Seismic response of monopile foundations for offshore wind turbines: from 3D to 1D modelling of soil-foundation interaction

Abstract

New offshore wind farms consisting of monopile-founded Offshore Wind Turbines (OWTs) are to be built in earthquake-prone areas. To design the monopile foundation and to accurately define the dynamic response of an OWT, the soil-monopile-superstructure interaction should be modelled properly. Due to the fact that the Three-Dimensional (3D) Finite Element (FE) analyses are complex and computationally expensive, research is focused on One-Dimensional (1D) FE models, in which the soil-monopile interaction is traditionally described via distributed translational springs representing the soil lateral load. Nevertheless, the increase of monopile diameter, followed by the monopile length-to-diameter ratio (L/D) decrease, implies the contribution of additional resistant components such as distributed moment, base shear and moment.

This thesis examines the 3D mechanisms to be accounted for in the 1D FE modelling of the soil-monopile-superstructure seismic response in case of a single-phased, linear visco-elastic soil layer. Both 3D and 1D FE analyses are conducted with the FE software OpenSees. The 1D analyses are simulated in two consecutive steps: first a Site Response Analysis is performed, next the recorded displacements over the soil layer depth are applied to the spring supports and the dynamic interaction of the system is simulated. Three different monopiles are considered with L/D equal to 26, 9 and 5. Two superstructures are examined, which are modelled as Single-Degree-of-Freedom systems. Distributed translational springs are assigned to the slender monopile (L/D=26), while for the stubbier monopiles the contribution of distributed rotational springs is examined as well. Lastly, the effect of considering the base moment and shear is also examined.

The stiffness of the soil reaction curves is calibrated by applying a monotonic lateral load and moment at the pile head, in case of the translational and rotational springs, respectively. The spring stiffness values are assumed uniform along the monopile length. As a next step, the dynamic response of the calibrated 1D models is examined in steady-state conditions, under the action of mono-harmonic excitation, and compared to the 3D results. Ultimately, the seismic response of the 1D models is examined in case of two earthquake excitations with different frequency contents.

In case of the monopiles with L/D=9 and 5, it is concluded that the use of monotonically-calibrated distributed translational and rotational springs provides a good match between 3D and 1D regarding the monopile head and superstructure response under seismic loading. Nevertheless, these 1D FE models cannot predict the base moment, for which a base rotational spring should be employed. In case of the stubbier monopile, with L/D=5, the base shear seems to positively affect the moment profile as well. Lastly, regarding the monopile with L/D=26, the employment of translational springs alone seems sufficient for the accurate prediction of the seismic response; however, the hereby monotonically-calibrated distributed translational springs result in a mismatch between 3D and 1D.