Optimization of Piles Supporting Monopile-Based Offshore Wind Turbines by Improved Foundation Models
MetadataVis full innførsel
With the focus on renewable energy sources over the last decades, offshore wind has become a popular source of energy harvesting. Among the different foundation concepts, the monopile is, by far, the most favourable for offshore wind applications (Page et al. 2018). However, a great extent of literature have indicated that the industry practice on monopile foundation design for offshore wind turbines (OWTs), fails to accurately predict the pile behaviour. Thus, excessive costs by overly conservative geometrical solutions are seen in the industry, and cost reductions in the reliability of the foundation design has been recognized as crucial for a further development. This thesis presents a study on the optimization potential of monopile OWT foundations, though using a more reliable macro-element foundation model, as an alternative to the industry practise of applying API p - y curves. Integrated time-domain simulations in 3DFloat has been used to simulate the load- and displacement response of the OWT, and the optimization potential has been assessed based on fatigue estimates. The focus has been on monopile-based OWTs situated at clay-dominated sites. By an assessment on the fatigue damage at the mudline, the macro-element model obtained an estimated fatigue life of 89.8% longer than the API p - y model. New geometries were suggested for the macro-element model to achieve similar fatigue damage estimates as the API p - y model, for geometry optimization. This resulted in a potential of steel savings on the monopile of 10 - 17%, by thickness reduction alone. Furthermore, the thesis includes a comparison with the different foundation models to measured data of an OWT installed in the North Sea. This analysis also includes an alternative p - y model, with curves extracted from finite element analyses (FEA). The macro-element model was seen to accurately predict all measured natural frequencies of the support structure, with a maximum deviation of 0.3%. In contrast, the industry practice of applying API p - y curves, under predicted all of the tower-bending frequencies by more than 10%. The FEA p - y model also provided good estimates on the measured natural frequencies. A fatigue damage assessment comparing the results from the macro-element model to the FEA p - y model was conducted to investigate the effect of soil damping on the fatigue estimates. A longer life expectancy of 29% was obtained for the macro-element model, and it was realized that neglecting soil damping (as the p - y models do), may limit the optimization potential of the monopile design. It was concluded that, by use of the macro-element model in foundation design and optimization, large potentials for cost reductions in the industry may be achieved. This was recognized both with regards to potential material savings, as the model predicted significantly less fatigue damage, as well as costs savings due to the greater reliability of the predictions, and consequently lowering risks.