Offshore wind energy has become a crucial element of the global energy transition, with the North Sea being a major hub for offshore wind farms. As first-generation farms approach the end of their operational life, decommissioning offshore wind export cables has emerged as a significant technical challenge. This thesis focuses on the feasibility and limitations of using the cable pullout method for decommissioning offshore wind export cables, particularly examining the influence of burial depth and soil conditions.

This research is structured in two main phases. The first phase involves a comprehensive literature review to identify critical knowledge gaps and establish an overview and understanding of existing decommissioning practices. The second phase employs simulations using OrcaFlex software to model and analyze various cable pullout scenarios. These simulations focus on determining the limits and constraints imposed by burial depth and soil relative density for sandy soils commonly found in the North Sea.

Key findings from the literature study underscore the complexities of decommissioning, which encompass legal, environmental, economic, and technical considerations. One of the main conclusions is the importance of adopting a 'design for decommissioning' approach during the initial cable installation. This involves designing cables and selecting burial depths that facilitate easier future removal, thus promoting sustainability and cost-effectiveness. Additionally, this approach can help mitigate potential environmental impacts and regulatory challenges associated with cable removal.

Soil modelling plays a crucial role in understanding how burial depth and soil conditions influence resistance forces during cable pullout. Factors such as shear strength and burial depth are analyzed to determine the forces that oppose cable recovery. The model includes scenarios for fully drained, fully undrained, and partially drained uplift resistance, implemented in a Python script to simulate real-time resistance during pullout operations. To address buried cables, an external soil model is integrated into OrcaFlex, allowing dynamic simulations of soil resistance during cable pullout.

The simulation results with a 525 kV HVDC export cable reveal how soil resistance significantly increases with greater burial depth and higher soil density. These findings highlight the critical importance of burial and soil conditions in planning decommissioning operations and suggest that additional deburial techniques may be necessary when cable pullout is not feasible.

This thesis provides valuable insights and recommendations for future research and practical applications, aiming to support the offshore wind industry's evolving needs and enhance the sustainability of decommissioning processes. These findings are crucial as the industry anticipates a substantial increase in decommissioning activities, with offshore wind capacity expected to continue to grow.

Offshore wind energy is considered one of the most promising solutions for sustainably meeting the society's growing energy needs. This is why there are major construction projects and massive expansion plans for new offshore wind farms worldwide. In the North Sea in particular, the technology has been used to generate electricity for several years. Looking at the oldest wind farms still in operation, it will soon be time to decommission them once the turbines have reached their service life of 20 to 25 years. However, especially the removal of the foundations is considered challenging because not many comparable projects have already been realised. The most commonly used type of foundation is the monopile due to its simplicity. During installation, these long steel piles are driven deep into the seabed to provide a secure foundation over the entire service life. One option for decommissioning is to cut the piles from the inside a few metres below the seabed and then pull them out. So far, this procedure has mainly been used for individual small piles. Scaling up to several large monopiles of a wind farm involves uncertainties. On the one hand, it is unclear how larger monopiles will behave during the cutting process, and on the other hand, the operational performance of several piles that are removed one after the other is uncertain.

When cutting monopiles, theoretically no complete cutting progress can be achieved because the pile breaks off beforehand. In this work, a calculation model was developed that predicts the cutting progress at which a monopile fails due to the hydrodynamic forces acting on it. Based on an existing project in which the internal cutting technology was used, a minimum progress was defined that must be reached before failure is allowed to occur. This makes it possible to determine which sea states are permissible in order to achieve the specified cutting progress. With the calculation model, a parameter study was carried out to find out which structural and environmental parameters have the greatest influence on failure. Additionally, two real wind farms were considered as case studies to investigate realistic parameter sets. One of these wind farms has relatively small monopiles as foundations, while the other has much larger ones. An operability analysis was also carried out for these case studies in order to determine the duration of the foundation decommissioning operation and to be able to analyse the weather downtime. In addition to calculating when a monopile fails during the cutting process, the forces required to pull the pile out of the seabed afterwards were calculated. This allows to determine the required crane capacity of the working vessels used for the operation.

The results of the calculations show that larger monopiles fail at an earlier cutting process than smaller monopiles. However, this difference is not large, which is due to the fact that together with the hydrodynamic forces, which increase with the pile size and length, also the wall thicknesses increase. This means that larger forces can generally be withstood. The operability is good for both wind farms from the case studies. Furthermore, it was found that it is important for the operability how the connection between monopile and transition piece was realised and whether ROVs have to be used. The crane capacity of the vessels used can be lower than that of the vessels used during installation. The additional force required to pull the monopiles out of the seabed does not make up for the weight lost by leaving a large part of the foundation in the soil.

The offshore wind energy market is expanding and the number of offshore wind turbines being installed in the near future is rising. Offshore wind turbines are being installed further offshore and in deeper waters. Besides, to lower the cost of wind energy offshore wind turbines are increasing in size and power output. Both wind turbines and their support structures are expected to keep increasing in size and weight in the coming years. After fabrication, wind turbine components and support structures have to be transported to onshore storage depots or to their offshore location. To enable safe transports, wind turbine components and support structures are constraint to heavy transport vessels or transport barges by sea fastening structures. As a result of wind turbine components and support structures increasing in size and weight, the reaction loads for which sea fastening structures need to be designed are increasing as well. Since increasing reaction loads have various negative consequences which are expected to become more critical for future transports, there is a need for an optimized reaction load calculation.

The current method of calculating the reaction loads which is widely used in the industry is often referred to as being a conservative method. The aim of this thesis was to enhance the existing calculation method of the reaction loads by shifting from a conservative approach towards a method of calculating reaction loads based on an acceptable probability of occurrence during transports. By calculating the reaction loads for an acceptable probability of occurrence it is avoided that sea fastening structures are designed for overly conservative reaction loads while the structural reliability of these structures will still be ensured.

In this thesis an existing sea fastening design project from the industry was used to perform a case study. Data and information from this project were used as input to perform motion analyses of a vessel which is transporting a jacket support structure. The obtained linear wave-induced accelerations of the jacket CoG were used as the main input for calculating the reaction loads. It was first investigated how these 6-DoF accelerations of the CoG are used in the current method of calculating the reaction loads. This was followed by introducing statistical extreme value theory with the purpose of using the accelerations of the jacket CoG as input for a probabilistic method of calculating the reaction loads.

The findings of this research show that optimized reaction loads can be obtained by replacing the current calculation method by a long-term probabilistic method. It was found that this long-term probabilistic method could be derived by combining 3-hour extreme value density functions of reaction loads with the probabilities of encountering the various sea states at the location on the route for which the most severe environmental conditions are expected. The long-term probabilistic method was used to perform a probabilistic investigation of the reaction loads calculated with the current calculation method. It was found that the return periods of the reaction loads calculated with the current method were significantly different for the individual jacket legs. Moreover, it was found that the sea fastening design for at least one of the jacket legs was expected to be over-conservative. By presenting the long-term probabilistic calculation method, a methodology was introduced which determines reaction loads based on acceptable return periods while avoiding over-conservative sea fastening designs.

This research has provided a new insight into the method of designing sea fastening structures. The long-term probabilistic calculation method can be applied in practice to determine optimized reaction loads incorporated in sea fastening designs. This research therefore makes a valuable contribution to preparing the reaction load calculation method for future transports which are expected to become more critical due to wind turbine components and their support structures growing in size and weight.

Bayesian system identification, including parameter estimation and model selection, is widely used to infer partially known, unobservable parameters of the models of physical systems when measurement data is available. A common assumption in the Bayesian system identification literature is that the discrepancy between model predictions and measurements can be described as independent, identically distributed realizations from a univariate Gaussian distribution. However, the decreasing cost of sensors and monitoring systems leads to more frequent structural measurements in close proximity to each other (e.g. fiber optics and strain gauges). In such cases, dependency in modeling uncertainty could be significant, both in space and time, and the assumption of uncorrelated Gaussian error may lead to inaccurate parameter estimation.

The aim of this thesis is to explore how Bayesian system identification can be feasibly performed using large datasets when spatial and/or temporal dependence might be present and to assess the impact of considering this dependence. A pool of models, each assuming a different correlation structure, is defined and Bayesian inference is performed. In particular, stress measurements obtained on a steel road bridge are used to update the parameters of the corresponding FE model and the parameters of the correlation structure. The results are compared to a reference model where only measurements of the response peaks are used under the assumption of independence. Nested sampling is utilized to compute the evidence under each model and Bayesian model selection is applied. The question of efficiently performing system identification for large datasets (N > 10² for temporal dependencies and N > 10³ for combined spatial and temporal dependencies) is investigated, and a novel approach for efficiently calculating the exact log-likelihood is derived. An approximation based on the Fisher information matrix is used to efficiently calculate the information content of measurements.

It is found that the choice of correlation function can significantly affect the posterior distribution of the model prediction uncertainty. Additionally, it is shown that using large datasets and considering dependence makes it possible to perform system identification for a larger number of parameters compared to the reference model. The results of the case study indicate that using measurements from multiple sensors under combined spatial and temporal dependence and additive model prediction error yields reduced uncertainty in the posterior and up to 29% reduction of the posterior predictive credible interval range compared to the reference case. Furthermore, the efficiency of the proposed likelihood evaluation method is assessed. Using this method, exact calculation of the log-likelihood can be performed for >10⁶ points in under a second in the case of correlation in one dimension. For combined spatial and temporal correlation it is shown to be approximately 900 times faster than naive evaluation for a 64 by 64 grid of observations. The results of the case study indicate that the described approach can be feasibly applied to real-world structures and can potentially improve parameter estimation and reduce prediction uncertainty. These findings suggest that further research into the approach could yield improvements over current methods.