Modal Parameter Identification based on the Probability Distribution of Frequency Response Function
Project no. 32090 · Status: Kapanmış (Yayın bekleniyor) · 2024
Summary
<p>Within the scope of this project, it is aimed at developing a Frequency Response Function (FRF) based input-output method that is able to determine the modal parameters of structures through multiple measurement setups that include response measurements collected from different points according to various input locations. For this purpose, the conditional probabilistic distributions of the measured FRFs are obtained depending on the modal parameters (natural frequency, damping ratio, modal displacement shape) by modeling the residual between the FRFs obtained from the response measurements in each measurement group and their theoretical values as a measurement error. Then, the posterior probability distribution of the modal parameters is analytically derived by using Bayesian inference over the joint probability distribution of all measurement setups. In the next stage, the posterior statistical parameters (variance, standard deviation, coefficient of variation, etc.) of the modal parameters over the derived probability distribution is obtained. For verification purposes, two experimental studies have been conducted. At first, the model parameters of a simply supported beam model is investigated using the response measurements obtained from multiple measurement setups. In the second study, the modal parameters of a laboratory model of a 3D torsional shear frame structure have been evaluated using the acceleration responses collected from different locations in each measurement setup considering the effect of exciter acting at the different locations and directions. The obtained results reveal that the presented methodology can provide innovative solutions with rather accurate results.</p>