Ilya Straumit

Cluster analysis of acoustic emission signals for 2D and 3D woven carbon fiber/epoxy composites:

Understanding the failure mechanisms in textile composites based on acoustic emission signals is a challenging task. In the present work, unsupervised cluster analysis is performed on the acoustic emission data registered during tensile tests on two-dimensional and three-dimensional woven carbon fiber/epoxy composites. The analysis is based on the k-means+ + algorithm and principal component analysis. Peak amplitude and frequency features – peak frequency for two-dimensional woven composites and frequency centroid for three-dimensional woven composites – were found to be dominant in cluster analysis. Cluster bounds were identified for both reinforcement types. These bounds do not differ for both reinforcement types and can be used as a starting point for acoustic emission analysis of other carbon fiber/epoxy composites. The statistics of high-frequency acoustic emission events are compared with the estimates obtained from a fiber bundle model based on Weibull fiber strength statistics. The number of acoustic emission events agrees well with the number of groups of carbon fibers that fail simultaneously. This finding may provide a new way to explain why the Weibull distribution predicts much more fiber breaks than measured by acoustic emission.

Internal Structure of the Sheared Textile Composite Reinforcement: Analysis Using X-Ray Tomography

X-ray micro computed tomography (Micro-CT) is a non-destructive technique that can provide information on the internal structure of materials. The purpose of micro-CT is to assess the presence of defects as well as characterizing internal structures and potential damage present in the produced part. Simple shear is an interesting deformation mechanism for woven fabric draping. The internal structure change of the carbon fibre twill fabric after shear deformation is chosen as a subject of this paper. Parameters of the mesoscopic internal structure of the woven fabric like cross section, shape, area, and middle line coordinates can be obtained from micro-CT images through image processing procedures. Details of the image data processing for sheared fabric cross sections are discussed. This paper illustrates the possibilities of micro-focus computer tomography in materials research, namely for defining geometrical properties of textile. Image processing is also used for the recognition of fibre direction in the yarns. Described methodology can be applied for determining structure of a fabric, and the results can be used for further micromechanical modelling. Identification of the fibres orientation is important for estimation of the mechanical properties of composites and can be achieved with image processing techniques.

Analysis and Segmentation of a Three-Dimensional X-ray Computed Tomography Image of a Textile Composite

Microfocus X-ray computed tomography allows obtaining highly detailed three-dimensional images of inspected objects. Regarding textile composites, resolution of this technique is enough to distinguish individual fibres. For the purpose of modelling, the micro-CT image of a composite must be segmented in order to separate materials components. This paper presents results of application of structure tensor and first-order statistics to compose a feature vector and segment the image. Results show that, depending on the choice of the variables used in the segmentation, the image can be segmented into the matrix, yarns and voids (pores) domains, or into the domains of matrix and yarns of different primary orientation.

Quantification of micro-CT images of textile reinforcements

VoxTex software (KU Leuven) employs 3D image processing, which use the local directionality information, retrieved using analysis of local structure tensor. The processing results in a voxel 3D array, with each voxel carrying information on (1) material type (matrix; yarn/ply, with identification of the yarn/ply in the reinforcement architecture; void) and (2) fibre direction for fibrous yarns/plies. The knowledge of the material phase volume and known characterisation of the textile structure allows assigning to the voxels (3) fibre volume fraction. This basic voxel model can be further used for different type of the material analysis: Internal geometry and characterisation of defects; permeability; micromechanics; mesoFE voxel models.Apart from the voxel based analysis, approaches to reconstruction of the yarn paths are presented.VoxTex software (KU Leuven) employs 3D image processing, which use the local directionality information, retrieved using analysis of local structure tensor. The processing results in a voxel 3D array, with each voxel carrying information on (1) material type (matrix; yarn/ply, with identification of the yarn/ply in the reinforcement architecture; void) and (2) fibre direction for fibrous yarns/plies. The knowledge of the material phase volume and known characterisation of the textile structure allows assigning to the voxels (3) fibre volume fraction. This basic voxel model can be further used for different type of the material analysis: Internal geometry and characterisation of defects; permeability; micromechanics; mesoFE voxel models.Apart from the voxel based analysis, approaches to reconstruction of the yarn paths are presented.