PSO-based design and optimization of jacket substructures for offshore wind turbines

Abstract

The search for more and better wind resources pushes offshore wind farms further into deeper waters, where jacket support structures become a competitive alternative to monopiles for bottom-fixed offshore wind turbines. In this context, this paper proposes a cost-effective methodology for the autonomous design that facilitates the generation of preliminary candidate designs of jacket substructures, reaching a level of detail suitable for the initial design phase. A Particle Swarm Optimization algorithm, with precomputed initial swarms, is used as a search and optimization tool. The proposed strategy is able to find candidate designs that satisfy, with a minimum use of material, a wide range of Ultimate Limit States, Fatigue Limit States and Geometrical Restrictions. It is shown that the use of a precomputed initial swarm generated taking into account a starting concept design of the structures with a low computational cost, instead of a standard random population, significantly improves the efficacy and efficiency of the algorithm. A reference case, based on the NREL-5MW wind turbine and its OC4 reference jacket support structure, is used for studying and illustrating the capabilities of the proposal.