In recent years, the interest in deep-water wind energy projects has drastically increased, driven by the considerable wind resource in deep-water locations. In such locations, floating wind turbines are more economically attractive than bottom-fixed wind turbines. Nevertheless, the floating wind industry faces numerous challenges. One major challenge is the high investment costs involved in the manufacturing of the substructure. Additionally, it is important to maintain the wind turbines’ power output within the required standards for grid connection. Therefore, the design and optimization of substructures for floating wind turbines has to account for both capital costs and power quality. To investigate the existing research on the optimization of floating wind turbines, a literature review was conducted. Evidently, there is limited research investigating both cost and power quality optimization. Furthermore, the majority of the optimization techniques presented are accompanied by a computationally expensive substructure analysis. Moreover, in the early stage of the substructure design, a variety of different alternative scenarios are tested, requiring an extensive number of optimizations. Thus, the computational resources required by the optimization become a major concern. One promising approach to address this issue is substituting a computationally expensive analysis with a surrogate model. This solution could significantly increase the computational efficiency of the optimization process.

This thesis investigates the trade-off between substructure capital cost and power quality optimization. To achieve that, a multi-objective optimization workflow is developed for the case of a semi-submersible substructure. In parallel with this task, the optimization’s analysis blocks requiring most of the computational resources are identified and substituted with a surrogate model. Two different surrogate model implementation approaches have been examined. The first approach substitutes the optimization’s substructure analysis tool (NREL’s RAFT) entirely. This approach is called the ”End-to-End” approach. The second approach substitutes a specific analysis block of RAFT that computes the floater’s hydrodynamic properties. This analysis block requires most of RAFT’s computational resources. This approach is defined as the ”PyHAMS” approach.

The thesis outcome shows that improving the quality of power output comes with the drawback of increasing the substructure’s cost. The Pareto front between these two competing criteria is dominated by designs of long pontoons and reduced chain thickness. This is the most economical approach towards reduced cost and improved power quality. These Pareto designs are mainly driven by the boundaries of the design variables. Additionally, the Pareto designs are affected by constraints on the maximum pitch, maximum surge, and the existence of vertical loads on the mooring line’s anchor. Regarding the surrogate model-based optimizations, there is a dramatic reduction in the computational resources required to conduct the optimization. Among the two surrogate model approaches, the ”PyHAMS” approach achieves a better accuracy at mimicking the original optimization workflow than the ”End-to-End”. Instead, the ”End-to-End” approach is the most computationally efficient one to conduct the optimization. Nevertheless, when factoring in the computational resources to produce the training dataset, the ”PyHAMS” approach is the most computationally efficient one. This is the outcome of utilizing a reduced training dataset compared to one utilized by the ”End-to- End” approach.

Site analysis to determine the loads experienced by wind turbines based on site-specific environmental conditions is typically done using either coupled aero-servo-elastic simulations for onshore wind turbines or coupled aero-servo-hydroelastic simulations in the case of offshore wind turbines. These simulations become computationally expensive when multiple load cases are needed to be taken into account, together with the numerous possible combinations of turbulence inflow patterns that result in the same mean inflow conditions. Probabilistic surrogate models offer a cheap alternative to these expensive simulations for predicting load statistics. This thesis explores a type of neural network called Conditional Generative Adversarial Networks (CGANs) as a potential candidate for such a surrogate model. Originally developed for image generation, CGANs have seen success in other applications. However, most applications of GANs to date are high-dimensional, with relatively low research focused on low-dimensional problems such as wind turbine load statistics. Multiple experiments are conducted using various multimodal and heteroscedastic datasets to assess its ability to model such characteristics
accurately. The conditional log-likelihood andWasserstein-1 distances were used as metrics. The results show that CGANs can indeed model such low-dimensional datasets. Finally, the CGANs are trained on data from simulations of onshore and offshore wind turbines in OpenFAST and compared with predictions from Mixture Density Networks (MDNs). The results from CGANs are comparable to MDNs, showing its potential as another alternative surrogate method, although more research needs to be performed.

Considering the goals set by the international community, the implementation of new energy sources has to increase considerably in the next seven years. In this thesis, the focus is on the acceleration and improvement of the application of offshore wind turbines. The power produced using this technology should become 3.6 times more before the end of this decade to comply with the set goals.

To achieve this target, new solutions have to be developed. To this aim, the implementation of model predictive control for wind turbines in the last years has been investigated. This would allow to optimize different parameters at the same time, such as maximization of energy production while minimizing the perceived loads. For this application, it is necessary to have forecasts of different features of the turbine, especially loads, with a higher frequency. As a result, the main research topic is defined as 'Can data-driven surrogate models be used for forecasting load time series on offshore wind turbines?'.

To answer this question, first environmental conditions are sampled within limits deducted from real data through Halton sequencing, and next simulations are run through OpenFAST to determine the resulting loads acting on the turbine. Within all the features resulting from the simulation, only five inputs and five target outputs are selected. This is the result of various considerations. Given the desire to develop a realistic methodology, the input variables are first filtered by assessing their availability from measurement devices. Next, the relationships between the variables are analyzed through cross-correlation to determine the degree of influence of each input on the output.

Using this data, a training database is created. It is used to train two different types of surrogate models, one linear and one non-linear, respectively ARIMAX and LSTM. These are implemented to generate a 30-second forecast of the moments acting at the root of the blade. To do so, the algorithms are trained using different variables as exogenous inputs to assess the models' performance in different cases. Given the wide range of target features, for LSTM two different behaviors are identified and the blade edgewise and flapwise moments are taken as examples. The hyper-parameters are tuned on the blade edgewise moment and lead to overfitting when applied to the blade flapwise and out-of-plane moment.

The obtained results show that the RMSE in ARIMAX is up to seven times larger than the one obtained from the application of LSTM. Within the non-linear models, the one resulting in the lowest percentage error for the blade edgewise, pitching, and in-plane moment considers the wind reference speed, the wind speed time series, and the corresponding tip deflection as exogenous inputs. Very low RMSE errors are obtained for all variables. Furthermore, it is concluded that while it is possible to implement LSTM in real-life, this is not achievable for ARIMAX.