Reduced-order model-inspired system identification of geometrically nonlinear structures

Abstract

There is a growing interest in more aesthetic-looking and slender engineering structures. These slender structures can undergo large-amplitude nonlinear vibrations, causing an unpredictable behaviour in their response. To predict their behaviour, often an accurate model of the nonlinear system is required. In this thesis, robust identification of geometrically nonlinear slender structures is addressed, by defining accurate mathematical models for them.
Accurate and robust mathematical models are derived for geometrically nonlinear systems, specifically those with large inertia, by taking inspiration from reduced-order modelling.
These models are used for Nonlinear System Identification (NSI). This inspiration is due to the similarity of the reduced-order modelling and system identification fields, i.e., usually both are used to replicate the dynamics of a system using a mathematical model with low complexity. Data from a Finite-Element (FE) model of a cantilever-type beam structure is used in system identification. It is shown that using the ROM-inspired model
in system identification improves the accuracy of the predicted response in comparison to a standard nonlinear model. This highlights the importance of defining accurate models
for NSI.
The research is then extended to experiments. Data is gathered from an experimental setup of a laboratory-scale cantilever-type beam, and the NSI approach developed in simulations is applied. The method is shown to be superior to the traditional nonlinear mathematical models. To account for the damping, a nonlinear mapping of non-conservative damping force is considered in the ROM-inspired model. The decay response of the identified ROM-based model closely matched the measured decay response.
Thus, the nonlinear effects of damping in the structural system is well-captured.
Finally, nonlinear FE model updating using the ROM-based model is discussed. The mapping between the estimated nonlinear parameters and the physical parameters of an FE model is demonstrated. Accurate FE parameter values are obtained using the most robust ROM-inspired parameters.

Date of Award	18 Jun 2024
Original language	English
Awarding Institution	University of Bristol
Supervisor	Tom L Hill (Supervisor), Simon A Neild (Supervisor) & Jason Zheng Jiang (Supervisor)