A fast reconstruction algorithm for time-resolved X-ray tomography in bubbling fluidized beds

Abstract

A new tomographic reconstruction algorithm is proposed for fast image reconstruction. The results are based on a high speed X-ray tomography system, consisting of 3 X-ray sources and 32 detectors for each source. The proposed algorithm combines void measurements of each X-ray beam into a triangular mesh, which is formed by the intersection points of all the beams. Simulations and real fluidized bed data are utilized to assess the quality of the proposed algorithm compared to the Simultaneous Algebraic Reconstruction Technique (SART). The influence of the number, position and diameter of the phantoms on the proposed reconstruction method is studied. The new method provides images with similar quality to SART reconstructions, although obtaining smaller bubble sizes. The low computing time needed to reconstruct each image with the new method, which is more than 5000 times faster than SART for a 40×40 mesh, encourages the use of the new method for the online image reconstruction of X-ray measurements.