Efficient parameter identification using generalized Polynomial Chaos Expansion – A numerical and experimental study

Research output: Chapter in Book/Report/Conference proceedingConference paper

Abstract

In order to validate numerical simulations, it is essential to compare calculated results with measurements. A common tool to increase the accuracy of the numerical simulation is model updating. The idea is to find better input parameters which describe the system more precisely by comparing the measurements and the numerical results, e.g. the modal properties. Finite element model updating is state-of-the-art and well described in literature. In this paper, an alternative computational method called Numerical Assembly Technique is used to calculate the natural frequencies, the mode shapes and the Frequency Response Functions. The generalized Polynomial Chaos Expansion is applied to increase the numerical efficiency of the model updating process. A stepped circular shaft is analyzed and it is shown, that the model updating process is not a straight forward technique. The input values of the parameter fit as well as the numerical model are limited. A solution which fits both criteria is described, and selected improvements are given.
Original languageEnglish
Title of host publicationProceedings of ISMA 2020 International Conference on Noise and Vibration Engineering and USD2020 International Conference on Uncertainty in Structural Dynamics
Pages1833 - 1846
Number of pages14
ISBN (Electronic)9789082893113
Publication statusPublished - 30 Oct 2020
Event ISMA 2020 International Conference on Noise and Vibration Engineering - Virtuell, Belgium
Duration: 7 Sep 20209 Sep 2020

Conference

Conference ISMA 2020 International Conference on Noise and Vibration Engineering
Abbreviated titleISMA 2020
CountryBelgium
CityVirtuell
Period7/09/209/09/20

Fingerprint

Dive into the research topics of 'Efficient parameter identification using generalized Polynomial Chaos Expansion – A numerical and experimental study'. Together they form a unique fingerprint.

Cite this