In recent years, the offshore wind energy industry has been facing significant barriers, ranging from supply chain disruptions to financial challenges due to rising commodity prices and increasing inflation. To keep offshore wind energy competitive, more efforts towards reducing costs and accelerating project timelines are needed. The optimization of load calculation processes, which are central services in the development of any wind energy project, would significantly contribute to the achievement of these goals.

Conventionally, load calculations are performed using high-fidelity numerical models that simulate the dynamic response of the turbine or wind farm. Although this approach is highly reliable, it is computationally intensive. For applications requiring a large number of load evaluations, this translates into high costs and long project timelines. Load response models have been developed as a data-driven alternative to the high-fidelity numerical models, offering sufficiently reliable load estimates at a cheaper computational cost. However, currently-existing load response models suffer from key limitations pertaining to their applicability to modern offshore wind farms.

This thesis aims to develop a load catalogue, consisting of a database of fatigue loads that can serve as a foundation for a load response model, that better represents the loading conditions in modern offshore wind farms. The database is constructed by running dynamic simulations for a variety of operating conditions tailored to represent those observed in offshore sites. In addition, the thesis seeks to use the catalogue to study the impact of employing the different wind and wake models recommended by the IEC 61400-1 standard on calculated loads. Specifically, it does so by comparing the 90th percentile and Weibull distribution turbulence intensity models, the Mann and Kaimal turbulence models, and the Frandsen and DWM wake models with respect to annual equivalent fatigue loads.

The analysis reveals that using a 90th percentile for the turbulence intensity overestimates the annual fatigue loads obtained when sampling the turbulence intensity from a Weibull distribution by an average of 20% for tower base loads and 6% for blade root loads. Moreover, it shows that the Kaimal and Mann turbulence models lead to significantly different fatigue load estimates. Furthermore, it finds that Frandsen's effective turbulence model consistently predicts higher loads than the DWM model for inter-turbine spacings smaller than 5 rotor diameters, but it can predict lower loads for inter-turbine spacings exceeding 5 rotor diameters.

The study concludes that using a 90th percentile turbulence intensity might lead to overly conservative designs, which can be avoided by using a Weibull distribution that accurately characterizes site turbulence. Also, it suggests that a combined modeling approach to long-term fatigue calculations that selectively toggles between the Kaimal and Mann turbulence models based on the short-term environmental conditions is expected to yield more accurate load estimates compared to employing a single model exclusively. Lastly, it implies that the effective turbulence model might not always be conservative, especially for wind farms with large inter-turbine spacings.

Ultimately, this research culminates in the delivery of a valuable instrument, the load catalogue, that enhances decision-making in the preliminary stages of project development and facilitates the development of better load response models. Additionally, it enriches the theoretical understanding of load calculation methodologies, paving the way for the development of better load calculation techniques.