Predicting non - linear bending behaviour of ultra - thin woven fibre composites

Abstract

Ultra thin woven composites are extensively used in deployable space structures, particularly on deployable booms which are responsible to deploy and hold key components in space missions. Due to high weight sensitivity of these applications, it is essential to achieve the maximum structural e ciency to reduce the payload. However the exural behaviour of these thin textile composites is still troublesome under high curvatures. Hence it limits the optimization of deployable structures to highest degree possible. Numerical modelling of these structures is considered as a promising tool in designing, considering the time consuming and costly nature of physical testing. Yet, most of the numerical models aimed at the macroscopic behaviour, su er from lack of accurate behavioural characteristic of non-linear geometric regime. This study is an attempt made to address the above problem by building virtual simulation techniques through micromechanical modelling. For this work a homogenized Kirchho Love plate model was developed with the identi ed unit cell of two-ply plain weave composite. The geometry was imported from TexGen textile modelling package and FEA simulation was done by ABAQUS commercial nite element package. A new logical framework was proposed to describe the behavioural characteristics of the tows at the interlacing points by means of cohesive behaviour. Material de nition for cohesive interaction was included through traction separation law maximum principal stress criterion for damage initiation. Required traction coe cients were extracted by a discrete FEA simulation due to unavailability of experimental data. The developed model was executed in the linear regime and then extended to non-linear geometric regime to predict the exural behaviour under high curvatures and it shows bending sti ness reduction as expected. Thus the proposed simulation technique can be utilized in designing process of deployable booms made of thin woven composites through the multiscale modelling approach after verifying the accuracy with experiments.