Dual-indicator early damage warning framework for composite wind turbine blade: integrating acoustic emission energy drift and frequency cluster evolution

This study presents a comprehensive acoustic emission (AE)-based structural health monitoring framework applied to a full-scale 51.5-m wind turbine blade containing an artificial cross-shaped crack. A staged crack extension procedure under variable-amplitude fatigue loading was employed to realistically simulate damage progression. Fourteen resonant AE sensors continuously monitored transient signals associated with matrix cracking, delamination, interfacial debonding, and fiber breakage. A robust multi-stage noise filtering pipeline was developed to ensure data integrity. Two physically interpretable early-warning indicators were introduced: a normalized energy drift index, tracking time-domain energy deviations, and the Jensen–Shannon distance, quantifying changes in peak-frequency cluster distributions. Results demonstrated that the dual-indicator framework reliably identifies both sustained damage progression and sudden shifts in failure mechanisms. Post-test inspections verified strong correspondence between AE-based predictions and actual crack propagation zones. This research bridges the gap between coupon-scale laboratory studies and practical, full-scale blade monitoring, providing a physically meaningful approach to early damage detection in large composite structures.