Application of machine learning and artificial intelligence for structural damage detection

Structural Health Monitoring (SHM) aims primarily to accurately identify the current condition of a structure, assessing damage levels and potentially enabling the prediction of its future performance. These problems can be difficult to solve using conventional computational techniques alone, which creates big data opportunities for the application of artificial intelligence methodologies.

This paper presents the integration of Machine Learning (ML) for pattern recognition with SHM systems through the development of statistical models. A case study of a steel tower is presented, tested in a laboratory environment to simulate structural damage by loosening bolted connections.

AUTHORS | Filipe Santosᵃ, Eduardo Gonçalvesᵇ and Abel J.P. Gomesᶜ

ᵃ VESAM Engenharia S.A., Zona Industrial de Cantanhede, Lote 69, 3060-197 Cantanhede, Portugal, ᵇ MIRA SYSTEMS Lda, Quinta Vale do Espinhal, EM558 1, 3230-343 Penela, Portugal, ᶜ Universidade da Beira Interior, Departamento de Informática, 6200-001 Covilhã, Portugal

Paper presented at the XIV Congress on Steel and Composite Construction