MSc defence Marco Guarnero

Supervisors: Mathias Stolpe, Alexander Verbart & Susana Rojas Labanda (DTU Wind Energy), Martin Kühn & Marjin Floris van Dooren (University of Oldenburg)

External examiner: Uffe Dal Eriksen, Rambøll

Abstract: Topology optimization of offshore wind turbine support structures
In this thesis the author identified in offshore wind turbines support structures an interesting area of cost reduction, which can be achieved through structural optimization. Structural optimization's aim is to find the optimal lay-out of a structure in order to carry the load in the best way. A two-steps method is proposed. The first step deals with topology optimization of ground structures, which concerns finding the optimal design starting from discrete structures consisting in a potentially high number of members by removal of unnecessary members. Classical minimum compliance topology optimization problems are solved using Solid Isotropic Material with Penalisation (SIMP) method, for several load cases. It is observed that topology optimized designs are not suitable for industry, because they may fail under the applied load or they may not satisfy structural requirements and post-processing is required. The second step concerns sizing optimization of the topology optimized designs under strength, buckling and eigenfrequency constraints. In the last part, a case study was carried out for the INNWIND reference jacket supporting the 10 MW DTU reference wind turbine. Results show that 50% mass reduction can be achieved through sizing optimization. The optimized INNWIND reference jacket is then compared to realistic designs obtained with the two-steps method. In general, it is observed that topology optimized structures are initially infeasible with respect to stress constraints and buckling, and performing sizing optimization of these structures results in a mass increment that matches the final optimal  mass of the INNWIND reference jacket after sizing optimization. Finally, three and four legged innovative designs obtained with this method present internal connectivities and might be taken into consideration for further analysis, which will include fatigue analysis and dynamic loading condition.



Mon 14 Aug 17
10:30 - 12:00


DTU Vindenergi


Technical University of Denmark
DTU Risø Campus, Building 118, Tycho Brahe
Frederiksborgvej 399, 4000 Roskilde
18 AUGUST 2017