Generation and evaluation of an artificial optical signal based on X-ray measurements for bubble characterization in fluidized beds with vertical internals

Abstract

The performance of fluidized bed reactors strongly depends on the bubble behavior, for which reason knowledge concerning the bubble properties is important for modeling and reactor optimization. X-ray measurements can be used to characterize bubbles within the cross-section of a fluidized bed on a laboratory scale, but cannot easily be extended to hot, pressurized large scale plants. For future measurements at hot conditions in a fluidized bed methanation reactor, we have developed an optical probing system that can be employed under these conditions. However, optical sensors are only able to investigate the local fluidization patterns at a defined position in the bed. The objective of this study is to characterize differences in bubble properties between local optical measurements and an X-ray tomography method that is able to detect bubbles over the entire cross-section. Therefore, an artificial optical signal is created out of existing hydrodynamic X-ray measurement data obtained at a cold flow model of a pilot scale methanation reactor. The determined bubble properties of both methods (i.e. evaluation of the derived artificial optical probe signal and image reconstruction based on the evaluation of original X-ray tomographic data) are compared with regard to the bubble rise velocity and the bubble size (for the X-ray method) or pierced chord length (for the optical evaluation method), respectively. The comparison shows that for the evaluation of the optical probe data, statistical effects have to be considered carefully. The detected mean chord length of the optical method does not immediately correspond to the mean bubble size determined by the X-ray method. Moreover, also differences regarding the bubble rise velocity were detected for some fluidization states. The reason for the discrepancies between both methods could be identified and corrected, amongst others by means of a Monte Carlo simulation in which rising bubbles in a fluidized bed were simulated and characterized by a local virtual optical sensor.