Buckling analysis of suction buckets: Influence of uncertainty in imperfections and soil parameters

Abstract

Suction buckets, also known as suction caissons, in combination with jacket substructures have become a prominent foundation concept for offshore wind turbines. Suction buckets are large cylindrical shell structures with relatively thin walls, that are installed by pumping out the water inside the structure, which creates a pressure difference and thus a downward force. The accom- panying circumferential compression on the cylinder wall causes a risk of shell buckling. The prediction of the buckling capacity of such large cylindrical shells is challenging, since it depends significantly on the initial geometric imperfec- tions resulting from the manufacturing process and on the boundary conditions imposed by the surrounding soil. Previous work on suction buckets revealed, that the choice of a representative imperfection form and amplitude is very chal- lenging, and that the calculated buckling pressure is sensitive to soil modeling choices. While it has not yet been possible to identify a generally applicable im- perfection form, probabilistic design approaches based on realistic imperfections were also not yet considered for suction buckets. In this work, a general understanding of the nonlinear buckling behavior of unstiffened cylindrical shells under circumferential compression loading is elab- orated by applying various imperfection forms and amplitudes to geometrically and materially nonlinear finite element models. Further, different soil param- eters and soil modeling approaches, including volumetric models and nonlin- ear soil springs, are applied and their influence on the buckling capacities are determined. The effects of global and local imperfection patterns in form of sinusoidal eigenmodes and weld depressions are evaluated and compared. The results suggest, that global imperfections are more detrimental than local ones. Consequently, a new approach based on sorting eigenmode-affine imperfections depending on the circumferential wave number to identify the most unfavorable imperfection pattern is developed. As a main novelty, stochastic imperfection modeling approaches based on measured data are further developed and the associated buckling behavior is compared to eigenmode-affine imperfections. Additionally, this thesis is the first to show that the inclusion of volumetric soil plasticity and contact effects has a significant impact on the buckling pressure. In contrast, it is demonstrated that soil springs, even when specifically adapted for this load case, are not able to capture the three-dimensional soil interaction due to localized deformations.