VUB researchportal: https://researchportal.vub.be/en/publications/practical-frontiers-in-applied-model-based-virtual-strain-sensing/Persistent thesis: https://hdl.handle.net/20.500.14017/4e20e183-d47a-4492-b4c3-0b1ea647c10f Offshore wind turbines are typically designed for 25 years of operation while enduring continuous wind and wave loading. Over time, these repeated forces cause progressive material degradation—known as fatigue—which inevitably will compromise the assets' structural reliability. Accurately predicting fatigue accumulation is therefore essential to the design process. However, due to modelling uncertainties and site-specific complexities, predictions differ from the actual behaviour observed in the field. Monitoring the structural response at critical locations is therefore key to improving confidence in these predictions, yet direct measurements in important regions such as below the seabed are often impractical or prohibitively expensive. This thesis investigates how fatigue accumulation can be estimated at inaccessible locations by combining limited sensor data from accessible parts of the turbine with structural models based on design documentation. This approach, known as model-based virtual sensing, enables strain and fatigue estimation without requiring physical sensors at every location. The work demonstrates that improving model fidelity—particularly in the modelling of soil–structure interaction—substantially enhances the accuracy of virtual sensing results. Validation on multiple turbines confirms that newer PISA-based foundation models yield significantly better fatigue estimates than conventional approaches. Additional studies address the role of blade flexibility, and the feasibility of minimal sensor configurations. By addressing critical challenges in modelling, data quality, and validation, this thesis advances the practical use of virtual sensing for structural health monitoring in offshore wind. It opens pathways to more reliable, cost-effective, and calable monitoring strategies.