- This event has passed.
November 25, 2020 @ 12:00 pm - 1:00 pmFree – $120.00
Cost-Effective Infrastructure Monitoring Using Data-Driven Methods:
Application of signal processing and deep learning-based computational models
Infrastructure monitoring and condition assessment is integral to asset management. Large‐scale civil infrastructure such as buildings, bridges, wind turbines, and pipeline systems are exposed to various external loads throughout their lifetime. Vibration caused by earthquakes, wind, temperature, or human‐made excitation initiates structural deterioration during their service lives and subsequently triggers catastrophic failure if not monitored continuously.
Structural Health Monitoring (SHM) is an emergent and powerful diagnostic tool for damage detection and disaster mitigation of large‐scale structures. The traditional SHM methods use global responses such as vibration and local responses such as strains or a combination of both to assess the structure during in‐service conditions or extreme climatic events. Most of these techniques primarily rely on acceleration measurements that require the installation of either contact or noncontact sensors collecting rich quality of data. The vibration-based acceleration measurements can be used for any structure irrespective of their material or topography.
Sandeep Sony proposes advanced signal processing methods and deep learning-based computational tools to detect, localize, and quantify the severity of the damage in the structures. An approach is also developed to minimize the number of sensors required to create a financially viable option for infrastructure monitoring with limited sensors. A stadium prototype laboratory structure and a well-known Z24 bridge dataset are used to validate the proposed algorithms.
- Discover how to utilize vibration-based data for effective structural health monitoring.
- Understand the difference between parametric and non-parametric methods for structural health monitoring.
- Learn about various advanced signal processing methods to identify and visualize damage in structures.
- Get acquainted with deep learning-based methods to classify various types of damage.
Sandeep Sony, Ph.D. Candidate at Western University SET BACK TO 2020(11/25/2021)
This initiative is offered through the Municipal Asset Management Program (MAMP), which is delivered by the Federation of Canadian Municipalities (FCM) and funded by the Government of Canada.