Computer-vision aided structural vibration tracking and analysis

Groot, Justin de (2024) Computer-vision aided structural vibration tracking and analysis. Research Project, Industrial Engineering and Management.

Preview	Text mIEM2024deGrootJ.pdf Download (46MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (131kB)

Abstract

Structural health monitoring (SHM) traditionally relies on accelerometers, but these come with practical limitations due to localized measurements and significant operational expenses. A new trend in SHM is non-contact video-based vibration analysis, allowing for comprehensive field analysis. This study investigates a novel deep-learning model named CoTracker for its efficacy as a comprehensive non-contact SHM tool. By recording the vibrations of a cantilever beam using a camera, various structural parameters are evaluated. A time-embedding DMD algorithm is employed for this analysis, and the results are then compared with an analytical solution and a FEM study by Cheng et al [10]. Findings reveal that CoTracker excels in measuring structural vibrations, showing a correlation of > 0.99 with accelerometer readings. Moreover, CoTracker identifies natural frequencies with a 0.7% error compared to accelerometer data. Both the analytical solution and FEM study affirm that CoTracker can extract the first two modal shapes, underscoring its significant potential as a full-field non-contact SHM algorithm.

Item Type:	Thesis (Research Project)
Supervisor name:	Cheng, L.
Degree programme:	Industrial Engineering and Management
Thesis type:	Research Project
Language:	English
Date Deposited:	18 Jul 2024 14:01
Last Modified:	25 Jul 2024 11:48
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/33478

Actions (login required)

View Item