Two-stage local optimization of lattice type support structures for offshore wind turbines

• An optimization method for offshore wind turbine support structures is presented.
• In each iteration a number of independent one-dimensional problems is solved.
• Fatigue utilization was maximized, thereby minimizing structural weight.
• Different initial designs resulted in more or less the same optimized design.
• The method converged using less than seven structural analyses.

Offshore wind turbines are exposed to stochastic dynamic loading that includes non-linear aerodynamic effects. Hence, time domain simulations are needed for the analysis of support structures, resulting in a time consuming and computationally demanding optimization process. Here, the structural optimization was performed assuming that changing dimensions of a structural member does not affect other members of the structure, a heuristic based on the principle of domain decomposition. This allows for a simple sizing algorithm, which optimizes area and stress concentration factors element-wise using a meta-model for the structural performance. In order to further speed up the algorithm, designs were evaluated based on performance data from earlier analyses, while searching for the most suitable changes in design variables. This was found to be very efficient, resulting in a nearly full utilization of fatigue resistance and a saving of analysis time of roughly 40%. Randomly chosen initial designs initiated sequences that converged to the same minimum weight using fewer than seven structural analyses. Fatigue damage was the design driver in this analysis, while ultimate and buckling performance were controlled within the given constraints.