Damage identification on a large-scale wind turbine rotor blade using sample-based deterministic model updating

Wind turbine rotor blades are among the most critical components of wind turbines, with their structural integrity directly affecting reliability, lifetime, and maintenance costs. Reliable damage identification is therefore essential for structural health monitoring (SHM) strategies in wind energy applications. In this context, the updating of numerical models represents an established method for vibration-based non-destructive damage identification, including damage detection, localization, and quantification. Naturally, the model-updating process is affected by different sources of uncertainty. On the one hand, the numerical model always represents an idealization that introduces unavoidable discrepancies between its basic assumptions and reality. On the other hand, the measurement data and identified modal parameters, typically serving as damage-sensitive features, are subject to uncertainty. Despite extensive research on uncertainty quantification and propagation in model updating, comparative studies of model-updating procedures applied to large-scale structures, particularly wind turbine rotor blades, remain scarce. Moreover, the level of model fidelity and the impact of different design variable configurations associated with the selected numerical model are seldom examined in the context of model updating, typically formulated as an optimization procedure. This study addresses this gap by systematically evaluating how model fidelity and design variable parameterization influence the model-updating results while considering input uncertainty associated with the measurement data and identification process. The investigations are conducted using measurement data from a 31 m rotor blade subjected to edgewise fatigue loading. A comparison of the results shows that all design variable configurations yield consistent results, confirming the robustness of the presented model-updating procedures. Model fidelity, however, strongly influences the outcomes, with higher accuracy and detail leading to distinctly improved damage identification.