This research focused on applying vibration-based model updating to estimate the structural properties of high-rise buildings using a discrete Timoshenko beam model. Previous studies by Moretti et al. [1] and Taciroglu et al. [2] showed limitations in estimating the structural properties of buildings using model updating with a uniform Euler-Bernoulli beam model or a uniform Timoshenko beam model. Both models were not able to describe the third measured bending mode accurately after updating. Moreover, with the uniform Euler-Bernoulli beam model, when only the lowest two bending modes were used to fit the model, the estimates of the rotational stiffness of the foundation around the y-axis, Kr,y, showed significant uncertainty after model updating, with a Coefficient of Variation (CV) of 46.9%. Therefore, this research aimed to improve the accuracy of structural property estimations for high-rise buildings by using a more detailed model.
The model updating method applied in this research was an indirect vibration-based technique, which adjusted the input parameters of the chosen model to minimize the difference between the model output and the measured data. Both the model output and the measured data were in the shape of natural frequencies and mode shapes. The technique was applied to two high-rise buildings: the residential tower New Orleans, which is bending-dominant in behavior, and the office tower Delftse Poort, which is shear-dominant in behavior and exhibits irregular stiffness across its height. The discrete Timoshenko beam models used to approximate the dynamic behavior of these buildings were created using the Finite Element (FE) models of the buildings.
The research findings highlighted several key aspects essential for obtaining more accurate estimations of the structural properties of high-rise buildings. For a more accurate estimation of parameter Kr,y, it is of importance to incorporate shear deformations in the model and to account for irregular stiffness along the height. For the New Orleans, using the discrete Timoshenko beam model led to estimates of parameter Kr,y with low uncertainty, indicated by a Coefficient of Variation of 6.91%. This was an improvement compared to the study by Moretti et al. [1], which showed a Coefficient of Variation of 46.9% using the uniform Euler-Bernoulli model.
Moreover, obtaining accurate values for the bending stiffness was crucial for achieving more precise estimations of the structural properties. It was challenging to determine the bending stiffness values using the FE models. These challenges posed a problem for model updating, as the initial structural property ratios were maintained for both high-rise buildings. Maintaining these ratios gives weight to the initial structural property values. These values must be correct. Otherwise, model updating is limited by the incorrect ratios and will not be able to accurately match the measured modal properties.
For the Delftse Poort, more pure bending modes were needed for better accuracy. The discrete Timoshenko beam model was unable to match the measured second bending mode in the y-direction, as it exhibited twisting, which the model could not represent due to its limitation to pure bending modes. Furthermore, since the discrete Timoshenko beam model requires a single displacement value per height but multiple sensors were used to measure displacements, the measurements had to be averaged. This averaging process may led to a mode shape that deviates from the actual behavior, resulting in an inaccurate representation of reality.

This research contributes to a novel non-contact installation method for offshore wind turbines, utilizing electromagnetism for positional control. Floating wind installation requires efficient control methods for payload positioning. Current methods use tuggerlines that are physically attached and work through tensile attractive forces. In contrast, electromagnetic control offers a non-contact solution and allows for control through both attractive and repulsive forces. This research explores the applicability of multiple data-driven control algorithms on a magnetically controlled pendulum system. These algorithms are evaluated for both positional control and motion attenuation under pivot-point excitation, using simulations of a numerical model of the electromagnetic pendulum. The model-free control algorithms considered in this study assumed that changes in the system’s state are solely due to the control output. These methods aimed to optimize the control output based on either the state error or estimates of the state gradient with respect to the control output. However, this assumption was found to be valid only for systems dominated by the control response with limited dynamic behaviour over time, which was not the case for the electromagnetic pendulum system.

The study culminates in the development of a controller based on the online estimation of the magnetic interaction force. This controller, named Proportional-Derivative Force estimating neural network controller (PD-FeNN), is a neural-network based state-dependent PD-controller. The neural network is trained during an operational learning phase on estimations of the magnetic interaction force, which are derived using the model of a linear undamped pendulum. By incorporating the neural network, this approach eliminates the requirement of the previous state-of-the-art controller to manual model the magnetic interaction force based on experimental data. The PD-FeNN controller is tested on both numerical and physical models of the electromagnetically controlled pendulum. The results demonstrate efficient control across a wide range of excitation frequencies and amplitudes for both motion attenuation and positional control. As a successor to the modified PD controller, the PD-FeNN controller improves upon its predecessor by enabling positional control without requiring a model of the magnetic interaction. This advancement enhances the applicability of non-contact motion control for payloads in the offshore wind industry.

This thesis investigates the estimation of period lengthening (T˜/T) in structures using model updating, which assesses frequency shifts due to changes in boundary conditions, such as flexible bases. A 2D finite element (FE) model based on Timoshenko beam elements was developed and optimized through Sequential Least Squares Quadratic Programming (SLSQP) to align simulated dynamics with measured natural frequencies and mode shapes of a steel structure.

The research was divided into two phases. Phase one utilized synthetic modal data to validate the model updating process in an error-free environment, showing that the algorithm effectively resolves stiffness-to-mass ratios rather than individual properties. Period lengthening predictions proved more accurate and stable with stiff springs, while soft springs demonstrated higher sensitivity and discrepancy. Phase two applied real measurement data, yielding results consistent with synthetic data but revealing a “plateau” in the cost function, where optimal parameter determination was challenging due to model and measurement uncertainties.

The findings indicate that model updating is a feasible method for estimating period lengthening, especially for flexible foundations. However, results are sensitive to modeling and measurement uncertainties, necessitating careful evaluation to ensure optimization convergence and parameter precision. The study recommends further research to mitigate uncertainties and enhance the method’s applicability in high-rise building design.

This thesis aims to investigate the feasibility of developing a successful unsupervised Structural Health Monitoring (SHM) methodology to detect damage in structures, specifically bridges. Detecting damage, especially in its earliest stages, is challenging, thus prompting the need for robust and effective methods. The success of such a methodology could lead to timely inspections and interventions, resulting in significant economic benefits and preventing further damage, including potential failure of the structure in use.

The approach involves a literature review to establish relevant background knowledge and useful concepts. From this, a methodology is developed utilizing unsupervised machine learning, specifically Gaussian Mixture Models (GMM), to identify abnormal behavior indicative of structural damage.

A Finite Element Method (FEM) model of a simple bridge is created and monitored over a three-year period, serving as a testing ground for the methodology and a primary source for data generation. Temperature data and its effects on the natural frequencies of the bridge model are used to establish a baseline for normal or healthy behavior. Synthetic damage, such as settlement and stiffness reduction, is then introduced to the model to create anomalies or abnormal behavior. The developed methodology is tested using three case studies, each with varying types of synthetic damage. By using both the healthy and unhealthy data generated from the model, the healthy behavior of the bridge is captured using GMM. The model then progressively incorporates unhealthy data into the proposed anomaly detection algorithm. The algorithm evaluates the likelihood of each incoming data point of belonging within the healthy distribution, resulting in data points being classified as either healthy or flagged as abnormal.

The case studies presented in this research underscore the efficacy of the proposed anomaly detection approach. In scenarios involving sudden or abrupt damage, the algorithm swiftly and accurately labels abnormal points. For gradual damage scenarios, such as settlement, the algorithm consistently identifies abnormal points, with the rate of abnormal point detection accelerating over time. This detection rate is contrasted with the rate of erroneous abnormal point labeling when processing an exclusively healthy data set through the anomaly detection algorithm. This comparison reveals a higher rate of abnormal point identification when actual damage is present, affirming the effectiveness of the unsupervised SHM methodology in pinpointing abnormal behavior within the modeled bridge structure.

Due to the increasing size of offshore wind turbines, the jack-up vessels and cranes installing them need to increase in size. The increase in length of the crane boom results in an increase of the dynamic response of the crane. This increase in response results in higher loads in the crane. The magnitude of the higher loads are unknown and not yet accounted for in crane design. To get a better insight in the dynamic response of the crane and higher loads, a probabilistic model is developed to quantify the annual probability of the crane.
To define the different crane operations, a measurement data set of crane operations is made available. From this data set, the different lifts types, exposure times and occurrences are identified. Multiple jack-up configurations are chosen with different water depth and wave heading. Next to the crane and jack-up configuration, long-term wave statistics are used for the environmental conditions, as input for the probabilistic model.
Time-domain simulations are performed in a rigid multi-body analysis software, to simulate the dynamic response of the crane, due to wave induced excitations. Due to the random nature of waves, convergence of the results is researched. For this, numerous time-domain simulations with different sea-states have to be performed, this is unwanted.
A methodology is developed, using the frequency domain, to limit the number of required time-domain simulations. Using the devised methodology, configurations with similar statistical description can be identified and grouped.
The annual probability of failure can be calculated using the devised methodology and probabilistic model. The resulting annual probability of failure is small, much smaller than required. A sensitivity analysis is performed, it is found that the annual probability of failure depends greatly on the input of the probabilistic model.
From the results and sensitivity analysis, multiple recommendations are given to decrease the sensitivity of the result. It can be said that from the devised methodology, a basis has been laid to quantify the annual probability of failure of the crane in normal operations.

Regular maintenance of civil engineering structures is essential for their safety. Current maintenance regimes involve periodic inspections at regular time intervals. In the time between inspections, there can be a critical development in the structural integrity of a structure, which can be expensive to repair or could even lead to structural failure. A more robust maintenance approach would involve continuous monitoring of the structure. This is the research area of Structural Health Monitoring (SHM). Vibration-based monitoring, which is a subset of SHM, aims to provide new cost-effective maintenance solutions that provide long-term life-safety benefits. Vibration-based monitoring uses vibration measurements from sensors to assess the "Health" of a structure. Damage in a structure will alter the structure's stiffness, mass and damping characteristics, which in turn will change the dynamic properties of the system. This change can then be discovered in the vibration data. The growth of the field is partly due to the significant advances in data-driven science and engineering in recent decades.

One of the issues in vibration-based monitoring is the presence of operational and environmental variability in the vibration data. With this variability present, it is challenging to determine from vibration data the characteristics of the underlying dynamic system. The aim of this thesis is to use Robust Principal Component Analysis (rPCA) to reduce or eliminate the operational variability from the traffic to allow for environmental and damage detection. rPCA is a matrix factorisation method that decomposes a data matrix into a low-rank matrix L and a sparse matrix S. The reconstructed low-rank matrix L contains the main correlations in the data that are robust to outliers and corrupt data that are contained in the sparse matrix S. By applying rPCA to the frequency representation of the vibration data, it is hoped that the underlying coherent structure corresponding to the dynamic system can be recovered.


The vibration data used for this thesis is from two measurement campaigns conducted on the Haringvlietbrug. The Haringvlietbrug is a steel box girder bridge in the Netherlands, and there are several fatigue cracks present in the bridge. This presented an opportunity for damage detection. The goal of the first measurement campaign is to conduct damage detection and discover if there is a difference between vibration data from a damaged area with fatigue cracks and a "healthy" reference area. In the second measurement campaign, the goal was to extract the underlying dynamics of the structure at different temperatures and see if it was possible to distinguish between the different structural states at different temperatures. After applying the rPCA on the vibration data, (regular) principal component analysis (PCA) is used to embed the data into the low-rank subspace of the PCs to distinguish between the different structural states.

The rPCA was successful in extracting the coherent structures in the vibration data corresponding to the underlying dynamic properties of the system. In the subsequent PCA, vibration data with underlying different structural states had different scores in the first three PCs. In other words, it was possible to distinguish between the different structural states based on the first three PCs, which correspond to the main correlation within the data. For the first measurement campaign, this meant it was possible to distinguish between vibration data in the damaged area and the "healthy" reference area and detect "damage". However, there was a difference in the structural configuration between the two areas, so it was not possible to conclude that the differences in vibration data were due to damage caused by the fatigue cracks. In the second measurement campaign, the dynamic system properties at different temperatures were recovered. With the low-rank vibration data from the rPCA, it was possible to distinguish between vibration data with a 1°C difference in the first three PCs from the (regular) PCA. This was not possible without lowering the regularisation parameter in the rPCA. Another method, the Sparse Sensor Placement for Optimisation (SSPOC), was used to determine the locations in the frequency spectrum that contained the largest differences between structural states. For both the first and second measurement campaigns, these locations were at specific natural frequencies of the system.

The offshore wind energy industry is quite literally reaching for the sky to meet the increasing demand for renewable energy in Europe. Offshore wind turbines are sprouting out of the waves off the coast of many European nations and their size is increasing rapidly. As the turbines become larger and the waters in which they are placed become deeper, the foundations that carry these behemoths must grow as well. The most frequently used foundation is the monopile foundation. Over the years, both the length and the diameter of these monopiles have steadily increased. Installing monopile foundations has, as a result of this, become more and more difficult.
An important aspect of the installation of monopile foundations that has become more challenging is the upending of the monopile. When the monopiles are transported to the site of installation, they are sea-fastened to the deck in horizontal orientation. When they arrive, they need to be rotated to a vertical orientation for installation. This can be done with the aid of an upend hinge. This thesis focusses on the modelling of such a hinge for operability studies.
The goal of these operability studies is to determine the weather and wave conditions in which the operation can be safely performed. Conventionally, the stiffness of the upend hinge is approximated by several linear springs between the vessel and the monopile. The aim of this thesis is to apply dynamic substructuring to existing Finite Element models of a hinge to more accurately describe its dynamic behaviour within hydrodynamic simulations. Hereafter, the response of the dynamically substructured model can be compared to the response of the conventional model.
The findings of the research show that it is possible to use dynamic substructuring to reduce the Finite Element models of an upend hinge and implement them into hydrodynamic simulations of an upend operation. When comparing the conventional model to the substructured model, it can be seen that there is a significant difference between the high-frequency response of the models. However, the responses of the models to prevailing ocean waves are very similar. For operability studies, this response is most relevant. Furthermore, the required computation time is significantly higher for the simulations employing the dynamically substructured models than for the conventional models. Therefore, it is not recommended to apply dynamic substructuring to model this upend hinge in operability studies. For situations where the high-frequency response is more relevant, however, dynamic substructuring may prove a valuable tool to more accurately describe the dynamic properties of an upend hinge or other marine equipment.

With the highly increased interest in offshore wind turbines and their technologies, the sector has witnessed rapid development in the past decade. For installed offshore wind turbines, there has been a lot of research conducted in the field of aero-, hydro- and structural dynamics, wind turbine-control, operation & maintenance, foundations and moorings. The research field of offshore wind turbine installation is on the other hand relatively new, and the studies regarding this topic are limited. The lifting of heavy objects is one of the most commonly performed offshore installation operations and has become more challenging due to the trend of increasingly larger and heavier payloads. Especially for substantial waves, the pendular motions of the payload may cause operations to be halted.

This thesis performs a study on a positioning control strategy for a complex lifting control scenario, i.e., position-keeping of a complex-shaped 6-DOF payload using a floating vessel equipped with multiple tugger winches. As the system is highly complex and contains non-linear and time-varying dynamic phenomena, it is an impracticable task to formulate a model that meticulously describes the actual system. For this reason, a fully integrated simulation model in Orcaflex has been used to capture the non-linear dynamic behaviour of the system.

The preassembly operation of a Jacket Lifting Tool on a monohull vessel is adopted as a case study to verify the proposed control strategy. Two scenarios are considered -installation and decommissioning- for which an outrigger configuration is used to position the tugger winches. Due to the difference in setpoint (i.e. the desired position) in the two scenarios, the proposed controller is solely implemented in the decommissioning scenario. Damping tuggers, the current state-of-the-art when it comes to motion mitigation, is considered suitable in the installation scenario.

The proposed controller does not consider the state–space equations of the system and only relies on real-time motion and tension measurements of the vessel and suspended payload. In addition, the controller considers the system's velocity tension and power limitations. The controller's impact is evaluated based on the positional error and verified by the peak reduction in the power spectral density spectra of the simulations. Despite its simple form, results show a significant reduction in the positional error, and therefore the possibility to extend the working conditions of the installation vessel. To improve the controller's performance it is recommended to involve derivative control, consider payload motion prediction and to optimise the tugger winch configuration. For further studies, experimental testing is needed to verify the effectiveness of the control scheme as it could appear that the controller does not exhibit similar performance in the real system. However, it is deemed unlikely that the latter would occur as a sensitivity study regarding measurement error indicates a stable response of the controller.

The Haringvlietbrug, which was delivered in 1964 and connects the island of Goeree-Overflakkee to the Dutch mainland, approaches the end of its lifetime. Rijkswaterstaat and Witteveen+Bos investigate the potential use of vibration-based structural health monitoring (SHM) to track deterioration. In vibration-based SHM, a change in the structure’s eigenfrequencies indicates probable damage but most current approaches focus on fundamental modes of the structure which are not or hardly affected by small structural impairments. In this study, vibration-based SHM was employed on local modes of vibration of the Haringvlietbrug, with the hope of detecting small-scale damage. New obstacles arose as a result of this: the influence of environmental and operational variabilities (EOV) on the eigenfrequencies of the structure was often larger than the influence of small damage. As a result, the EOVs should be filtered out of the recorded acceleration data before damage sensitive features can be extracted.
To investigate this, 32 accelerometers and 16 temperature sensors were installed in two segments of the Haringvlietbrug. The sensors yield a large data set which was analysed extensively. This made it clear, among other things, that the dynamic response of the bridge deck showed large variability when comparing different vehicle passages.

The first research objective was to investigate the applicability of similarity filtering (SF) to filter out operational variabilities from the Haringvlietbrug acceleration signals to extract damage-sensitive features. SF amplified similarities and damped differences between samples of vehicle passages. This way, only consequently excited modes should remain in the signals which were subsequently used as damage-sensitive features for SHM.
This study found that using SF to filter the operational variabilities from the Haringvlietbrug data set was ineffective. Three reasons were identified for this. First, the method was not robust as a single deviating sample or a poorly chosen filter coefficient significantly influenced the results. Secondly, missing closely-spaced modes might have caused inconsistency of the results. Lastly, SF was not able to converge to consistent behaviour as the variability of response of the bridge deck might be too great for different passages. The latter reason was investigated further in the remaining of the research.

The second research objective was to improve the understanding of the effect of vehicles on local bridge vibrations of the Haringvlietbrug for the application of vibration-based SHM. First, the eigensystem realisation algorithm (ERA) was used to identify and compare consistent mode shapes in order to better understand bridge deck vibrations and find patterns in the eigenfrequencies. ERA is an operational modal analysis (OMA) and output-only system identification technique.
ERA was able to identify two consistent modes in the data but they had a large variance in both shape and eigenfrequency. The uncertainties of the results were too large to draw firm conclusions based on ERA because two of ERA’s key assumptions were not perfectly met and the input samples were not optimal.

Next, a semi-analytical model of a segment of the Haringvlietbrug was built to simulate the important sources of the variability as found in the acceleration recordings of the Haringvlietbrug. The goal of the model was to investigate the sensitivity of the response of the beam to variations of input parameters. The model was both used for time-history analyses and eigenfrequency analyses. Parametric studies showed that the response of the beam was highly sensitive to the time delay between moving masses (dependent on the axle configuration and vehicle velocity), the unevenness (describing any source of vibration of the interaction between vehicle and bridge deck) and the vehicle velocity. These parameters influenced both the amplitude and the shape of the acceleration response of the Haringvlietbrug bridge deck in the time and frequency domain.
The effect of the three dominant parameters was further investigated by simulating four characteristic cases from the Haringvlietbrug data set. Hypotheses on the source of the measurement variability were formed by qualitatively comparing the results of the model to the measured response of the bridge. The model parameters were able to describe a large part of the variability but it was concluded that it is likely that some factors that were not included in the model also play a role in the variability of the measurements.

The above findings led to the following recommendations for further research. Firstly, it is recommended to only select similar vehicle passages for SF because this would improve the ability to converge to consistent modal behaviour. Secondly, should someone want to further explore ERA, it is recommended to equip a section of the bridge with a high spatial density of accelerometers to improve the reliability of the results.
Next, some recommendations related to the model were made. The first step of follow-up research would be to verify the influence of the model parameters on the dynamic response of the Haringvlietbrug by experimenting with different test vehicles. Subsequently, depending on the importance of the parameters, possible extensions of the measurement set-up for a new campaign were proposed. Secondly, the current model could be improved by introducing more parameters, like acceleration of the moving mass or by implementing a more realistic vehicle model, to be better able to explain the dynamic variability. Thirdly, more parameters could be investigated by building a 3D finite element model. This would make it possible to investigate the influence of the presence of multiple vehicles on the bridge and 3D wave propagation.
Upcoming research into data-driven approaches of vibration-based SHM of bridge decks is recommended to focus on similar passages as not all samples contain the same information on the dynamic behaviour of the structure. Decreasing the variability of the input samples might improve the performance of the algorithms.

Due to increasingly efficient engineering of the last decades, the design process for civil structures is heading in the direction of optimal solutions. Among other things, the reduction in self-weight enhances the susceptibility to dynamic loads. As the process for the determination of realistic modal parameters from a digital model often diverges from the reality, the uncertainties should be investigated. This research aims to gain knowledge on the accuracy and stability of such modal parameter estimates.

A case study is investigated to determine modal parameters for a realized tied-arch railway bridge. The desired modal properties include the natural frequencies, mode shapes and damping parameters. The frequency domain decomposition (FDD) method is applied on a set of accelerational measurements to determine the modal parameters.
Four dominant frequencies are identified and investigated, from which the most dominant operational deflection shapes are extracted. The overall response of the structure shows great similarities with the expected mode shapes for similar bridge types. The computed modes show signs of complex behaviour, due to the characteristics of the load or the structure. An effective method to reduce uncorrelated noise is the application of an autocorrelation function (ACF) to the time domain signals, before execution of the FDD. Prior to the computation of the final results related to the case study, a simplified test case is considered to validate correctness of the FDD implementation.

Wind turbines are typically designed for an operational life of 20-25 years. The operation of the assets can be extended beyond their design life if structural components have sufficient reserves left. One approach is to monitor fatigue loads and compare these with design assumptions to determine the remaining useful lifetime of the assets. A major challenge is that sensors for measuring the stress history, such as strain gauges, only deliver local information. Monitoring of every hot spot is technically and financially not feasible due to cost and access restrictions. In addition, strain gauges only have a limited lifetime when compared to accelerometers. Several response estimation or extrapolation methods have been proposed in literature to tackle this problem. All of them are Kalman filter based methods, with the exception of the Modal Decomposition and Expansion Method. A new Kalman filter based method has been recently proposed in literature called Gaussian Process Latent Force Model. The aim of this work is to assess this new method with respect to existing ones both theoretically and numerically. Theoretically, the Kalman filter based methods always rely on white gaussian noise assumptions for the unknown loads, and the modal decomposition and expansion method disregard measurement imperfections. The new method improves upon these assumptions by providing a flexible stochastic definition for the unknown load, and by taking into account measurement imperfections. The numerical analysis is restricted to a comparison with respect to the modal decomposition and expansion method given the different theoretical backgrounds. The comparison is realised upon a simulation which illustrates and validates the methods, and upon a real-life onshore wind turbine equipped with accelerometers and strain gauges. Measured strains are compared with estimated strains. The accuracy of each method is quantified using the mean absolute error and by the correlation between measurement and estimation. The higher accuracy obtained shows that the new method is an improvement upon existing methods. This is further extended in the calculation of Damage Equivalent Loads. The result shows a relative error that, depending on operational conditions, ranges within 20-40[%] for the modal decomposition and expansion method and less than 10[%] for the new method. These results show that the novel Gaussian Process Latent Force Model method should be taken into account for response estimation when accuracy is relevant. Future works should aim on developing a mechanical model that better capture the real behaviour of the wind turbine, as the accuracy of the response estimation methods is mainly controlled by the validity of the underlying assumptions of the mechanical model. Furthermore, the strain estimations should be sought in the whole frequency range, this can be realised by including measurements that deliver this information: GPS sensors or inclinometers for example.

Technology improvements and growing maturity of the offshore wind industry have resulted in significant cost reductions and rise in demand. More can be achieved by focusing on improving the understanding of key design areas of an offshore wind turbine (OWT). Since fatigue is one of the main design criteria for offshore structures and little is known about fatigue cumulative development in time under operating conditions, it formed the basis of this thesis. However, fatigue is a complex phenomenon that is dependent on numerous interlinked parameters, such as damping, loading type, and stiffness of support structure (SUS). For the purpose, data analysis was performed on measurements which were gathered over a period of four years from a monopile-supported OWT site. The measurement data included information about the environmental and operational conditions of the OWT; and strain, acceleration and inclinometer readings from its SUS. Use of different data acquisition systems during the measurement campaign inherently required significant efforts to synchronise and preprocess raw data to appropriate state for data analysis. Time-dependent lag was identified via computation of cross-correlation sequences between data segment pairings, and the lag was successfully corrected. With the use of a standardized rainflow cycle counting algorithm, strain-derived moment time series were converted to constant amplitude events, which could be used to derive damage equivalent bending moments, M-N curves and fatigue damage accumulation frequency spectra. Short-term damage equivalent bending moment (STEL) was revealed to have a linear dependency with turbulence intensity when the OWT was in the run-up operational state. Additionally, the fatigue damage was demonstrated to be higher for downstream turbines due to wake effects. Also, fatigue life consumption of OWT SUS increases above rated wind speed conditions despite decrease in load magnitude due to increase in SUS response frequency. The latter was confirmed with fatigue damage accumulation frequency spectra, which revealed that the importance of high frequency band (1 – 5 Hz) increases in conjunction with wind speed at above rated wind speed conditions. However, most of the overall fatigue damage is accumulated at the low frequency band, which associated with wind and wave loading, first SUS bending modes, 1P and 3P rotor harmonics.

Over the last decade, the demand for offshore power cables has increased significantly due to, in particularly, the emergence of offshore wind. Despite this growth, industry-wide rules and guidelines to analyse fatigue during cable installation operations are yet to be developed, even though fatigue can become an important parameter in the installation design.  In similar offshore operations, e.g. installation of pipelines or umbilicals, fatigue is assessed based on fundamental fatigue theory like rainflow counting, S-N curves and Miner's rule and therefore these principles form the basis of analysis in this research. However, submarine power cables consist of numerous layers of different materials that result in complex cable cross-sections. The fatigue behaviour of the cross-sectional elements due to external loads has been investigated to determine a maximum stand-by time during installation operations. In this regard, a cross-sectional analysis was performed to establish the stress-strain response of the individual cable components to the global cable deformations. For bending behaviour, the analysis was shown to be consistent with existing test data. However, this type of data is scarce and the analysis was based on a limited sample size. Moreover, geometrical cross-sectional data of submarine cables is rarely provided by suppliers and hence several assumptions were made in the analysis. For future work, it is recommended to develop in-house test data of cables such that the cross-sectional model can be accurately verified and improved. The cross-sectional stresses were implemented in a fatigue assessment model, in which the local stresses were calculated based on the output of global modelling software OrcaFlex and subsequently analysed with rainflow counting and Miner's rule. The model was applied to three cable types and it was found hat lead, used for cable sheaths, is the critical cable component in terms of fatigue. When no lead is used in the power cable, the conductors are the critical component. Furthermore, for mild loading scenarios, i.e. waves with significant wave heights H_s≤1.5 m, the model yields components infinite life for all cable components. For higher load cases, i.e. H_s≥2.5m, the model showed that fatigue mitigation measures are required to stay within the fatigue budget. Several mitigation methods were implemented, from which it was found that increasing the layback length of the cable combined with adjusting the vessel heading to favourable conditions most effectively reduces fatigue damage as all load cases with H_s=2.5 m resulted in maximum stand-by times of t_sb≥ 125days. However, for higher loading scenarios, i.e. Hs=4 m, all researched mitigation measures were insufficient to stay within fatigue budget. For future work, it is therefore recommended to improve the mitigation measures such that severe load cases can be survived. Lastly, the wave conditions were simulated as JONSWAP waves. These simulations are time-consuming and therefore not many loading scenarios were modelled. It is recommended to research methods to approximate wave conditions with regular waves, as the duration of OrcaFlex simulations will decrease exponentially. An increase in number of load cases will result in more accurate limit states for the cable components.  

In the Netherlands there are many bridges that crosses waterways, which are key nodes in the transportation system. The safety and integrity of these bridges is nowadays assessed by (visual) inspections at regular intervals. This approach cannot provide information about damage development in between inspections, leading to potential failure or high intervention costs. A maintenance strategy for bridges based on continuous vibration-based monitoring provides a possible cost-effective solution that may replace the periodic (visual) inspections. Vibration-based structural health monitoring of civil engineering structures is receiving increasing attention in recent years. This is due to new developments in related areas such as sensing technology, system identification, data mining and condition assessment. A particular damage scenario that frequently occurs in steel orthotropic deck structures is the development of fatigue cracks in the welds between the longitudinal U-shaped stiffeners and the transverse beams or the steel deck plate. This damage scenario was also observed in the steel deck structure of the Zwartewaterbrug, located in Hasselt in the Netherlands. Vibration tests were performed on the bridge to develop a damage detection algorithm for early detection of the fatigue cracks in the bridge deck. In order to simulate the effect of damage, small masses were attached to the bottom of the bridge deck. The measurements were performed over a relatively short time period to exclude – in as far as possible – the effects of environmental variability. However, the bridge was kept open to traffic during the testing and structural response was primarily triggered by traffic load. The vibration data (accelerations) obtained during these tests was used for validation of the proposed methodology in this thesis. The aim of the thesis is to detect the added masses from the vibration data. To achieve this, a data-based approach for damage detection was proposed, which, besides the data acquisition, consists of four main stages: data preprocessing, feature extraction, pattern recognition and decision making. The analysis is focused on the high-frequency acceleration response, because the associated higher-frequency local modes are more sensitive to the small changes induces by the added mass. The similarity filtering preprocessing procedure was applied to filter out the variability related to the operational loading, and the Frequency Domain Decomposition system identification technique was used to extract damage indicators. Subsequently, a support vector machine algorithm was applied to “learn” patterns in the extracted damage indicators, which enables the algorithm to assign a damage state to a given measurement vector. Alternatively, a novelty detection technique was used to distinguish data from the “healthy” structure from data corresponding to a “damage” state. It was not possible to reliable detect the added masses from the high-frequency acceleration data due to the large variations in the damage indicators of a mass class. These variations are attributed to the changing environmental conditions over the measurement period. The principal component analysis was used in an attempt to remove these variations from the damage indicators, but the results were still not satisfactory. In conclusion, it was not possible to properly test the proposed methodology for damage detection using the data of the Zwartewarterbrug, mainly because the data was not suitable for the machine learning algorithms employed. Recommendations are given related to the methodology for the similarity filtering, determination of the high-frequency range, and machine learning algorithm to be used. Additionally, recommendations are given for future measurement campaigns to obtain data that is more suitable for successful application of the methodologies employed in the present project.

The offshore wind industry has grown rapidly over the last decade and drivetrains are increasing in size to reduce the cost of energy. These turbines are operating in a harsh environment. Adopting a preventive maintenance strategy is important to achieve an as high as possible availability of the farm and reduce the cost of maintenance. A well performing condition monitoring system that utilizes SCADA data from the wind farm can enable this strategy without the need in additional cost in hardware. This master thesis focusses on the development of a framework that can be utilized for this task. This framework can process raw operational SCADA data collected at the Egmond aan Zee offshore wind farm to create a clean dataset to train supervised machine learning models on. This work provides an insight in the correlation between different SCADA signals using a mathematical approach and from a understanding of the system integration of drivetrain components. Bearing temperatures are modelled using a data driven approach to describe the temperatures under healthy conditions. Several models are evaluated for this task and it was concluded that a decision tree supervised machine learning regression model resulted in the lowest error between predicted and measured values. Anomalies are detected and tracked with a normal behaviour model and a Sherward and CUSUM control chart that are applied on the residual error between modelled and measured temperature signals. 4 anomalies could be identified in the gearbox bearings using the developed framework. Abnormal behaviour of the drivetrain could be identified as early as 1 month before the turbine was taken out of productions. This highlights that temperature based condition monitoring that utilizes SCADA data can be used for early detection of faults by combining the accuracy of supervised machine learning methods with different fault detection methods like the CUSUM control chart. This work also investigates the relation between experienced wake of a wind turbine and the influence on the drivetrain component temperatures. The wake conditions at Egmond aan Zee, modelled with an Ishahara wake model, and the component temperature measurements from the SCADA data are used for this analysis. The bearing temperature distributions under different operational and wake conditions can be compared by clustering over the wind speed and the velocity deficit or turbulence intensity at turbine level. It is concluded from this work that wake effects do not result in a change in drivetrain component temperatures. The effects of asymmetric wake conditions opposed to wake experienced over the entire rotor is analysed by comparing the temperature distributions under these conditions in a cluster where the turbine is partially waked. A small shift towards higher component temperatures can observed on a limited amount of data for turbines under asymmetric loading conditions.

Fatigue is often a governing design factor for offshore wind turbines. Since the design of offshore wind turbines includes conservatism, the actual accumulated fatigue damage can be lower than what the turbine is designed for. In this case, the operator can make a decision on life time extension of existing wind turbines. Therefore, it is important to estimate the actual accumulated fatigue damage to support decision making on life time extension, and for optimization of support structure design. However, fatigue critical locations are located near mudline where it is unfeasible to install strain gauges to measure the accumulated fatigue damage. The first purpose of this thesis is to investigate if data-driven approaches (linear regression and feed-forward neural network) can be applied to estimate the accumulated fatigue damage both in individual turbines and farm-wide levels. The second purpose is to determine the minimum number of sensors and quantity of data required for accurate estimation. Towards this goal, real measurement data of two offshore turbines in the same wind farm have been used. Specifically, the data-driven approaches have been applied with real measurement data from the SCADA system, measurements at the top and bottom of the tower, and data from a wave measurement system. This data was used to estimate the accumulated fatigue damage at multiple locations (tower bottom, transition piece and two levels on the monopile) in the form of damage equivalent loads. Throughout the study, 10 min. statistical properties of the measurement data have been used as input to the learning algorithms. One remark is that the estimation has not been performed for the fatigue critical location near mudline itself, but it is expected that estimation with these approaches can be expanded to the fatigue critical location if accurate response estimation at multiple locations on the support structure is possible. The results of this thesis show that the data-driven approaches can give accurate estimates damage equivalent loads on individual turbine level at multiple locations on the support structure when moment or inclination signals at tower bottom is used. For farm-wide level load estimation as well, it has been proven that the data-driven approaches can give quite accurate estimates the damage equivalent load. However, it should be noted that the turbines used in this study have similar dynamic properties. Therefore, the farm-wide level load estimation with the data-driven approaches should be further investigated in the future.

Over 450 oil platforms need to be removed from the North Sea in the next decade. Each of these platforms is custom designed for installation and service lifetime, but not for decommissioning. Allseas' Pioneering Spirit is the largest heavy lift vessel in the world, capable of lifting complete topsides in a single-lift operation. The vessel’s lift system places up to eight lift points against the bottom of the platform and uses hydraulic cylinders to transfer the weight from the substructure to the vessel in a matter of seconds. But even with this vessel, removing an oil platform is technically challenging; the structural integrity needs to remain guaranteed, while minimizing the lift preparation cost as the structure will be scrapped. The goal of this thesis is to define a general approach that uses cost optimization for the platforms’ lift preparation scope.

The costs of the platform’s lift preparation scope consists mainly of two things; installation of lift points and reinforcement of the structure to avoid exceedance of their design resistance. In order to minimize the cost, the challenge is to find a lift configuration which minimizes the combined cost of the number, type and location of the lift points and the amount of reinforcement. A solution should not exceed constraints such as the allowable load per lift point, the number of lift points and the relative height difference between lift points. A lift configuration is statically undetermined, and the relative height difference between lift points is used to get different load distributions for the same configuration. In this way, forces are redistributed to strong points in the structure.

Detailed finite elements models are used to prove the structural integrity of a topsides during decommissioning. The number of degrees-of-freedom of such models is far too large for them to be used in an optimization algorithm. Model order reduction is used to condense the mass and stiffness matrix to the possible lift points, while keeping internal forces of interesting elements related to these points to check the structural integrity of a solution. The reduction allows for hundreds of lift configurations to be solved per minute.

The minimal costs for a lift configuration in the defined domain is found with the genetic algorithm. The optimization can include an excess amount of lift locations of different possible types. The location and load distribution over the lift points are used to define the cost function. The algorithm makes a selection that suits the constraints. The cost function gives the sum of the cost of the number and type of lift points and the number and type of structural failures.

This thesis presents a systematic approach for finding the optimal lift configuration for any offshore structure. The method is tested on a case study topsides by comparing the standard configuration with the optimized configuration. It is shown that by using the proposed method, the areas that need structural reinforcements are reduced from ten to zero and the amount of lift points is reduced from eight to seven, leading to a cost reduction of 27% on the offshore preparation scope.

The algorithm shows promising results for this topsides removal project. It is possible to include a new finite element model into the algorithm, which together with a generic cost function, makes sure the algorithm is conveniently applicable to other topsides removal projects.

Vibration-based structural health monitoring of civil engineering structures is receiving increasing attention in recent years. This is due to the development of more robust system identification techniques as well as improvements with regard to the practicality of installing the necessary instrumentation. Health monitoring systems are more often replacing static deflection tests and detailed visual inspections, where continuous monitoring of vibration data aims for early damage detection.

This thesis can be summarized as the development of a structural health monitoring method for bridges. It forms a part of the Zwartewaterbrug project, related to the Zwartewater bridge located in the city of Hasselt, in the Dutch province of Overijssel. The Zwartewater bridge served as a case study for validation of the proposed early damage detection algorithm.

The wider purpose of this work was to address the problem of having a large number of bridges in Europe that are facing the end of their service life and should either be retrofitted or decommissioned.. The chosen approach consisted of a low-cost vibration monitoring method, that is suitable for continuous structural health monitoring.

The initial challenge was to introduce a small change to the structure, representative of early damage, without actually damaging the structure. For this an added mass approach was taken, where weights of respectively 25, 50, 75, and 100 kg were added below the bridge deck. The vibration data (accelerations) measured on the ”damaged” structure were then used to solve an inverse problem where the aim was to detect the induced damage without closing the bridge to traffic.

Since the data was gathered with only ambient excitation by wind and traffic, a new challenge arose, which is considered to be the core of this thesis. The structural differences between vibration measurements related to the ”healthy” and ”damaged” structure were concluded to be smaller than the structural changes due to the differences in unknown traffic loading. The traffic loads were namely varying from 200kg motorcycles to 50t trucks.

In this thesis, the method of Probabilistic Filtering was developed to face the aforementioned challenge. The method involves mainly a data pre-processing step, and the probabilistically filtered signals are subsequently processed with two output-only identification methods: Frequency Domain Decomposition and data-driven Stochastic Subspace Identification. The full process was integrated into an automatic analysis algorithm, performed in MatLab.

It was concluded that both detection and localization of the induced damage is possible with the proposed methodology. Further work was then performed to quantify the detected added mass in terms of structural damage. A quantitative finite element analysis was performed with the substructure approach (super-element, static boundary conditions to the substructure of interest). The given analysis concluded that the developed structural health monitoring method is able to detect anomalies comparable to a 35cm crack in the welds securing the longitudinal stiffeners to the transverse beams.

Finally, recommendations were made regarding an expansion of the method to structure service life predictions and more computationally efficient calculation set-ups.