Automated model updating of multiple cracked cantilever beams for damage detection. (November 2017)

• Record Type: Journal Article
• Authors: Altunışık, Ahmet Can; Okur, Fatih Yesevi; Kahya, Volkan
• Is Part Of: Journal of constructional steel research. Volume 138(2017)
• Page Start: 499
• Page End: 512
• Publication Date: 2017-11
• Subjects: Ambient vibration test -- Automated model updating -- Cracked beam -- Damage detection -- Finite element method; Steel, Structural -- Periodicals; Building, Iron and steel -- Periodicals; Acier de construction -- Périodiques; Construction métallique -- Périodiques; 624.1821
• DOI: 10.1016/j.jcsr.2017.08.006
• Languages: English
• ISSNs: 0143-974X

Abstract

Abstract: This paper presents a detailed study on the automated model updating of a cantilever beam with box cross-section including multiple cracks for damage detection. To consider multiple crack effects, a cantilever beam with cracks is considered under six damage scenarios. Finite element models are first constituted in ANSYS software for numerical solution. The results are validated by experimental measurements. For this aim, ambient vibration tests are performed to extract the dynamic characteristics using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification (SSI) methods. Calculated and measured natural frequencies and corresponding mode shapes for undamaged and damaged beams are compared. For damage detection, automated Model Update is carried out using the modal sensitivity method based on the Bayesian parameter estimation to minimize the difference between the calculated and measured results. Modal Assurance Criterion (MAC) and Coordinate Modal Assurance Criterion (COMAC) factors that show the correlation between the measured and calculated mode shapes are also employed to identify and locate damaged elements within the beam. Highlights: Investigation on box sectional steel cantilever beam with multiple cracks is presented. Ambient vibration tests are conducted to extract the dynamic characteristics. Damage identification is studied for undamaged and damaged conditions. Model updating procedure is considered for damage detection.