Observability, Identification and Lie Symmetries in Structural Health Monitoring Problems

Add to your list(s) Download to your calendar using vCal

• Prof Manolis Chatzis, Oxford University
• Friday 08 November 2024, 16:00-17:00
• JDB Seminar Room, CUED.

If you have a question about this talk, please contact div-c.

Non-linear dynamic systems arise naturally in several applications either due to the underlying constitutive laws and geometric non-liearities, or simply because of how a researcher approaches a problem.

A special category of non-linear systems is that of non-smooth dynamics systems, i.e., systems whose equations of motion, or measurement equation, include terms that are not infinitely differentiable with respect to some of their states. Such systems often arise in engineering and are often related to some form of damage, e.g., sliding, impact, hysteresis or fracture.

The accurate modeling of non-smooth systems becomes an important task because of the connection of such behaviors to failure and the increased inherent uncertainty. An added challenge to Structural Health Monitoring applications is that the excitations to systems representing infrastructure elements are often impossible to measure, e.g., as in the case of the wind excitations to bridges and wind turbines.

An important means to tackle the previous challenges and reduce the effects of other sources of uncertsainty is through system identification, i.e., estimating the behavior and properties of the system using data obtained from sensors.

Prior to using an algorithm for identifying the system it is worth to investigate if this would be possible even under ideal conditions, i.e. investigating the theoretical observability and identifiability of the system. In situations where the system is not observable for an examined sensor setup, it is also interesting to investigate what are the equivalent transformations of the dynamic states that one may be tracking, i.e. the Lie Symmetries of the system. In doing so, the non-smooth nature of a system, the high-dimensionality of the models corresponding to real-life applications, and the common reality of not measuring all of the inputs pose challenges.

In this seminar, developments on the Observability, Lie Symmetries and Identification of non-linear systems is discussed. Improvements on Observability algorithms to handle large linear- and non-linear systems and unknown inputs are presented. Motivating examples from structural systems including bridges and offshore wind turbines are provided.

Short Bio: Manolis Chatzis is an Associate Professor in the Department of Engineering Science at the University of Oxford and a Tutorial Fellow for Hertford College. Prior to joining Oxford in 2013 Manolis was a Post-Doctoral Research Scientist at Columbia University in the City of New York. He holds a Diploma in Civil Engineering and an MSc. in Structural Engineering from the National Technical University of Athens. He was awarded a PhD from Columbia University in the City of New York in 2012. His research focuses on the study of non-linear dynamic systems with emphasis in developing algorithms for their modeling and identification.

This talk is part of the Engineering - Dynamics and Vibration Tea Time Talks series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge University Engineering Department Talks
• Cambridge talks
• Centre for Smart Infrastructure & Construction
• Civil Engineering Talks
• Computational Continuum Mechanics Group Seminars
• Engineering - Dynamics and Vibration Tea Time Talks
• Engineering - Mechanics, Materials and Design (Div C) - talks and events
• Featured lists
• Interested Talks
• JDB Seminar Room, CUED
• School of Technology
• Trust & Technology Initiative - interesting events
• bld31

Note that ex-directory lists are not shown.