par Deraemaeker, Arnaud 
;Preumont, André ;Kullaa, Jyrki
Référence IMAC-XXIV(30 January - 2 February 2006: Saint-Louis, Missouri, USA), 24th Conference and Exposition on Structural Dynamics 2006, Society for Experimental Mechanics, Bethel, Connecticut
Publication Publié, 2006
;Preumont, André ;Kullaa, JyrkiRéférence IMAC-XXIV(30 January - 2 February 2006: Saint-Louis, Missouri, USA), 24th Conference and Exposition on Structural Dynamics 2006, Society for Experimental Mechanics, Bethel, Connecticut
Publication Publié, 2006
Publication dans des actes
| Résumé : | This paper presents a methodology to perform automated damage detection based on output-only vibration measurements using large arrays of sensors. The methodology combines different techniques. For feature extraction, modal filtering together with wavelet analysis and peak detection are used. Variability of the structure due to environmental effects is modeled and removed using factor analysis. The technique relies on a long time monitoring of the intact structure. Finally, multi-variate control charts are used in order to detect deviation from the normal condition. The methodology is fully automated. The efficiency of the method is illustrated on a numerical example of a three-span bridge. |
