Answers are grounded in this paper and cite their sources.
X-ray radiography measures particle densities and dynamics in granular systems, indicating new understanding for industrial applications.
In nature we observe avalanches, landslides and debris flows with disastrous destructiveness. From a physical point of view these events can be described as granular flows and research in this field is essential for an effective protection. The understanding of related systems is also important for industrial applications such as the flow behavior of particles in a fluidized bed reactor, or in the field of transportation and mixing of bulk cargo. Examples for industrial relevant goods are for instance charcoal, gravel, sand, coffee or pharmaceuticals. Despite their widespread occurrence, the dynamics of such systems are just poorly understood. In this work measurement techniques to investigate particle densities and dynamics of granular system are established. Such systems appear highly opaque to classical optical methods, restricting observations to boundary effects. In contrast, X-rays pass through matter in straight lines and allow a look inside. The X-rays are attenuated when they penetrate matter. High particle densities in the sample lead to a strong attenuation and result in a lower X-ray intensity, in comparison to low particle concentrations. The X-ray intensity behind the sample carries the integrated information of the particle density distribution in propagation direction of the X-rays. The final gray value in a radiogram is an integrated quantity, which makes a quantitative interpretation challenging. The full three-dimensional information of a sample can be accessed via X-ray tomography, but reconstructing a full tomogram requires several hundred radiograms captured from different directions. Classical X-ray setups, as used in scientific laboratories, are too slow to study dynamic granular systems via tomography. In the presented work, measurement methods are established, which allow an investigation of the particle densities and dynamics of granular systems utilizing X-ray radiography. The results are validated quantitatively. Subsequently, an overview to these methods is outlined. Correction of beam hardening in X-ray radiograms The intensity of monochromatic X-rays decays exponentially in dependence of the penetration length, when passing through a material. This is described by Beer-Lambert’s law. The thickness of the material can be calculated from the decrease of the X-ray intensity via the attenuation coefficient μ. However, classical X-ray tubes do not emit a monochromatic X-ray spectrum, but the polychromatic broad bremsstrahlung-spectrum. The attenuation coefficient depends on the photon energy. Photons with high energy are attenuated less than photons with low energy and the spectrum hardens, as the beam passes through a slab of material. Beam hardening must be considered in order to determine the penetration length of a polychromatic beam through the sample. Beer-Lambert’s law remains only valid, if μ is extended to an effective attenuation coefficient μeff (x), which considers the spectral change in dependence of the penetration length x. In this work, a method is presented quantifying μeff (x) by fitting a phenomenological model to measurement data in dependence of the material thicknesses x. Measurements are captured for varying X-ray settings and for multiple materials on two different setups. We compare the accuracy of our model function to other models from the literature. Among these, our model describes the measurement data with highest precision. This correction allows the determination of the penetration length from single radiograms. In the case of a granular sample the particle volume fraction Φ can be calculated from that. X-ray Digital Fourier Analysis We aim to measure the granular dynamics in a liquid-fluidized bed. For this purpose Digital Fourier Microscopy (DFM) has been extended to X-ray radiography. DFM is a Fourier image analysis technique, that extracts the intermediate scattering function (ISF) of a particle system from time averaged difference images. The ISF contains information on the structure and the dynamics of the particle system. Different model functions are used for the ISF depending on the particle dynamics, to quantify the mean particle motion in the system. DFM-analysis is well established for investigating colloidal or bacteria systems using light microscopy techniques. In comparison to the case described in this work, the imaging physics differs fundamentally: Here, we use X-ray radiography instead of light microscopy. We call this extension, from optical to X-ray imaging, X-ray Digital Fourier Analysis (X-DFA). The single steps of the analysis are performed analogous to the original DFM, but applied to X-ray radiograms. As a first approach to the new technique, we generated simulation based synthetic radiograms, on which the single steps of X-DFA were carried out. This works as a sample-system, helping to get familiar with the analysis and to study the effect of synthetic X-ray imaging. The specification of the particulate system, i.e. the dynamics, mean particle speed and number of particles, are input parameters to the simulations. This allows a parameter study with quantitative verification. We give a theoretical introduction on DFM, outline the single steps of the analysis on the simulation based images and compare its results to the expected outcome of the particle simulations. The values coincide within 2% relative difference. Validating X-DFA on a sedimenting suspension of granular particles In a next step we apply our findings on X-DFA to experimental data. For this purpose, we used an experimental model system, which allows a reliable reference measurement of the particle dynamics with well established methods, and can be compared to the X-DFA measure. We observe the sedimentation of particles in a liquid-fluidized bed of glass beads immersed in a glycerin-water mixture. The mean sedimentation velocity of the particles is measured in two ways: First, we track the position of the sedimentation front, which marks the upper edge of the particle bed, and calculate the sedimentation velocity from that. Second, we apply X-DFA to different regions of the particle bed in the radiograms. A first validation of X-DFA on experimental data is achieved by a comparison of both results, hvifront and hvixdfa . We perform experiments for different particle densities of the bed, 0.45 < Φ < 0.56, which leads to a range of sedimentation velocities from 0.04 mm/s to 0.13 mm/s. Both velocity measures coincide within a mean relative deviation of 9.4%. hvixdfa is systematically larger than hvifront . We address this systematic deviation to the contribution of a creeping boundary flow of particles at the walls of the system. The boundary flow is considered less when obtaining hvixdfa in comparison to hvifront , and leads to systematically larger values of hvixdfa . The thickness of the boundary layer can be estimated from the deviation between both velocities hvifront and hvixdfa. We get a thickness of ≈ 0.5 mm, which corresponds to ca. 3 particle diameters. X-DFA allows a quantitative study of the dynamics of opaque systems, that cannot be accessed with optical methods. In this work X-DFA is validated on a system of sedimenting particles. This technique opens the possibility of studying more complex particle dynamics in opaque systems, e.g. the granular dynamics in a fluidized bed or complex multiphase flows.
No takes yet. Share an insight, caveat, or question.
Manuel Baur (2020) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: