Skip to Main content Skip to Navigation
Theses

Probabilistic single fibre characterisation to improve stochastic strength modelling of unidirectional composites

Abstract : Fibre reinforced composite materials are widely used for high-performance and critical structural applications.The design of composite material structures can be assisted by computational models for predicting their mechanical properties at both component and structural level. These models require very accurate constituent properties as input to make reliable predictions about the strength and lifetime of composite structures. However, a good understanding of the constituent properties of fibre-reinforced polymer composites is still lacking. For e.g., despite their wide use, very large differences are observed in the literature for properties of T700 carbon fibers.This thesis aims to advance the understanding of the constituent properties of fibre reinforced composites. To achieve this, an extensive experimental and statistical study has been conducted to understand the variations in the tensile strength of fibres and their morphology. These results have been analysed to identify and evaluate the critical parameters which contribute to errors or uncertainty in determining the parameters of the fibre strength Weibull distribution. This knowledge of uncertainties in constituent properties has also been used to highlight the variabilities they introduce on the output structural behaviour predicted by composite strength models. To further improve the fibre strength characterisation process, a data-reduction technique based on Bayesian approach of using prior knowledge of deficiencies in experimental data has been developed to model the tensile strength variation of brittle fibres. Additionally, new directions to accurately characterize the in-situ microscale properties of the matrix have been proposed.
Document type :
Theses
Complete list of metadata

https://pastel.archives-ouvertes.fr/tel-03272250
Contributor : ABES STAR :  Contact
Submitted on : Monday, June 28, 2021 - 11:18:10 AM
Last modification on : Thursday, April 28, 2022 - 11:28:51 AM
Long-term archiving on: : Wednesday, September 29, 2021 - 6:28:00 PM

File

2020UPSLM066_archivage_ISLAM.p...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03272250, version 1

Citation

Faisal Islam. Probabilistic single fibre characterisation to improve stochastic strength modelling of unidirectional composites. Mechanics [physics]. Université Paris sciences et lettres, 2020. English. ⟨NNT : 2020UPSLM066⟩. ⟨tel-03272250⟩

Share

Metrics

Record views

128

Files downloads

26