Application of machine learning classifiers to elastic wave signal analysis and diagnostics of bolted joints

openaccess, Vol. 634 (6) 2025 / czwartek, 26 czerwca, 2025

Zastosowanie klasyfikatorów uczenia maszynowego do analizy sygnałów fal sprężystych oraz diagnostyki połączeń śrubowych

(Open Access)

DOI: 10.15199/33.2025.06.10

citation/cytuj: Bróż N., Nazarko P. Application of machine learning classifiers to elastic wave signal analysis and diagnostics of bolted joints. Materiały Budowlane. 2025. Volume 634. Issue 06. Pages 91-97. DOI: 10.15199/33.2025.06.10

A laboratory model of a portal frame was tested. Piezoelectric transducers were used to excite and measure elastic waves. As a result, 2,100 patterns were collected for various connection state scenarios (7 classes). Subsequently, a multi‑level diagnostic system for fault detection and classification was developed. After training, it achieved flawless classification results, offering an alternative to previously applied approaches for this task SSN, ANN. Furthermore, it was demonstrated that the significance of the principal components does not correspond to their normalised order.

Badaniom poddano laboratoryjny model ramy portalowej. Do wzbudzania i pomiaru fal sprężystych zastosowano przetworniki piezoelektryczne. W rezultacie zebrano 2100 wzorców różnych scenariuszy stanu połączenia (7 klas). Następnie zbudowano wielopoziomowy system diagnostyczny do wykrywania i klasyfikacji usterek. Po jego wytrenowaniu możliwe było uzyskanie bezbłędnych wyników klasyfikacji, co stanowi alternatywę dla wcześniej stosowanych w tym zadaniu podejść SSN, ANN. Ponadto wykazano, że istotność składników głównych nie jest tożsama z unormowaną ich kolejnością.
non‑destructive tests; elastic wave propagation; classification models; machine learning; sensitivity analysis.

badania nieniszczące; propagacja fal sprężystych; modele klasyfikacji; uczenie maszynowe; analiza wrażliwości.
  1. Cha YJ, You K, and Choi W. ‘Vision‑based detection of loosened bolts using the Hough transform and support vector machines’, Automation in Construction, vol. 71, pp. 181–188, 2016, DOI: 10.1016/j.autcon.2016.06.008.
  2. Chena R, Chena S, Yanga L, Wangb J, Xua X, and Luoa T. Looseness diagnosis method for connecting bolt of fan foundation based on sensitive mixed‑domain features of excitation‑response and manifold learning’, Neurocomputing, pp. 376–388, 2017, DOI: 10.1016/j.neucom.2016.09.041.
  3.  Yang J and Chang FK. Detection of bolt loosening in C–C composite thermal protection panels: I. Diagnostic principle, Institute of Physics Publishing, 2006, DOI: 10.1088/0964‒1726/15/2/041.
  4. Yang J and Chang FK., Detection of bolt loosening in C–C composite thermal protection panels: II. Experimental verification, Institute of Physics Publishing, 2006, DOI: 10.1088/0964‒1726/15/2/042.
  5. Stepinski T, Uhl T, and Staszewski W. Advanced Structural Damage Detection: From Theory to Engineering Applications. Wiley, 2013.
  6.  Yun CB and Min J. Smart sensing, monitoring, and damage detection for civil infrastructures, KSCE Journal of Civil Engineering, vol. 15, no. 1, pp. 1–14, Dec. 2010, DOI: 10.1007/s12205‒011‒0001‑y.
  7.  Su Z and Ye L., Identification of Damage Using Lamb Waves. From Fundamentals to Applications. Springer, 2009.
  8. Abdi H and Williams LJ. Principal component analysis, Wiley Interdisciplinary Reviews: Computational Statistics, vol. 2, no. 4, pp. 433–459, Jun. 2010, DOI: 10.1002/wics.101.
  9.  Hernandez‑Garcia MR and Sanchez‑Silva M. Learning Machines for Structural Damage Detection, in Intelligent Computational Paradigms in Earthquake Engineering, N. D. Lagaros and Y. Tsompanakis, Eds., Idea Group Publishing, 2007, pp. 158–187.
  10. Ziaja D, Turoń B, and Miller B. Detection of Anomaly in a Pretensioned Bolted Beam‑To‑Column Connection Node Using Digital Image Correlation and Neural Networks, Applied Sciences, vol. 10, no. 7, p. 2400, 2020, DOI: 10.3390/app10072400.
  11.  Ziaja D and Nazarko P. SHM system for anomaly detection of bolted joints in engineering structures, Structures, vol. 33, pp. 3877–3884, Oct. 2021, DOI: 10.1016/j.istruc.2021.06.086.
  12. Nazarko P. Structural diagnostics using elastic waves and artificial neural networks (in Polish). Publishing House of Rzeszow University of Technology, 2019.
  13. Nazarko P, Soft computing methods in the analysis of elastic wave signals and damage identification’, Inverse Problems in Science and Engineering, vol. 21, no. 6, pp. 945–956, 2013, DOI: 10.1080/17415977.2013.764295.
  14. Waszczyszyn Z and Ziemiański L., Neural networks in mechanics of structures and materials – new results and prospects of applications., Computers and Structures, vol. 79, pp. 2261–2276, 2001.
  15.  Nazarko P and Ziemiański L. Damage detection in aluminum and composite elements using neural networks for Lamb waves signal processing, Engineering Failure Analysis, vol. 69, pp. 97–107, 2016, DOI: http://dx.doi. org/10.1016/j.engfailanal.2016.07.001.
  16. Zima B and Rucka M. Detection of debonding in steel bars embedded in concrete using guided wave propagation, Diagnostyka, vol. 1, no. 3, pp. 27–34, 2016.
  17. Zima B and Rucka M., Guided waves for monitoring of plate structures with linear cracks of variable length, Archives of Civil and Mechanical Engineering, vol. 16, no. 3, pp. 387–396, May 2016, DOI: 10.1016/j.acme.2016.01.001.
mgr inż. Natalia Bróż, Politechnika Rzeszowska, Wydział Budownictwa, Inżynierii Środowiska i Architektury
ORCID: 0000-0002-2719-7566
dr hab. inż. Piotr Nazarko, prof. PRz, Politechnika Rzeszowska, Wydział Budownictwa, Inżynierii Środowiska i Architektury
0000-0002-6135-2486

mgr inż. Natalia Bróż, Politechnika Rzeszowska, Wydział Budownictwa, Inżynierii Środowiska i Architektury ORCID: 0000-0002-2719-7566

Correspondence address: n.broz@prz.edu.pl

Full paper:

DOI: 10.15199/33.2025.06.10

Article in PDF file

Received: 07.01.2025 / Artykuł wpłynął do redakcji: 07.012025 r.
Revised: 03.03.2025 / Otrzymano poprawiony po recenzjach: 03.03.2025 r.
Published: 24.06.2025 / Opublikowano: 24.06.2025 r.