Understanding the fatigue load history of wind turbines is critical for taking decisions regarding the lifetime of a project. However, direct measurement of fatigue loads at each turbine in a wind farm is unfeasible. For this reason, surrogate models offer a useful alternative. In this thesis, a methodology for creating surrogate models for emulating fatigue loads of offshore wind turbines is presented. The methodology is unique in that it accounts for the variability of site-specific conditions that may be present between wind turbines of the same class. First, a method for creating simplified structural models which depends only on a few degrees of freedom is derived. After this, a database of simulation data is assembled by varying the geometric, dynamic, and environmental degrees of freedom within ranges which capture a high degree of variability of possible site-specific conditions. This database is then used to train surrogate models using neural networks.

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.

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.