Vision-based bridge monitoring using displacement curvatures

Chidiebere Brendan Obiechefu, Fouad Mohammed, Zakwan Arab, Rolands Kromanis

Research output: Contribution to conferenceAbstractAcademic

35 Downloads (Pure)

Abstract

Resilience is about improved operational performance and safety by ensuring integrity and redundancy of infrastructure systems. Structural health monitoring (SHM) of old infrastructure like bridges is crucial to ensuring resilience. The fusion of affordable vision-based structural health monitoring (VBSHM) systems with effective damage detection techniques has the potential to provide cost-effective solutions to support condition assessments of such old bridges, - a necessary step towards ensuring their safety and consequently resilience. With VBSHM, distributed sensing along a bridge is attainable, after which image-processing can be used to obtain bridge response. One of such responses is curvature. The curvature technique involves fitting a curve to response from tracked targets, extracting their quadratic coefficients across all loading timesteps, and taking the maximum coefficient as bridge response. Feasibility of this technique is demonstrated on a numerical model of a truck-loaded bridge girder subjected to multiple damage scenarios. Noise is also added to replicate real-world scenarios. Damages can be detected and localised but are influenced by damage extent and measurement noise. The technique shows potential for field applications.
Original languageEnglish
Pages120-123
Number of pages4
Publication statusPublished - 2020
EventJoint International Resilience Conference, JIRC 2020: Interconnected: Resilience Innovations for Sustainable Development Goals - Online conference
Duration: 23 Nov 202027 Nov 2020

Conference

ConferenceJoint International Resilience Conference, JIRC 2020
Abbreviated titleJIRC
Period23/11/2027/11/20

Fingerprint

Dive into the research topics of 'Vision-based bridge monitoring using displacement curvatures'. Together they form a unique fingerprint.

Cite this