Model updating and damage detection of frame structures using output-only measurements / Hooman Monajemi

Hooman , Monajemi (2018) Model updating and damage detection of frame structures using output-only measurements / Hooman Monajemi. PhD thesis, University of Malaya.

[img] PDF (The Candidate's Agreement)
Restricted to Repository staff only

Download (1227Kb)
    [img]
    Preview
    PDF (Thesis PhD)
    Download (5Mb) | Preview

      Abstract

      Model based damage detection and localization is one of the most efficient non-destructive health assessment methods in civil structures. It is based on the fact that damage in the structure alters structural and consequently modal properties of the system. There are many ways of employing modal properties for damage detection, including reconstructing flexibility matrix using modal data i.e. modal flexibility matrix. Among various structural systems, utilizing this method in frame structures is relatively more difficult. The main reason is their complex geometry and subsequently complex flexibility matrix. The second concern is that mass normalized mode shapes are required to reconstruct flexibility matrix and they are not so easy to obtain, especially in operational modal testing. The third issue is incomplete measurements. In frame structures like jacket platforms for example, it is not possible to measure all degrees of freedom and it is important to find a solution to detect damages on unmeasured part of the structure. This study aims to address these concerns and other issues that are related to damage detection of frame structures using operational modal analysis (OMA). A scaled model of a steel frame structure was constructed and tested in the laboratory. The model was excited using two shakers that were mounted on top of the structure. Although the input forces were available, but they were just used to validate the results of operational modal analysis. Since input forces were assumed to be unavailable in OMA, an alternative scaling method based on change in mass was used to normalize mode shapes. Among various flexibility based detection methods, damage locating vectors were found to be one of the most suitable methods considering the complex geometry of the frame structure and were used as the primary detection method in this study. The method was tested by a number of damage scenarios and the results were showing that damage locating vectors (DLV) is always certain on indicating the undamaged members, but it sometimes fails to indicate the damaged member(s) with an acceptable certainty. To solve this problem, a second damage indicator was suggested based on cross model cross mode (CMCM) model updating method. The advantage of using this indicator was that it has the opposite type of error compare to DLV and so combining these two methods resulted on a more certain damage indicator. Cross model cross mode model updating was also used to address the problem of incomplete measurements by updating a finite element model of the frame structures in respect to the experimental data. Then each unmeasured member was damaged in updated FE model which provided a range of frequencies for each damage case. Comparing the calculated frequencies with the frequencies obtained from the experiment and also using the extra information provided by the measured DOFs, it was possible to approximate the location of the damaged member.

      Item Type: Thesis (PhD)
      Additional Information: Thesis (PhD) – Faculty of Engineering, University of Malaya, 2018.
      Uncontrolled Keywords: Model based damage detection; Operational modal analysis (OMA); Non-destructive health assessment methods; Complex geometry
      Subjects: T Technology > TA Engineering (General). Civil engineering (General)
      Divisions: Faculty of Engineering
      Depositing User: Mr Mohd Safri Tahir
      Date Deposited: 19 Oct 2018 08:16
      Last Modified: 19 Oct 2018 08:16
      URI: http://studentsrepo.um.edu.my/id/eprint/8563

      Actions (For repository staff only : Login required)

      View Item