Identification, reduced order modeling and model updating of nonlinear mechanical systems

Kurt, Mehmet

Permalink

https://hdl.handle.net/2142/50609 Copy

Description

Title

Identification, reduced order modeling and model updating of nonlinear mechanical systems

Author(s)

Kurt, Mehmet

Issue Date

2014-09-16

Director of Research (if dissertation) or Advisor (if thesis)

Vakakis, Alexander F.

Doctoral Committee Chair(s)

Vakakis, Alexander F.

Committee Member(s)

• Bergman, Lawrence A.
• McFarland, Donald M.
• Alleyne, Andrew G.

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Nonlinear System Identification
• Nonlinear Model Updating
• Reduced-Order Modeling

Abstract

In this dissertation, we propose a new method for global/local nonlinear system identification, reduced order modeling and nonlinear model updating, applicable to a broad class of dynamical systems. The global aspect of the approach is based on analyzing the free and forced dynamics of the system in the frequency-energy domain through the construction of free decay or steady-state frequency-energy plots (FEPs). The local aspect of the approach considers specific damped transitions and leads to low-dimensional reduced order models that accurately reproduce these transitions. The nonlinear model updating strategy is based on analyzing the system in the frequency-energy domain by constructing Hamiltonian or forced and damped frequency-energy plots (FEPs). These plots depict the steady-state solutions of the systems based on their frequency-energy dependencies. The backbone branches, branches that correspond to 1:1 resonances, are calculated analytically (for fewer DOFs) or numerically (e.g., shooting method). The system parameters are then characterized and updated by matching these backbone branches with the frequency-energy dependence of the given system by using experimental/computational data. The main advantage of our approach is that we do not assume any type of nonlinearity model a priori, and the system model is updated solely based on numerical simulations and/or experimental results. As such, the approach is applicable to a broad class of nonlinear systems, including systems with strong nonlinearities and non-smooth effects, as will be shown in this dissertation. For larger scale systems, model reduction techniques (e.g., Guyan reduction) are applied to construct reduced order models of the system to which the aforementioned methods are applied.

Graduation Semester

2014-08

Permalink

http://hdl.handle.net/2142/50609

Copyright and License Information

Copyright 2014 Mehmet Kurt

Owning Collections