Quantitative Identification Of Pore Defects In C/SiC Composites And Their Impact On Flexural Resistance

Ceramic Matrix Composites(CMCs)are a new type of lightweight material with high temperature resistance and have been widely used in hypersonic vehicle thermal protection systems,high thrust-to-weight ratio aero-engines,high-speed braking systems,and other fields.However,limited by the internal structure of the material and the complexity of the process,components with large size and complex shape will produce the problem of poor homogeneity of the material structure,which seriously affects the performance of the material.However,at present,the quality of the product is mainly judged by the apparent density in the process of material preparation,and the overall density cannot effectively reflect the homogeneity of the material structure.Based on this,this paper focuses on the material structural homogeneity,and takes the plain woven ceramic matrix composite heterogeneous members as the research object,and carries out the following research from the perspective of internal pore defects:Pore defects are an important constituent phase of fiber-reinforced composites and their distribution has an important impact on the material properties,in order to obtain their real condition inside the material high precision defect detection is performed using industrial CT.An adaptive dual-threshold OTSU algorithm combining median filtering and grayscale transformation is proposed for the problem of strip artifacts and uneven grayscale distribution in tomographic images of composite heterogeneous components,which leads to difficulties in extracting defects inside the material.The improved algorithm is compared with OTSU algorithm and Fuzzy C-mean class clustering segmentation algorithm,and the over-segmentation and under-segmentation phenomena are significantly improved,and the accuracy of pore defect segmentation is enhanced.Aiming at the problem that the overall porosity cannot effectively characterize the structural homogeneity of the material,a spatial cell segmentation method based on the pixel marking method is proposed to calculate the percentage of defects on each spatial cell by combining the segmentation results of internal defects,while the defect content is mapped in three dimensions using the light projection algorithm to intuitively present the defect content on each spatial cell of the heterogeneous member.Finally,the product of the standard deviation and the mean value of the unit porosity is used as the quantitative index of homogeneity to complete the intuitive quantitative characterization of the material structure homogeneity.As the local porosity is certain,the inhomogeneity will also exist on different spatial units due to different pore types.To address the problem that the local porosity cannot reflect the characteristics of internal defect shape,size and distribution position,the morphological characteristic quantity of defects is extracted by the connectivity analysis,and the classification identification and spatial localization of defects on spatial units are realized from two directions of parallel and perpendicular fiber layups by combining Kmeans and fractal dimensions.The corresponding finite element model is also established by combining the shape and spatial location information of defects,and the effects of different defect types on the flexural properties of the material interlayer are investigated by numerical simulations based on the material parameters obtained from the mechanical experiments of composite specimens with equal porosity.