Stiffness evaluation of polymer textile composites subjected to internal geometry variability

The variability in the mechanical properties of textile composites is investigated from virtual samples subjected to internal geometry variability. Each virtual specimen is generated using the WiseTex software given as input statistical information collected from experiments. These data are acquired in prior work by inspecting a unit cell sample using X-ray micro-computed tomography to identify the geometrical scatter. The tow path is statistically characterised in terms of non-stochastic periodic trends and non-periodic stochastic fluctuations. The statistical behaviour is expressed in terms of standard deviation and correlation length. In a next step, virtual samples are generated using these systematic trends added with deviations produced by a Markov Chain algorithm. The transition matrix of this procedure is calibrated with the statistical information. This methodology guarantees that the specimens possess the same statistical properties as the considered sample. By producing many virtual replicas and applying homogenisation, statistical data about the mechanical properties are obtained. In the current work, the statistics of the stiffness properties of a twill 2/2 woven carbon fabric in an epoxy matrix are analysed. Copyright 2013 by Aurora Flight Sciences.status: publishe