PSI - Issue 2_A

M. Wicke et al. / Procedia Structural Integrity 2 (2016) 2643–2649 M. Wicke et al./ / Structural Integrity Procedia 00 (2016) 000–000

2645

3

3.1. 3D reconstruction of casting pores by μ -CT X-ray computed tomography (x-CT) is a non-destructive and non-invasive imaging method using radiographs, which are recorded from different viewing directions, to reconstruct the internal structure of a volume. Due to the spatial and contrast resolution involved here, the use of high-resolution X-ray CT, also termed micro-computed tomography (μ-CT), is required as these technique can be exploited for reconstructing sub-millimeter casting defects. On the basis of 2D frontal scanning of the samples, the internal reconstruction is performed by reconstructing the cross-sectional images (i.e. slices) using a back projection algorithm. All measurements were carried out with ZEISS Xradia Versa 520 X-ray microscope, equipped with a 180 kV / 9 W ultrahigh performance nanofocus X-ray tube at a voltage of 70 kV / 6 W. 1601 radiographs taken over 360° rotation were acquired with a resolution of 1.0 μm and an exposure time of 2.5 s per image. In order to reconstruct the morphology of the internal defects, pores within the differently stacked slices provided by μ-CT (Fig. 1a), whose number is dependent on the volume analyzed as well as the slice-to-slice distance, were automatically identified first. Due to the proportionality of material density and grey level of each pixel, the distinction whether a pixel belongs to the material or a cavity is possible. Fig. 1b illustrates the identification of a pore, which appears as a dark spot in a light background representing the surrounding matrix. After applying an evolution algorithm, a 3D reconstruction of the pores was obtained. The outer surface of the tortuous 3D shape, which results out of the stacking of the pore cross sections, was then smoothed as shown in Fig. 1c using Gaussian smoothing filters with a standard deviation of 0.8 and the geometry finally exported using the STL (STereoLithography) standard.

Fig. 1. Scheme for pore reconstruction: (a) stack of slices; (b) pore identification in each slice; (c) smoothed pore surface

The volume investigated contained numerous pores of various sizes and geometries. These pores are typically classified either as gas pores, which exhibit a comparatively spherical geometry, or as shrinkage pores, characterized by highly irregular shapes with multiple branches as they result from the volume contraction in course of the solidification. Three shrinkage pores obtained by the cross-sectional reconstruction are depicted in Fig. 2.

Fig. 2. Shrinkage pores of highly variable shapes revealed by μ-CT