Abstract
Fossil-fuels are a major cause of global carbon emissions, leading to significant environmental deterioration and possible depletion of naturally sourced materials. Wind power is a sustainable source of green energy and has helped mitigate the greenhouse effects in Europe. Offshore locations offer ample wind resources for greater power output; therefore, wind technology is reaching several countries outside Europe. Many countries located in earthquake-prone regions are planning their contribution to the global offshore wind market with substantial investments. Typically, wind and wave are design driving loads for offshore wind turbines (OWT). However, earthquakes may additionally dominate the design depending on the local geology and seismicity. The history of OWTs in seismically driven sites is short; therefore, the literature and design guidelines are relatively brief on the subject. The design and installation feasibility of wind turbines is assured using the existing codes in consultation with a certifying agency. Both design- and certification-based guidelines quantify earthquake loads with simplified approaches that may lead to underestimation of the total design loads. This can further affect the economic feasibility of the wind farms during the production and service life of OWTs.Earthquake analysis requires proper quantification of the expected hazard scenarios and the structural model. Aleatory and model uncertainty caused by earthquake variability and idealisation of structural parameters (mass, stiffness, damping) can affect the reliability of the computed response. OWTs are dynamic complex structures due to large rotor-nacelle-assembly (RNA), slender tower-support structure, and flexible foundations. The dynamic behaviour can differ in different operational conditions, and rotor directions. The current practice considers OWTs as a symmetric system based on the tower geometry while ignoring the importance of RNA. A point-mass rigid RNA is often used, suppressing the local blades modes, and affecting the higher mode response in OWTs. There are explicit instructions on the determination of the design earthquake, its selection procedure, and the minimum required number for seismic simulations. Criteria by local building codes are adopted, resulting in misleading seismic demand with the impression that OWTs may not be critical to earthquake loads.
This thesis aims at addressing these issues, starting with investigating various finite element models (FEM) for nonlinear dynamic analysis of OWTs. A qualitative baseline criterion is presented for choosing an appropriate numerical model based on desired analysis type (global or local) and computational efficiency without significant loss on accuracy. Given an initial model selection, various RNA configurations are studied to understand the implications of its simplifications on the dynamic response of wind turbine. RNA models with or without rotor eccentricity, asymmetry, blades’ rotary, and flexibility are considered. A generalised closed-form solution is discussed to estimate the simplified mechanical properties of blades for their explicit modelling. The blade properties are optimised using a genetic algorithm to achieve the target modal response.
Performance-based earthquake engineering (PBEE) is employed to estimate the seismic vulnerability of OWTs under pulse-like and non-pulse like records. The choice of an efficient IM is scrutinized using a probabilistic linear regression model (cloud analysis) for serviceability (SLS) and ultimate limit states (ULS). The effects of wind turbine size, earthquake characteristics, and RNA parameters on the dynamic behaviour, tower failure mechanisms, and failure probabilities are investigated. The results reveal that the interplay between vertical accelerations and the higher mode response can cause tower damage (buckling) at multiple locations. Seismic fragilities showed the high vulnerability of OWTs in shallow crustal regions, even at low-to-moderate intensity levels. It is pointed out that larger wind turbines can be more vulnerable to earthquakes due to greater RNA mass, tower flexibility and more active higher-mode response. It is also shown that the lack of accurate blade consideration can underestimate the failure probabilities of OWTs. Results are intended to be used as a baseline, using which more detailed seismic design guidelines could be developed.
| Date of Award | 28 Sept 2021 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Anastasios Sextos (Supervisor) & Raffaele De Risi (Supervisor) |