LAUSR

Abstract In this article, the elastic wave propagation phenomenon is applied as the basis for a system designed to detect anomalies in the pre-tensioned connections of engineering structures. A series of laboratory tests were carried out on a steel frame. This approach, where the connection is a fragment of the complex structure, and not an isolated part, enabled us to analyse behaviours similar to those in real engineering structures. The problem of the similarity of signals corresponding to different types of anomalies was considered. Thanks to the application of principal component analysis and a procedure employing a combination of artificial neural networks, the extraction of the most significant features of the registered signals was possible. The proposed procedure proved to be very sensitive and accurate for the proper detection of anomalies, as well as for the determination of their location and type, even if the loosening of bolts was very slight (i.e., invisible to the naked eye). The most important element of the presented approach is the fact that the combination of the assessment of the condition of local parts along with the application of artificial intelligence techniques can result in a Structural Health Monitoring (SHM) system that allows for the global assessment of structural integrity.