Nonlinear Kalman filtering based damage quantification for civil infrastructure

Simple item page
dc.contributor.authorGhorbani, Esmaeil
dc.contributor.examiningcommitteeEl-Salakawy, Ehab (Civil Engineering) Muthukumarana, Saman (Statistics) Zhou, Wenxing (Civil and Environmental Engineering, University of Western Ontario)en_US
dc.contributor.supervisorCha, Young-Jin (Civil Engineering) Svecova, Dagmar (Civil Engineering)en_US
dc.date.accessioned2021-05-11T20:17:35Z
dc.date.available2021-05-11T20:17:35Z
dc.date.copyright2021-05-01
dc.date.issued2021-04-30en_US
dc.date.submitted2021-05-01T18:06:36Zen_US
dc.degree.disciplineCivil Engineeringen_US
dc.degree.levelDoctor of Philosophy (Ph.D.)en_US
dc.description.abstractDespite advancements in Kalman filtering and the dozens of published works on the subject, the performance of Kalman filtering in terms of structural damage quantification is limited to structural systems with low dimension state vectors. Furthermore, Kalman filters cannot accurately identify damage in structures with unknown excitations. Therefore, This dissertation introduced two methods to overcome these fundamental limitations. The first method is called the iterated modified cubature unscented Kalman filter (IMUKF). It is developed by a careful combination of sigma points deriving from the unscented Kalman filter (UKF) and cubature points from the Cubature Kalman filter (CKF) and adding iterative convergence criterion to the algorithm. Extensive numerical analysis reveals that the IMUKF outperforms the traditional nonlinear Kalman filtering by increasing the number of unknown parameters. The second method eliminates the uncertainties related to the unknown excitations by extracting the free vibration response of the underlying structure. This idea introduces an output-only damage quantification method. The free vibration response is fed into the UKF, which estimates the unknown states without any need to input information for damage quantification. This proposed output-only method, referred to as RD-UKF, is evaluated using numerical and full-scale experimental data. Bridge scour is a leading cause of catastrophic bridge failures in North America. In order to accurately estimate the level of scour, a new numerical model for a bridge pier is derived based on the Euler beam theory. The new model is able to simultaneously consider soil-structure interaction, bridge bearings, deck mass, water force, and traffic loading. The model is then integrated with the proposed RD-UKF method for bridge scour quantification. In addition to the numerical simulation studies, the method is implemented on a real bridge in the province of Manitoba, Canada. The estimated soil height is verified with two different independent bathymetries in two different seasons. Results of the numerical and experimental investigations indicated that eliminating input excitation from the measured response decreases the uncertainty in the estimation process. Furthermore, it shows that the introduced RD-UKF method can be used for damage quantification in different structures without any information about the input excitation.en_US
dc.description.noteOctober 2021en_US
dc.identifier.urihttp://hdl.handle.net/1993/35508
dc.language.isoengen_US
dc.rightsopen accessen_US
dc.subjectStructural health monitoringen_US
dc.subjectKalman filteren_US
dc.subjectStructural damage identificationen_US
dc.subjectScour monitoringen_US
dc.subjectRandom decrementen_US
dc.subjectOutput-only structural damage quantificationen_US
dc.titleNonlinear Kalman filtering based damage quantification for civil infrastructureen_US
dc.typedoctoral thesisen_US
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
PhD Thesis_Esmaeil Ghorbani.pdf
Size:
7.65 MB
Format:
Adobe Portable Document Format
Description:
PhD thesis, Esmaeil Ghorbani
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
2.2 KB
Format:
Item-specific license agreed to upon submission
Description:
Download
Collections
FGS - Electronic Theses and Practica