(149g) Time-Resolved X-Ray Tomography of a Fluidized Bed

Research on fluidization and other multiphase reactors is handicapped by the opaqueness of these multiphase systems. The well-known laser-based techniques used in single phase flow (Laser Doppler Anemometry, Particle Image Velocimetry, etc.) are of very limited use. This holds especially for fluidized systems, where the particle phase is dense and light based techniques can not be used. As alternative a number of techniques has been developed, of which the tomographic ones have caught quite some attention over the last decade. The advantages of tomographic techniques are obvious: they provide 2D or even 3D information without disturbing the flow. The tomographic techniques can be classified according to the underlying measurement technique. On the one hand there are the electrical techniques, like Electrical Capacitance Tomography (ECT, Dyakowski et al, 2000) or Electrical Impedance Tomography (EIT). These are characterized by high speed: in a 2D version easily several hundreds of image per second can be produced. This is attractive, as that allows a study of the transient nature of the flow, e.g. the passage of voids or bubbles can be easily followed. However, the measurements rely on so-called soft fields: a perturbation anywhere in the system is influencing all measurements. This makes tomographic reconstruction of the measured data very difficult. Consequently, the spatial resolution of these techniques is its weak point. On the other hand, there are the nuclear techniques based on γ-radiation or X-rays (Kumar et al 1995). These are so-called hard fields. Here, the radiation penetrating through the object is only influenced by what is directly on the path traversed by the radiation. However, these techniques require relatively long measuring times, making them less suitable for studying transient behaviour. In the present contribution, a tomographic system will be presented that is based on X-rays and that has the capability of measuring images with a frequency of several hundred images per second (Mudde et al 2005). The system is made of three medical X-ray sources, that simultaneously send X-rays through a 23cm-diameter bubbling fluidized bed. The set up is sketched in fig.1. The fluidized bed has a diameter of 23cm and is filled with polystyrene particles (0.56mm diameter, Geldart type B). Each X-ray source generated a fan beam that is detected by 32 detectors per fan. Medical X-ray sources (with maximum photon energy of 150keV) are used together with CdWO4 detection crystals (Hamamatsu, 10mm diameter). The radiation is attenuated upon passing through the fluidized bed, according to the Lambert-Beer law: I(x)/I0 = exp(-∫ μ(x)dx) in which I(x) is the intensity of the radiation at a distance x from the source, I0 is the initial intensity and μ is the attenuation coefficient which is depending on the absorbing material and the X-ray energy. A problem with using X-rays compared to γ-radiation is that X-rays are not monochromatic but have a wide energy spectrum, making Lambert-Beers law more complicated to use. Proper calibration can circumvent this as will be discussed in the paper. The measured beam attenuation can directly be turned into the mean chordal void fraction, , called the ray sum, pi. As the attenuation is a local phenomenon, the ray sum is the integral effect along the beam path. This can easily be discretized using pixels, resulting in the system with ?Ñk the averaged void fraction of pixel k and Wik the length of the beam i through pixel k. For the reconstruction the iterative method SART (from the Algebraic Reconstruction Technique family) is used. An example of static phantoms is shown in fig.2 (from a 5-source set-up): the gray circles are the original phantoms; the black pixels are the reconstructed ones. This shows that a spatial resolution of 5mm is achievable with a measuring frequency of 200 frames/s. Further experiments were done on a moving egg-shaped object. The egg is hollow with a thin shell of expanded polystyrene. It is moved through the bed at velocities ranging from 10 to 60cm/s. The shape and diameter can be reconstructed with good accuracy. Finally, the 3-source set-up is used to measure passing bubbles in the fluidized bed. The system can measure several seconds continuously and captures the passage of the bubbles. Figure 3 shows a snapshot of a passing bubble in the fluidized bed. Our system is capable of measuring several seconds consecutively, generating a movie of the passing bubbles. In the contribution we will high light the strong and week points of the scanner and show some results in the form of movies. It will be concluded that an X-ray based tomographic scanner has both the spatial and temporal resolution to study the void distribution in fluidized beds. This technique allows the study of what goes on inside the fluidized beds. References 1. Dyakowski, T., Jeanmeure, L.F.C., Jaworski, A.J., Applications of electrical tomography for gas-solids and liquid-solids flows - A review, Powder Technol. 112, 174-192, 2000. 2. Kumar, S.B., Moselmian, D., Dudukovic, M.P., A gamma-ray tomographic scanner for imaging voidage distribution in two-phase systems, Flow Meas. Instrum., 6,1, 1995. 3. Mudde, R.F., Bruneau, P.R.P., van der Hagen, T.H.J.J., Time-resolved gamma-densitometry imaging within fluidized beds, Ind. Eng. Chem. Res. 44, 6181-6187, 2005.