Theoretical modeling and identification method for surface crack damage in wind turbine blades

Abstract Wind turbine blades are susceptible to damage from multiple factors during extended operation, and early detection of surface damage is critical to ensuring the stable performance of the turbine system. This study integrates fracture mechanics with full-scale fatigue test data analysis to propose an efficient theoretical model and identification method for surface crack damage in wind turbine blades. A discrete dynamic model of cracked blades is constructed based on the structural and dynamic characteristics of wind turbine blades. The breathing behavior of surface cracks under fatigue-induced vibration is simulated using a bilinear spring, resulting in a refined and more accurate dynamic model. Comprehensive mining and comparison of strain data across the blade’s entire service life are conducted to extract key damage features, including first-order natural frequency reduction, variations in strain coupling behavior, and shifts in mean strain values. Based on the extracted features, a Gaussian mixture model is constructed to identify surface cracks in wind turbine blades. In the experimental validation stage, blade design parameters such as segmented stiffness and mass were utilized to confirm the effectiveness of dynamic response deviations caused by cracks, and surface cracks were successfully identified in full-scale fatigue tests. The results demonstrate that the proposed method offers high accuracy and localization capability in the early detection of surface cracks, providing a novel technical pathway and theoretical foundation for the structural health monitoring of wind turbine blades. The work also contributes to the field of measurement science, applying data-driven techniques to improve real-time damage detection, and establishing a foundation for further research in non-destructive testing methods for wind energy systems.