Behaviour of Axially Loaded Bucket Foundations in Sand
Author
Grecu, Sorin
Term
4. term
Education
Publication year
2018
Submitted on
2018-06-08
Pages
311
Abstract
Havvind flytter længere ud på havet for at sænke omkostningerne, men det giver nye økonomiske og tekniske udfordringer. Fundamenter til møller er en af dem. Sugefundamenter (suction buckets) kan reducere installationsomkostningerne og bruges bl.a. under gitterkonstruktioner (jackets) til store møller i overgangsdybder, hvor belastningen primært er aksial (op/ned). Der mangler dog lettilgængelige metoder til at forudsige deres opførsel. Denne afhandling udvikler t-z-kurver for sugefundamenter i sand ved hjælp af numerisk modellering. T-z-kurver beskriver sammenhængen mellem forskydningsmodstand (t) langs skørtets kontaktflader og den lodrette forskydning (z) af spanden. Med PLAXIS 2D og Python opbygges 100 aksialsymmetriske finit-elementmodeller. Der påføres kontrollerede lodrette forskydninger i begge retninger under både drænede (vand kan slippe ud) og delvist drænede forhold. De simulerede spande har diametre på 10–20 m og et indspunsningsforhold på 1. Jorden antages at være Frederikshavn-sand (kohesionsfri) beskrevet med Hardening Soil small strain-modellen, og der indgår kompaktionstilstande fra løs til meget tæt. Forskydningsspændinger langs skørtets grænseflader plottes mod spandens forskydning og normaliseres i forhold til maksimal forskydningsspænding og maksimal forskydning. Maksimalværdierne bestemmes ved regressionsanalyse for hver kombination af belastningsretning og dræningsforhold. På basis af de normaliserede kurver opstilles en matematisk model for hver af de fire sæt. Ved at bruge Winkler-metoden, hvor jordens reaktion idealiseres som fjedre, anvendes t-z-formuleringerne til at beregne den samlede friktionsrespons af sugefundamenterne. Fjederstivhederne afhænger af lodretspænding, spanddiameter, skørtlængde og friktionsvinkel – alle standardparametre i geoteknisk design. Verifikation af modellerne peger på visse begrænsninger og viser, at de drænede formuleringer klarer sig bedst. Afhandlingen afslutter med forslag til forbedringer af t-z-udtryk og anbefalinger til videre forskning.
Offshore wind is moving farther from shore to improve cost efficiency, which brings new economic and technical challenges. Turbine foundations are a key issue. Suction bucket foundations can cut installation costs and are used under multi-legged jacket structures for large turbines in transitional water depths, where the buckets are mainly loaded axially (up and down). However, there is limited guidance on how to predict their performance. This thesis develops t-z curves for suction buckets in sand using numerical modeling. T-z curves describe the relationship between shear resistance (t) along the bucket skirt interface and vertical displacement (z). Using PLAXIS 2D and Python, 100 axisymmetric finite element models are built. Prescribed vertical displacements in both directions are applied under drained (water can escape) and partially drained conditions. The modeled buckets have diameters of 10–20 m and an embedment ratio of 1. The soil is cohesionless Frederikshavn sand represented by the Hardening Soil small strain model, with density states from loose to very dense. Shear stresses along the skirt interfaces are plotted against bucket displacement and normalized by peak shear stress and peak displacement. These peak values are obtained by regression analysis for each combination of loading direction and drainage condition. Based on the normalized curves and fitted parameters, a mathematical model is created for each of the four cases. Using the Winkler method, which idealizes soil reaction as springs, the t-z formulations are employed to calculate the total frictional response of suction buckets. The spring properties depend on vertical stress, bucket diameter, skirt length, and friction angle—standard parameters in geotechnical design. Model verification highlights limitations and shows better performance for the drained formulations. The thesis concludes with suggestions to improve the t-z expressions and directions for further research.
[This abstract was generated with the help of AI]
Documents