The structural response and thus the service life of offshore wind turbines is significantly affected by their first natural frequency. However, field measurements from offshore wind farms show that the field response of monopile-founded wind turbines is stiffer than expected, suggesting that improved modelling could lead to optimized cost-efficient structures.

This discrepancy may be associated to conservative foundation modelling. The current industry-standard finite element (FE) approach assumes monopiles as wished-in-place (WIP), thus neglecting installation effects. The goal of this thesis is to calibrate a 3D FE model to better capture the lateral response of impact and vibratory-driven monopiles in predominantly sandy soils under monotonic loading, particularly in the small-strain range that governs the dynamic response of offshore wind turbines, by accounting for installation effects.

PLAXIS 3D is used to model the soil-structure interaction with Hardening Soil elastic-plastic constitutive model that can capture small-strain stiffness (HSsmall). The study has a twofold scope: namely, establishing an interpretation scheme for initial soil properties to model the WIP response and incorporating installation effects through a practical approach that captures the effects of the installation on the soil state, and consequently the lateral capacity, without explicitly simulating the pile installation.

The WIP models, validated against a number of field tests, show that the current modelling approach can accurately predict the lateral response of vibratory driven monopiles, while it underestimates the stiffness of impact driven ones. Therefore, this thesis proposes to artificially incorporate installation effects into the established WIP FE models for impact driven monopiles, by either imposing volumetric strains to the soil plug or by modifying the coefficient of lateral earth pressure at rest. Calibration of these installation parameters is performed against the global monopile response and the post-installation horizontal stress profiles. Both methods lead to increased horizontal stresses around the monopile and result in stiffer global response and improved agreement with the field data.

This thesis offers a comprehensive framework for modelling the lateral response of monotonically loaded monopiles, including installation effects, that could potentially be adopted by industry thanks to its simplicity, computational efficiency and reliance on commonly available data in offshore wind projects.

The rapid expansion of offshore wind energy is increasingly pushing projects into seismically active regions, necessitating monopile foundation designs that account not only for typical environmental loads but also for seismic displacement demands. To accurately assess the dynamic response of an offshore wind turbine (OWT) and design its monopile foundation accordingly, it is essential to model the soil-monopile-superstructure interaction with high fidelity. Three-dimensional (3D) finite element (FE) analyses represent the most advanced numerical approach for capturing this complex behaviour. To address this need, this thesis develops a comprehensive FE framework in PLAXIS 3D that integrates free-field site response and soil-structure interaction (SSI) under both linear elastic and nonlinear elastoplastic soil to evaluate the seismic performance of monopile-supported OWTs.
A structured four-step approach is adopted. Step 1 performs one-dimensional (1D), two-dimensional (2D), and 3D site response analyses (SRA) in linear elastic soil, validating the numerical setup against analytical solutions. Step 2 introduces the pile and superstructure, modelled with linear elastic material, in the same medium to assess the dynamic characteristics and their response under steady-state monoharmonic excitation, allowing verification of the SSI model against benchmarks from the literature. To capture more realistic soil behaviour, the framework then incorporates the SANISAND-MS constitutive model, which accounts for nonlinear cyclic sand response, including strain accumulation and stiffness degradation. In Step 3, SRAs with displacement input motions of varying amplitude are performed under both broadband Ormsby and steady-state monoharmonic excitation to evaluate the influence of nonlinearity on free-field response. Finally, Step 4 couples the nonlinear soil with the structure to examine the fully integrated dynamic characteristics under nonlinear SSI conditions.
Overall, the analyses establish best practice boundary conditions and numerical setups for seismic SRA and SSI modelling in PLAXIS 3D for linear elastic soil (Steps 1 and 2), and quantify the transition from inertial to kinematic dominance in the i response of a flexible pile (Step 2). In the nonlinear domain, the results from Step 3 highlight the critical role of loading type in capturing key features of nonlinear free-field response. More specifically, with steady-state monoharmonic excitation of increasing amplitude, the transition from linear to nonlinear behaviour is captured, along with a shift of predominant frequencies to lower values compared to the linear elastic SRA. In contrast, when the same procedure is applied with the broadband impulse load Ormsby wavelet, post-impulse stiffening effects are revealed, expressed as an upshift in resonance frequency. In Step 4, the dynamic characteristics of the monopile-superstructure system are assessed under nonlinear SSI conditions. However, computational cost prevents the execution of full amplitude sweeps, underscoring the practical trade-off between accuracy and run time in high-fidelity 3D nonlinear SSI modelling. Instead, the 1995 Kobe earthquake record is employed to assess the system’s behaviour under real, multi-harmonic excitation. Altogether, the work demonstrates the potential of advanced 3D FE tools to enhance the seismic design of monopiles.

Understanding the dynamic behaviour of soil is crucial for the design of geotechnical structures, particularly for offshore wind turbine foundations. This study evaluates the performance of a nonlinear numerical model in accurately capturing soil response under dynamic shear loading, comparing its simulated response to experimental data obtained from Resonant Column Tests (RCT). The research systematically investigates the role of higher-order odd harmonics in the numerical solution, assesses the differences between linear and nonlinear models, and validates the nonlinear model against experimental results.

The findings reveal that higher-order harmonics contribute negligibly to the numerical response, simplifying the model's implementation. Furthermore, the nonlinear model effectively captures the resonance shift observed in experimental data, outperforming the linear model, particularly at higher strain levels where soil softening effects become significant. While the model demonstrates a strong correlation with experimental results, discrepancies arise in high-strain scenarios, primarily due to limitations in damping representation. To address this, corrective approaches such as a damping correction factor and strain radius adjustments were explored, improving model accuracy but failing to eliminate errors entirely. The study highlights the necessity of incorporating a strain-dependent damping formulation to enhance predictive capabilities for high-strain soil behaviour.

The results confirm that nonlinear numerical modelling is a valuable tool for capturing essential soil dynamics, but further refinements in damping representation are required to improve alignment with experimental data. Future research should focus on advanced damping models to ensure greater accuracy in high-strain dynamic soil simulations.

The renewable energy sector is experiencing rapid growth, with offshore wind energy gaining significant global importance. Designing bottom-fixed offshore wind turbines (OWTs) with monopile (MP) foundations in seismic-prone regions, particularly in coarse-grained soils, presents challenges due to the risk of soil liquefaction during earthquake excitations. Current design practices for soil-structure interaction modelling address seismic pore pressure effects by accounting for the net decrease in soil shear modulus resulting from the development of excess pore water pressure.
This thesis examines the cyclic contour diagrams framework (CDF) for its capacity to predict excess pore water pressure (EPP) build-up in coarse-grained soils under seismic conditions. The framework is assessed by employing PM4Sand in PLAXIS2D to produce the cyclic contour diagrams for a coarse-grained material. This material is then used to perform site response analysis (SRA) on a soil column in PLAXIS2D. Subsequently, a total stress SRA is conducted using DEEPSOIL V7.0 to obtain the equivalent cyclic stress ratio (CSR) and the number of cycles by employing a best-match soil material similar to the one utilized in the PLAXIS analysis. By analyzing the cyclic stress history obtained from the SRA, the equivalent number of cycles and cyclic stress ratios are applied to the material's cyclic contour diagrams. Finally, the predictive capabilities of the method to predict the EPP calculated in the PLAXIS model are assessed.
Two case studies illustrate the framework's practical application. The first case focuses on Akita Port, Japan, which experienced significant damage during the 1983 Nihonkai-Chubu earthquake. The CDF was able to accurately predict liquefaction severity, validating the method. The second case study involves an offshore wind farm in the Netherlands, examining how liquefaction impact monopile foundation designs. The study highlights the importance of incorporating soil degradation because of liquefaction into the design of offshore wind turbines.
The findings indicate that the CDF is a viable tool for evaluating liquefaction potential in coarse-grained soils under seismic loading in engineering practice.