Gradient-based design optimization of semi-submersible floating wind turbines

A wide variety of semi-submersible floating wind turbine designs have been proposed, though few studies have attempted design optimization with more than a handful of design variables. Gradient-based design optimization can enable optimization with many design variables to define the tower, substructure, and structural scantling design and allow for the inclusion of increasingly realistic design constraints based on ultimate and fatigue limit states of the structure. A surrogate-assisted optimization model centered around a linearized frequency-domain aero-hydro-servo-elastic model of a semi-submersible wind turbine was developed and applied to find optimal designs in three hypothetical locations. The optimal designs vary significantly based on the environmental conditions, and show the potential for significant cost-savings by employing site-specific design for floating wind turbines. For all of the hypothetical locations considered here, the tower design is driven by fatigue damage in near-rated wind speed conditions at the tower base and top, and buckling utilization in the rated-wind speed extreme condition for the remaining tower sections. The substructure design is driven by fatigue damage in near-rated wind speed conditions for the central column and above-rated wind speed conditions for the pontoons. Rules-based buckling constraints in the rated-wind speed condition drive the outer column and scantling design. The optimization model has also been used to investigate the impact of the bounds on tower design variables on the integrated tower-substructure structural design, and has the potential to be adapted to help answer a range of other concept-level design questions.