| dc.description.abstract | Offshore wind energy plays an important role in sustainability-focused international policies and experiences one of the fastest growth rates of all renewable energy sources. Although the cost of offshore wind energy has decreased dramatically in the last years, further cost reduction can be achieved by improving the accuracy of the analysis tools used in the design.
The design of offshore wind turbines (OWTs) relies on time-domain integrated load analyses tools that simulate the response of the entire OWT under combined aerodynamic and hydrodynamic loading. The numerical modelling of the foundation is an essential part due to its impact on the global dynamics; however, it is poorly represented in the current industry practice. In particular, it fails to accurately reproduce the foundation stiff ness and hysteretic damping observed in pile tests and in OWT field measurements.
The present PhD-work proposes novel foundation models that can provide accurate foundation stiffness and hysteretic damping in integrated time-domain simulations of OWTs, and demonstrates their impact on the dynamic structural response. It focuses on monopile foundations, the most common support structure and foundation solution for bottom-fixed OWTs.
The developed foundation models follow the macro-element approach, where the response of a pile and the surrounding soil is condensed to a force-displacement relation at seabed. The models are formulated as multi-surface plasticity models with kinematic hardening, and they represent a new application of this framework. The model formulation is based on results from 3D finite element analyses (FEA) of the soil volume and the pile, supported by symmetry conditions. The foundation
models are verified against cyclic FEA and validated against large-scale pile tests, providing good agreement. It is concluded that they are suitable to represent the foundation response in integrated analyses of OWTs because: (1) they model the non-linear load-displacement response and incorporate hysteretic foundation damping observed in pile tests; (2) they include the e
ect of multidirectional loading in the computed response; (3) they allow for a simple and exible calibration based on FEA which has a physical interpretation; and (4) they simulate the pile and soil behaviour with almost the same level of accuracy as FEA, but with a considerable reduction in computational effort. This is important for practical applications, where the response in time-domain analyses has to be computed for thousands of load cases. The foundation models are implemented as dll (Dynamic Link Library) and are freely available for download. Up to date, they have been incorporated into two integrated analyses tools (3DFloat and SIMA).
In addition, the impact of the foundation stiffness, damping and the e
ects of long term cyclic loading on the dynamic structural response has been investigated and quantified. Focus is set on the fatigue life, since fatigue is often the design-driver for monopile-based OWTs. The results indicate that the foundation stiffness and damping can have a substantial impact on the dynamic response and the fatigue life of OWTs. This is specially important during idling cases, where aerodynamic damping is small.
Moreover, the impact of the foundation model on the predicted response has been analysed by comparing simulations and full-scale field measurements of an OWT installed in the North Sea. The comparison indicates that the developed foundation models agree better with the measured response than the current industry practice in foundation modelling. In particular, the developed model calibrated to FEA matches the natural frequency (with an error lower than 1%), and overpredicts fatigue damage equivalent loads by approximately 16%. This contrasts with the current industry practice, which underpredicts the natural frequency by 11%, and overpredicts the fatigue damage equivalent loads by 160%. A more accurate modelling of the foundation response in integrated analyses of offshore wind turbines will reduce the uncertainty in the predicted loads and in the estimated fatigue life and therefore reduced risk in the design. | nb_NO |
