Clustering acoustic emission activities in concrete using unsupervised pattern recognition methods

Tayfur S., Tayfur S., Ercan E. , Alver N.

Hezarfen-International Congress of Science, Mathematics and Engineering, İzmir, Turkey, 8 - 10 November 2019, pp.1-10

  • Publication Type: Conference Paper / Full Text
  • City: İzmir
  • Country: Turkey
  • Page Numbers: pp.1-10


Acoustic Emission (AE) is one of structural health monitoring methods used in different fields of engineering to detect defects such as cracks and leakages. The method is based on detection of elastic waves released from local sources in a stressed material. In civil engineering, by means of AE, location, type and orientation of the damage in concrete are obtained utilizing different algorithms. Pattern recognition, which is a subfield of artificial intelligence based on classifying objects, is also a proper tool for identifying types of AE activities. In this study, AE activities obtained from cracking sources of concrete material were clustered using two approaches of unsupervised pattern recognition: k-means and Gaussian Mixture Model. The results were evaluated and compared with each other to reveal effectiveness of these two processes.

Keywords: pattern recognition, acoustic emission, concrete, k-means, Gaussian mixture model.