Loading...

Modal Identification of Structures with Phase-Based Motion Magnification and Upsampled Cross Correlation Techniques

Poorghasem, Sina | 2020

457 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53376 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Rahimzadeh Rofooei, Fayaz; Mahdavi, Hossein
  7. Abstract:
  8. Civil engineering structures and infrastructures including buildings and bridges are exposed to various externally applied loadings such as traffic, wind and earthquake during their operational lifetime. In this regards, structural static/dynamic displacement monitoring and their charactristics by mounted transducers is the mostly utilized and popular strategy. This may provide quantitative information for both structural safety evaluation and maintenance purposes. Such practice, however, is highly expensive to be operateddue to wiring, power signal transmission issues, limitations of sensor configuration to capture the entire structural response, and so forth. In addition, this may affect the dynamic characteristics by adding mass or stiffness of such sensors, especially in lightweight structures. In order to overcome afformentioned problems, noncontact vision-based displacement measurement systems have been developed and implemented recently, which are primarily enabled by the template matching techniques. The upsampled cross correlation technique which is based on a vision system, can provide the dynamic displacement time-histories of a structure with subpixel accuracy. However, the upsampled cross correlation technique is limited by camera resolution and intrinsic noise of the acquired images. Moreover, exciting a structure with only a small amount of energy may result in a subtle structural response that is hardly perceptible by observing the raw data of the conventional optical measurement systems (i.e., upsampled cross correlation). In this thesis, at the first step, a low frame rate video of the vibrating 4 story lab-scaled structure is taken when forced into a vibration. The performance of the upsampled cross correlation technique is evaluated by comparing displacements of each degree of freedom with accelerometers data in both time-domain and frequency-domain. Afterwards, the phase-based motion magnification and upsampled cross correlation techniques, are adopted in conjunction with metaheuristic optimization algorithms to automatically perceive the resonant frequencies and operating modal shapes corresponding to higher frequencies of the 4 story benchmark structure. Moreover, the upsampled cross correlation (UCC) technique, resonant frequencies and modal shapes of the structure are recovered quantitatively and are validated by accelerometer measurements. It is cocluded that, the proposed synthesis approach could be very practical in dealing with higher frequencies, in which that, higher frequencies are satisfactorily identified and corresponding mode shapes are confidently captured. The reliability of the results provided by phase-based motion magnification and upsampled cross correlation is demonstrated for identification of the structural modal parameters. It is also included that the proposed algorithm is promising enough in dealing with real applications, and the identified vibrational modal shapes and resonant frequencies can be valuable information for design enhancement, as well as damage detection, finite element model updating, and model validation
  9. Keywords:
  10. Template Matching ; Modal Identification ; Meta Heuristic Algorithm ; Phase-Based Motion Magnification ; Upsampled Cross Correlation

 Digital Object List

 Bookmark

No TOC