Combined hydrodynamic and reaction analysis of a bubbling to turbulent Fluidized Bed Reactor

Abstract

There are many large-scale contacting methods for gas reactions requiring a solid catalyst. The catalytic gas-solid Fluidized Bed Reactor (FBR) is one of the popular methods in industry. In FBRs the bulk of the gas throughput is present as lean bubbles, mostly deprived of solids, bubbling through a solids-rich emulsion phase. The movement of gas into and out of the emulsion often dictates the performance of an FBR. During the past five decades major contributions have been made towards the understanding of FBRs, although numerous gaps still exist, especially at higher bubbling regime velocities. This work follows an integrated approach for the simultaneous measurement of hydrodynamics and reactor performance. Hydrodynamics are measured using fast X-Ray Tomography (XRT), pressure analysis techniques and an optical fibre probe. Reactor performance is measured by utilizing the ozone decomposition reaction. Performance is quantified using a basic two-phase reactor model with an apparent overall interphase mass transfer (K0) parameter. Two 14 cm (ID) fluidized bed columns are used, one setup supporting the ozone decomposition reaction and the other installed within a fast XRT facility. Special emphasis is placed on superficial velocities (U0) spanning the entire bubbling regime up to the onset of the turbulent regime (Uc). The particle types employed are Geldart B sand particles and highly dense ferro-silicon (FeSi) particles. Fines are added to both particle types, resulting in a total of four particle systems (sand baseline; sand with fines; FeSi baseline; FeSi with fines). Time constraints on the XRT equipment limited the tomography measurements to the sand baseline particle system. The hydrodynamics of the other particle systems were limited to the pressure signal and optical probe measurements of the ozone decomposition setup. The results of the sand baseline system suggest that a distinction should be made between the low-interaction bubbling regime and the high-interaction bubbling regime. A change in mass transfer behaviour occurs around a U0/Uc value of 0.25. Reactor performance increases up to U0/Uc = 0.7, after which a decreasing trend is observed. An empirical correlation is proposed for the specific interphase mass transfer (kbe) of the higher velocity bubbling regime. This correlation is based on the integration of the hydrodynamics determined by means of XRT and reactor performance: 4-12 The hydrodynamic parameter β gives the best fit for the entire velocity range with an average error of 8%, although it is not recommended for U0/Uc<0.17. It is observed that the classical approach of penetration theory for interphase mass transfer, performs exceptionally well at low velocities (U0/Uc<0.34). The addition of fines to the FeSi particle type decreases the overall reactor performance, despite decreased bubble sizes. The solids fraction, however, unexpectedly increases with the addition of fines and a collapse of the emulsion phase is measured. It is therefore postulated that though flow in the emulsion phase is much higher for the FeSi baseline system and decreases with the addition of fines. For the sand particle type, the behaviour expected from literature is observed: reactor performance increases, bubble sizes decrease and the solids fraction decreases. Very distinct hydrodynamic behaviour is observed for all the fluidization regimes with XRT. Probability density distributions show there are still two phases present in the turbulent regime and that the emulsion-phase solids fraction remains independent of velocity until fast fluidization sets in. The turbulent regime has unique hydrodynamic behaviour, although voids appear to be a transient structure between the structures of the bubbling and fast fluidization regimes.