How does “listening” help fracture understanding?

Rosemere de Araujo Alves Lima, R. Tao, A. Bernasconi, Michele Carboni, Nicolas Carrere, S. Teixeira De Freitas

Research output: Contribution to conferencePosterScientific


Understanding the relationship between the sensors’ outputs and the damage evolution within the joints is becoming increasingly crucial to improving structural health monitoring systems and collecting data to improve the joint’s design. Therefore, a study of the acoustic emission method associated with visual fracture evaluation was proposed to give insights into the toughening of composite bonded joints and better understand the relationship between the acoustic emission features and the damage mechanism involved. Thus, two different layups were proposed for the substrates: [0]8 and [0/902/0]S. In addition, a toughened epoxy adhesive with an embedded carrier (AF163-2k) was used to bond the substrates. Five specimens of each stacking sequence were tested under quasi-static mode I loading conditions. A travelling microscope and a regular digital camera were used on the lateral sides of the specimens to track the crack propagation paths. One piezoelectric sensor linked to the AMSY-6 Vallen system was used to assess the acoustic emission features produced within the joints during the tests. Unsupervised machine learning algorithms based on artificial neural networks and the Morlet continuous wavelet transformation were used to pattern recognition of the acoustic emission data. Self-organising maps, together with k-means algorithms, were used for data clustering. Following that, the acoustic emission features of each cluster were associated with the insights obtained from the crack propagation images. Finally, it was observed that the different layups triggered simultaneous toughening mechanisms. The combination of the acoustic emission and the visual evaluation was crucial for a deeper understanding of the underlying phenomena.
Original languageEnglish
Number of pages1
Publication statusPublished - 2023
Event 3rd Benelux Network Meeting and Workshop on Damage and Fracture Mechanics - Antwerp, Belgium
Duration: 25 May 202325 May 2023
Conference number: 3


Conference 3rd Benelux Network Meeting and Workshop on Damage and Fracture Mechanics
Abbreviated titleBDFM-2023
Internet address


  • acoustic emission
  • Toughening mechanisms
  • Adhesively bonded joints
  • Artificial Neural Networks


Dive into the research topics of 'How does “listening” help fracture understanding?'. Together they form a unique fingerprint.

Cite this