Options
Reliability based design of unstiffened fibre reinforced composite cylinders
Citation Link: https://doi.org/10.15480/882.1380
Publikationstyp
Doctoral Thesis
Date Issued
2017-04-01
Sprache
English
Author(s)
Advisor
Referee
Title Granting Institution
Technische Universität Hamburg
Place of Title Granting Institution
Hamburg
Examination Date
2016-12-14
TORE-DOI
Number in series
13
Citation
TUTECH Verlag, ISBN 978-3-946094-14-2, 2017
Publisher
TUTECH Verlag
Unstiffened lightweight cylinders are especially interesting for the space industry where high requirements in terms of controlling uncertainties and weight have to be met. Due to lack of experimental data, a deterministic design method based on a data basis containing only metallic cylinders is currently employed. Metallic cylinders show different characteristics as compared to cylinders made of carbon fibre reinforced plastics (CFRP) and hence the applicability of this method is restricted and the margin of safety achieved is unknown at early design stages.
Within this thesis, 11 CFRP cylinders of equal layup are tested in axial compression and extensive measurements are taken on the micro, meso and macro level. These are used to develop a probabilistic design method that incorporates material, test and model related uncertainties. After identifying relevant parameters through a sensitivity analysis, a Bayesian framework is set up to update those parameters as soon as new measurements are available. These are used as input for a Monte Carlo analysis to compute a distribution function of the load bearing capacity of the structure. From this distribution function a safety factor covering material- and structural related uncertainties is calibrated. By employing a Bayesian multiplicative error model an additional safety factor is derived to address the uncertainty related to this method.
The developed method offers the possibility to calibrate safety factors in dependence of the required reliability level and the a priori knowledge about material properties. Thus, a less conservative design can be pursued.
Within this thesis, 11 CFRP cylinders of equal layup are tested in axial compression and extensive measurements are taken on the micro, meso and macro level. These are used to develop a probabilistic design method that incorporates material, test and model related uncertainties. After identifying relevant parameters through a sensitivity analysis, a Bayesian framework is set up to update those parameters as soon as new measurements are available. These are used as input for a Monte Carlo analysis to compute a distribution function of the load bearing capacity of the structure. From this distribution function a safety factor covering material- and structural related uncertainties is calibrated. By employing a Bayesian multiplicative error model an additional safety factor is derived to address the uncertainty related to this method.
The developed method offers the possibility to calibrate safety factors in dependence of the required reliability level and the a priori knowledge about material properties. Thus, a less conservative design can be pursued.
Subjects
buckling
reliability
CFRP
probabilistic analysis
Bayesian updating
DDC Class
600: Technik
Loading...
Name
170511_Schillo_Dissertation_2.pdf
Size
11.02 MB
Format
Adobe PDF