Statistical and Modelling Uncertainties in the Design of Offshore Wind Turbines

Abstract

This thesis aims to contribute to lowering the cost of offshore wind energy by minimizing design conservatism. By increased confidence in the structural design, the total cost of energy can be reduced through less material consumption or extending the operational lifetime. Both model- and statistical uncertainties in the design of offshore wind turbines have been systematically addressed in this thesis to evaluate the potential design conservatism. Two major research objectives (RO) have been formulated: RO1 Assessment of environmental load effects and their impact on the reliability of a monopile-mounted offshore wind turbine RO2 Develop methodologies to reduce excessive design conservatism by means of highfidelity models and time-domain simulations in a probabilistic analysis framework To date, all commercial offshore wind farms are bottom-fixed, and the monopile foundation is currently the most cost-effective design at shallow to intermediate water depths. Hence, the case studies in this thesis are based on the next-generation monopile-mounted offshore wind turbines with a rated capacity of 10 MW. Numerical, fully coupled, aero-servo-hydroelastic simulations and statistical analysis have been the main methodologies in this thesis. As future wind farms are being planned in deeper water far from shore, the impact from wave loads on the final design of the sub-structure is increasing. Hence, more accurate methods are needed to capture important load and response phenomena, such as wave diffraction and ringing. It has been observed that new and improved methods may lead to both lower and higher characteristic response values to be used in the design phase. In this thesis, the model-specific hydrodynamic load effect uncertainties have been identified and their impact on the partial safety factors are discussed and exemplified. An improved probabilistic description of the ocean environment is proposed. The joint probability distribution includes state-of-the-art models for wind speed and wind sea, and is extended to include their respective directions, separation of swell, and astronomical tide. For offshore wind turbines, directional descriptions are important due to the large aerodynamic damping induced by the rotor in the fore-aft direction. The impact of directional models on the foundation fatigue and structural reliability have been investigated. It has been shown that separation of wind sea and swell reduces the foundation fatigue damage. An engineering problem will often consist of model uncertainties, which may be improved by research, and statistical uncertainties that can only be handled by gathering more data and performing more simulations. It has been shown in the present work that both short- and long-term statistical uncertainties can be reduced by smart simulation techniques, efficient use of computational resources and structural reliability analysis.