Search

Mesoscale simulations of woven composites using parameterized analytical geometries offer a way to connect constituent material properties and their geometric arrangement to effective composite properties and performance. However, the reality of as-manufactured materials often differs from the ideal, both in terms of tow geometry and manufacturing heterogeneity. As such, resultant composite properties may differ from analytical predictions and exhibit significant local variations within a material. We employ mesoscale finite element method simulations to compare idealized analytical and as-manufactured woven composite materials and study the sensitivity of their effective properties to the mesoscale geometry. Three-dimensional geometries are reconstructed from X-ray computed tomography, image segmentation is performed using deep learning methods, and local fiber orientation is obtained using the structure tensor calculated from image scans. Suitable approximations to composite properties, using analytical unit cell calculations and effective media theory, are assessed. Our findings show that an analytical geometry and sub-unit cell geometry provide reasonable predictions for the effective thermal properties of a multi-layer production composite.

Characterizing the connection between material design decisions/parameters and their effective properties allows for accelerated materials development and optimization. We present a global sensitivity analysis of woven composite thermophysical properties, including density, volume fraction, thermal conductivity, specific heat, moduli, permeability, and tortuosity, predicted using mesoscale finite element simulations. The mesoscale simulations use microscale approximations for the tow and matrix phases. We performed Latin hypercube sampling of viable input parameter ranges, and the resulting effective property distributions are analyzed using a surrogate model to determine the correlations between material parameters and responses, interactions between properties, and finally Sobol' indices and sensitivities. We demonstrate that both constituent physical properties and the mesoscale geometry strongly influence the composite material properties.

Image-based simulation, the use of 3D images to calculate physical quantities, fundamentally relies on image segmentation to create the computational geometry. However, this process introduces image segmentation uncertainty because there is a variety of different segmentation tools (both manual and machine-learning-based) that will each produce a unique and valid segmentation. First, we demonstrate that these variations propagate into the physics simulations, compromising the resulting physics quantities. Second, we propose a general framework for rapidly quantifying segmentation uncertainty. Through the creation and sampling of segmentation uncertainty probability maps, we systematically and objectively create uncertainty distributions of the physics quantities. We show that physics quantity uncertainty distributions can follow a Normal distribution, but, in more complicated physics simulations, the resulting uncertainty distribution can be both nonintuitive and surprisingly nontrivial. We also establish that simply bounding the uncertainty can fail in situations that are sensitive to image segmentation. While our work does not eliminate segmentation uncertainty, it makes visible the previously unrecognized range of uncertainty currently plaguing image-based simulation, enabling more credible simulations.