Wind energy is a growing industry, and in an effort to reduce costs and increase turbine efficiency, rotor blades are becoming increasingly large in size. To facilitate this effort, the SmartBlades2 research project has designed, built, and tested a set of prototype research blades. As part of the SmartBlades2 project, high sensor density modal testing has been conducted on the research blades. The analysis of the modal tests showed good agreement of the global vibration modes with the finite element model predictions. However, the test analysis also identified low frequency vibration modes, referred to as breathing modes, which were not predicted by the finite element models. These vibration modes were found on all of the blades and are characterised by out-of-plane trailing edge panel motion. The objective of this thesis is to identify and predict the aforementioned breathing modes using finite element analysis. To achieve this, three model characteristics are analysed to determine their influence on the breathing mode prediction, namely, model topology, shell element configuration, and material properties. To characterise the affect of model topology, a cut section from the SmartBlades2 prototype blade is modelled with shell elements and continuum element glue joints. To validate the blade section model, a modal test is conducted which identifies breathing modes analogous to the full blade. Various topology features are investigated with the focus on the shell glue joints and spar web joints of the blade section. The analysis shows that while these changes significantly effect the mode shapes and frequencies, none of them predict the experimentally identified breathing modes. To investigate the source of this discrepancy, the modal behaviour of a sample plate structure with the same materials is used to remove the variability of topology. The effects of shell element size and configuration are analysed with mutual comparisons. The analysis shows that higher fidelity element configurations offer no advantage over linear shell elements for prediction of modal behaviour, while the element size shows higher sensitivity. Furthermore, the effects of material properties are examined using the sample plate, subject to modal and flexural tests. It is found that the specified properties are stiffer than measured, and new predictions of the properties are made which better fit the plates experimental results. Finally, the topology, element, and material investigations are then applied to an improved finite element model of the complete blade and correlated with the experimental modal tests. It is found that the improved blade model has closer correlation with the experimental modal tests for global modes, however is unable to predict the identified breathing modes for the blade. It is hypothesised that cause of this may relate to the connection of the spar web with the glue flanges.

The reconstruction of the elastic deformed shape of a structure from strain measurements is a field which has received considerable attention over the years. This work aims to suggest some improvements for wing-like structures trying to limit as much as possible the amount of strain measurements needed. In particular, a simple beam model is proposed based on the framework of the inverse Finite Element Method (iFEM). Then, the performances of iFEM using shell elements will be enhanced pre-extrapolating the strain field and the results will be compared with another shape sensing method, the so-called Modal Method (MM). In the final part of the work the external loads under the form of a pressure field are recovered using both the reconstructed displacements and directly the strain measurements. Both static and dynamic analyses will be carried out, so recovering the load both in space and in time. The results obtained show that the beam model developed allows to obtain a satisfactory bending reconstruction of the structure, while the twist is not always accurate. Computing the full displacement field with iFEM brings to a relatively good representation, even though not as satisfactory as the one delivered by the Modal Method. Finally, recovering the static external loads directly from the strain measurements seems to perform better compared to the reconstruction from the full displacement field, but it is significantly affected by noise and uncertainties. The dynamic load reconstruction is in general much more challenging and the results obtained often show a significant error compared to the reference solution.