Sampling-based integrated design optimization of floating wind turbine substructures with Pywind

Design optimization of floating wind turbine (FWT) substructures remains computationally demanding due to the extensive time-domain simulations required to explore high-dimensional design spaces. This study builds upon the previously developed Pywind-based integrated design optimization (IDO) framework, extending it by comparing two sampling strategies: Latin Hypercube Sampling (LHS) and Maximin LHS. A 15 MW semi-submersible FWT is employed to assess the performance of these methods in capturing design variability and coupled hydrodynamic and structural responses under stochastic environmental conditions. The analysis shows that Maximin LHS significantly improves space-filling uniformity, reducing the Voronoi coefficient of variation by 41.5% compared to Standard LHS. Additionally, Maximin LHS shows a 46.8% improvement in the mean Voronoi area, resulting in more even parameter-space coverage. This highlights the capability of the Pywind-based IDO framework to conduct efficient, scalable, and statistically robust time-domain optimization. The findings underscore the IDO framework's potential for probabilistic design exploration, supporting intelligent decision-making for the next generation of FWT systems.