Dynamic Finite Element Analysis of Impact Pile Driving Centrifuge Tests

More Info
expand_more

Abstract

Many farms of wind turbines are being installed worldwide as a response to clean energy targets. These farms are usually installed in offshore environments since limitations of space can be encountered or due to public pressure against the visual disruption that farms can cause in urban areas. These offshore wind turbines are commonly founded on monopiles, which have been traditionally installed in shallow waters. Nevertheless, current trends show preference towards turbines of greater capacity which are installed in deeper waters. This trend poses uncertainties and new challenges to the offshore industry since most of the gained experience lies on the installation of monopiles in shallow waters. The hydromechanical response of soil and soil-structure interaction during the installation is complex and far from being fully understood yet. Thus, research projects involving advanced physical and numerical modelling techniques could be proposed as a first step to gain a better insight in the impact pile driving technique and how different soil conditions can influence the installation procedure of such monopiles. The current study is part of a project lead by Royal IHC, which has been developed at Delft University of Technology.This numerical study aims to study the dynamic behaviour of saturated sand and pile response by means of numerical simulations of a series of centrifuge tests using the finite element method in the time domain. These experiments were going to be performed in open-ended piles using the centrifuge facility of Delft University of Technology as part of a separate research project. Special attention has been given to the prediction of boundary effects in the centrifuge experiments since limitations of space are present in the facility. Two types of drainage conditions have been considered, i.e. fully drained and fully undrained. Two relative densities have been chosen to cope with loose and dense sand response, i.e. RD = 30% to 90%, respectively. A hypoplastic model has been used to model the surrounding soil, while soil-structure interaction has been addressed using a zone of degraded material properties as a continuum.Numerical predictions have shown that boundary effects may be usually stronger in denser soil conditions. In addition, liquefaction has been predicted to take place in the surroundings of the pile when driven into loose sand. In this regard, the effects of an initial soil liquefaction triggering event have been predicted to not influence in a significant manner pile drivability. In general, boundary effects have been shown to be problem dependent, i.e. on the geometry, stress state and soil packing state. In addition, drainage conditions and the amount of degradation of strength and stiffness properties have shown to play a role on numerical predictions of impact pile driving response. In addition, a numerical validation has been performed by comparison with preliminary centrifuge test results on dry dense sand. Finally, it has been concluded that the use of lateral viscous boundaries may be a suitable strategy to mitigate boundary effects in impact pile driving centrifuge tests. A first estimation of the damping demand of these Soft Boundaries (SB) has been presented in this study. Future studies should aim in exploring different numerical techniques in the sake of improving the predicting capabilities, e.g. by considering particle based methods such as MPM to extend the current research findings within multiple-blow centrifuge tests and by incorporating more complexities within the simulation of the hammer impact.

Files

Delft_University_of_Technology... (.pdf)
(.pdf | 15.2 Mb)
- Embargo expired in 01-11-2019