Assessment of monopiles for lifetime extension of offshore wind turbines

Abstract

Offshore wind is a fast-growing industry contributing to reduce CO2 emissions. Although most offshore wind turbines are still young today, we will face a significant fleet of ageing assets in the near future. The question arises whether these offshore wind turbines can continue to operate beyond their 20-25 years of design lifetime. This thesis investigates methods to assess if monopiles have structural reserves left for lifetime extension. In addition, a better understanding of the end-of-life decision addressing lifetime extension and repowering was targeted. Structural reserves can be analysed using (i) updated numerical simulations (‘analytical assessments’), (ii) measurements of fatigue loads (‘data-driven assessment’), and (iii) inspections (‘practical assessments’). Analytical assessments were performed for two case studies with multiple offshore wind turbines using aero-hydro-servo-elastic simulation software. Results show that the sensitivity of remaining useful lifetimes to environmental and operational conditions varies for each site. Since simulations were computational demanding, the elementary effects method was applied to efficiently screen a large parameters space. The core of this thesis is a novel fatigue load monitoring concept. It tracks loads at the entire monopile while only requiring strain gauges at the transition piece as sensors. The concept utilizes the correlation between damage equivalent loads at different location of the structure. A k-nearest neighbour regression algorithm was applied to extrapolate measured damage equivalent loads. Validation with measurement data from two offshore wind turbines showed promising results. Monthly damage equivalent loads were predicted with errors below 5%. Paris’ law was applied to simulate fatigue crack growth at a circumferential weld of a monopile. Results proved to be sensitive to uncertain parameters, such as initial crack size and material properties. Furthermore, the effect of load sequence and weather seasonality was demonstrated. An investigation of inspection methods confirmed that lifetime extension decisions for monopiles suffer from low probability of detection of fatigue cracks. Bayesian analysis was applied to update results from probabilistic analysis of fatigue crack propagation with uncertain inspection outcomes. A framework for a decision model aiming to optimize the time to switch from lifetime extension to repowering was developed. It demonstrated that end-of-life decisions are sensitive to wear-out of components and future electricity market prices. Results of this thesis contribute to best practises for low-cost and reliable methods to assess structural reserves of monopiles and to make decisions at the end-of-life. This is an important step towards safe lifetime extension of offshore wind turbines to increase the amount of carbon-free energy existing that these existing assets produce for our society.