Signal-level clustering of acoustic emission streaming
Résumé
Clustering of acoustic emission signals aims at interpreting the dynamical materials behaviour through solicitations. It is usually posed as problem of finding clusters with good shape and well separated in the feature space, where features are extracted from AE signals. We here propose an alternative which does not use such features. The methodology relies on anomaly detection and includes a criterion that optimises the spread of onsets of clusters in place of shape.