Papers by Keyword: Quantitative Image Analysis

Abstract: The work pertains to research of inhomogeneity level of the composite compacts manufactured from flaky metal particles and polymer binder. In the present investigation, magnetically hard Nd-Fe-B powders were used. The composite matrix was in all cases polystyrene. Volume fraction of the powders in composites was 40 ÷ 54%. As the forming method injection moulding was used. Granulates were obtained from powders and polystyrene dissolved with toluene. Produced cylindrical specimens had 10 mm in diameter and 4 mm in height. The structure of specimens was determined on the basis on quantitative analysis of the micrographs of axial sections. The analysis of binary images of metallographic specimens allowed to determine homogeneity and directions of flake distribution. Standard deviation of surface fraction is a measure of inhomogeneity which shows relatively well both local and general differences in the distribution of particles. The greater the volume of powder V_p in composites, the lower standard deviation S. The size of particles also influences the value of standard deviation. The greater the size of particles in a composite, the greater the values of S. In the next research, information about inhomogeneity will be useful in estimation of distribution of mechanical and magnetic parameters of the compacts (dielectromagnets) formed in this way.

Abstract: Syntactic functionally graded metal matrix composites (SFGMMC) are a class of metallic foams in which closed porosity results from the presence of hollow ceramic microspheres (microballoons), whose spatial distribution varies continuously between the inner and the outer section of the part, thus resulting in a continuous variation in properties. In this work, aluminiumbased SFGMMC rings were fabricated by radial centrifugal casting. The graded composition along the radial direction is controlled mainly by the difference in the centrifugal forces which act on the molten metal matrix and the ceramic particles, due to their dissimilar densities. In this case where the density of the SiO2-Al2O3 microballoons is lower than that of molten aluminium, the particles show a tendency to remain closer to the inner periphery of the ring. Thus the microballoon volume fraction increases along the radial direction of the ring from the outer to the inner periphery; in other words, the particle-rich zone is limited to an inner layer of the ring. Precursor conventional MMCs were prepared by stir-casting from the constituent materials, by homogeneously dispersing commercial SiO2-Al2O3 microballoons (particle size: 50 µm; particle volume fraction: 5 and 10 %) within a molten commercial Al-7Si-0.3Mg (A356) alloy. The resulting MMCs were then re-melt and centrifugally cast in order to produce the functionally graded composites. Particle gradients in the centrifugally cast composites were investigated by quantitative image analysis of optical micrographs (for the estimation of the particle volume fraction, mean particle diameter and porosity volume fraction).

Abstract: The macrosegregation is one of the most recurrent phenomena in continuous steel casting of slabs. It often impedes the further processing and has detrimental effect on the quality of rolled strip. The macrosegregation can be explained by the interpretation of unit steps of solidification in continuous casting slabs as well as on the basis of the interaction of the inclusions at the interface of solid and liquid phases. Development of solid shell and interaction between growing dendrites and MnS segregation has been presented. Only the pushing effect could be the real cause of MnS segregation among the three possible interactive mechanisms. The sulphur print could give misleading information on sulphur segregation.

Abstract: In the present paper, work of a preliminary nature pertaining to the development of a simple tool for easy identification of reinforcement clustering in particle-reinforced MMCs is reported. The tool consists of an image processing and analysis routine performed with the aid of a public-domain software. Results obtained were in the form of distribution curves pertaining to the Euclidian distances between particle centroids within sample images. Analysis of that distribution allowed the identification of cluster occurrence. The technique was successfully applied to computer-generated images of virtual MMCs, as well as to micrographs of rheocast MMCs. Apart from particle clustering, the distribution curves were found useful to roughly assess the matrix cell size, for cases where this is significantly higher that the size of the reinforcing particles.