damage labels per section from three distinct 22 MW FOWT tower designs, generated through 19,404 detailed OpenFAST simulations (6,468 for each tower, with 1,078 specific wind/wave operating scenarios). This dataset seeks to fill the gap for a standardized surrogate benchmark in the field, where varying simulations, splits, and metrics hinder method comparisons. Accurate predictions of fatigue damage are essential for certification, design refinement, and cost efficiency as the industry progresses towards 22 MW class models. The research is available on arXiv (2605.25717).
Key facts
- FLOATBench is a public tabular benchmark for floating offshore wind turbine tower fatigue.
- It includes 582,120 per-section fatigue-damage labels.
- Data covers three 22 MW FOWT tower geometries.
- Derived from 19,404 high-fidelity OpenFAST simulations.
- Each tower has 6,468 simulations: 1,078 aligned wind/wave operating conditions.
- Addresses lack of shared surrogate benchmark in the field.
- Accurate fatigue-damage prediction is critical for certification, design optimization, and cost reduction.
- Published on arXiv with ID 2605.25717.
Entities
Institutions
- arXiv