Parametric identification via ambient vibration signals for mechanical systems

Abstract

Random signals from environmental vibrations enable determination of system parameters without knowing input force In civil engineering system parameters such as stiffness and damping ratio of structures can hence be approximately estimated through output-only recordings The same idea can be applied to measure mechanical properties of nano-scale objects under thermal fluctuations In this study natural frequencies and damping ratios are used as the mechanical indicators of the structures Finite element software COMSOL was adopted to simulate a cantilever beam under Gaussian white noise to demonstrate the methodologies such as the random decrement (Randomdec) method Furthermore oscillators with one and two degrees of freedom were analyzed for validation In addition the displacement responses of the Kao-Ping-Hsi cable-stayed bridge from field tests were analyzed with the random decrement technique and Ibrahim time domain (ITD) Both methods are able to produce similar natural frequencies and damping ratios with the assistance of autoregressive vector (ARV) analysis

Date of Award	2014 Aug 29
Original language	English
Supervisor	Yun-Che Wang (Supervisor)