On the material characterization of a composite using micro CT image based finite element modeling

摘要

Novel techniques for generating robust and accurate meshes based on 3-D imaging data have recently been developed which make the prediction of macro-structural properties of composite structures based on micro-structural composition straightforward. The accuracy of reconstructions is a particular strong point of these new techniques with geometric accuracy only contingent on image quality. Algorithms developed and used are topology preserving, volume preserving and multi-part geometric models can be handled straightforwardly. In addition to modeling different constituent materials as separate mesh domains, material properties can be assigned based on signal strength in the parent image thereby providing a way of modeling continuous variations in properties for an inhomogeneous medium. These new techniques have been applied to the analysis of a ceramic matrix composite which was micro-CT scanned and the influence of imaging parameters on both predicted bulk properties and localized stresses has been explored. This paper utilizes the Computed Tomography (CT) as the NDE technique to characterize the initial matrix porosity's locations and sizes in a Ceramic Matrix Composites (CMC) test specimen. Further, the Finite Element (FE) method is applied to calculate the localized stress field around these pores based on the geometric modeling of the specimen's CT results, using image analysis, geometric modeling and meshing software, ScanIP/ScanFE. The analyses will simulate experimental loading conditions where scanned specimens are then tensile tested to a 0.07% total strain to identify the matrix cracking locations in relation to the original pores. Additional work is carried out combining the image processing and finite element to investigate the applicability of some novel meshing techniques. Finally, the calculated Finite Element localized stress risers are compared with the observed matrix cracking locations. This work is expected to show that an FE model based on an accurate 3-D rendered model from a series of CT slices is an essential tool to quantify the effects of internal macroscopic defects of complex material systems such as CMCs.