Compressive sensing spectral estimation for output-only structural system identification

TauSiesakul, B., Gkoktsi, K. & Giaralis, A. (2014). Compressive sensing spectral estimation for output-only structural system identification. Paper presented at the 7th Stochastic Computational Mechanics Conference, 15-18 June 2014, Santorini, Greece.

[img]
Preview
Text - Accepted Version
Download (380kB) | Preview

Abstract

In this paper a compressive sensing (CS), sub-Nyquist, non-uniform deterministic sampling technique is considered in conjunction with a computationally efficient power spectrum estimation approach for frequency domain output-only system identification of linear white noise excited structural systems. The adopted CS sensing spectral estimation approach assumes multi-band input random signals/stochastic processes without posing any signal sparsity requirements and therefore it is applicable to linear structures with arbitrary number of degrees of freedom and level of damping. Further, it applies directly to the sub-Nyquist (CS) measurements and, thus, it by-passes the computationally demanding signal reconstruction step from CS measurements. Numerical results pertaining to the acceleration response of a damped structure with closely-spaced natural frequencies are provided to demonstrate the effectiveness of the considered approach to provide reliable estimates of natural frequencies by means of the standard frequency domain peak-picking algorithm of operational modal analysis using up to 90% fewer measurements compared to the Nyquist rate sampled data. It is envisioned that this study will further familiarize the structural dynamics community with the potential of CS-based techniques for vibration-based structural health monitoring and condition assessment of engineering structures.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Compressive Sensing, Power Spectrum Estimation, Output-only System Identification, ARMA Filter, Multi-band Stationary Random Processes, Multi-coset sampling
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Divisions: School of Engineering & Mathematical Sciences > Engineering
URI: http://openaccess.city.ac.uk/id/eprint/13620

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

View more statistics