A Single-Sensor attenuation fingerprint mapping framework for autonomous acoustic emission source identification in composite wind turbine blades

This research article presents a single-sensor correlation-mapping framework (SCMF) that conducts intelligent frequency-domain analysis for acoustic emission (AE) source identification in glass-fibre composite wind turbine blades without relying on traditional multi-sensor time-of-arrival methods. This structural health monitoring (SHM) method decomposes registered AE signals into narrow partial power (PP) bands to capture unique attenuation fingerprints across different frequency bands. The SCMF primarily comprises of two key components: an unsupervised learning pipeline and a cosine-similarity-based mapping. The former is applied to narrow PP bands to cluster AE signals into distinct damage or impact source zones within the blades. The latter employs zone acoustic vectors and cosine-similarity, enabling precise and noise-resilient AE-source identification. Results demonstrate that SCMF provides robust and accurate SHM performance for composite wind turbine blades using only a single sensor. The method has the potential to be scalable, transferable, data-efficient, and sensor-minimal, offering a promising pathway towards next-generation intelligent SHM of complex engineering structures and systems.