Oscillations in gas-fluidized beds: Ultra-fast magnetic resonance imaging and pressure sensor measurements C.R. Müller ⁎ , J.F. Davidson, J.S. Dennis, P.S. Fennell, L.F. Gladden, A.N. Hayhurst, M.D. Mantle, A.C. Rees, A.J. Sederman Department of Chemical Engineering, University of Cambridge, Pembroke Street, Cambridge, CB2 3RA, United Kingdom Received 7 June 2005; received in revised form 20 November 2006; accepted 20 February 2007 Available online 27 February 2007 Abstract Ultra-fast Magnetic Resonance Imaging (MRI) and pressure sensor measurements have been applied to study: (i) pressure fluctuations, (ii) the eruption of bubbles at the top of a bed and (iii) the formation of bubbles in a gas-fluidized bed. Ultra-fast MRI has been applied for the first time to study the formation and eruption of bubbles; the technique is non-intrusive and provides measurements with good temporal and spatial resolutions. The MRI measurements revealed that bubbles are formed periodically, rather than randomly at a distributor, which in this case was a perforated plate. The frequency at which bubbles erupted from the top of the bed matched the frequency of the pressure fluctuations measured just above the distributor, where the measured pressure is predicted very well for the case of slug flow by Kehoe and Davidson's [P.W.K. Kehoe, J.F. Davidson, Pressure fluctuations in slugging fluidized beds, AIChE Symp. Ser. 128 (69) (1973) 34–40] correlation, originally developed for locations high up a bed. Both findings lead to the conclusion that the passage and eruption of bubbles at the top of a bed are the dominant cause of the pressure fluctuations, which are subsequently propagated downwards through the bed. Two new correlations are proposed for predicting the frequency of pressure fluctuations in a bubbling bed; both correlations agree well with experimental measurements. A modification of Baeyens and Geldart's [J. Baeyens, D. Geldart, An investigation into slugging fluidized beds, Chem. Eng. Sci. 29 (1974) 255–265] correlation predicts the frequency of pressure fluctuations when slugs are formed, but are not fully developed. The frequency of bubble formation, as measured by MRI, is equal to or higher than both the frequency of bubble eruption at the top of the bed and the frequency of pressure fluctuations, depending on the depth of the bed. The frequency of bubble formation is significantly lower than predicted by Davidson and Schüler's [J.F. Davidson, B.O.G. Schüler, Bubble formation at an orifice in an inviscid liquid, Trans. Inst. Chem. Eng. 38 (1960) 335–342] equation, originally developed for gas–liquid systems. © 2007 Elsevier B.V. All rights reserved. Keywords: Bubble eruption; Bubble formation; Pressure fluctuations; Magnetic Resonance Imaging; Gas-fluidized beds 1. Introduction Fluctuations in the pressure measured at some height up a fluidized bed have been attributed to: (i) the formation of bub- bles at the distributor, (ii) the coalescence of bubbles and (iii) the passage and eruption of bubbles at the top of the bed [1–3]. Some existing correlations for predicting the frequency of pressure fluctuations are summarised in Table 1. The theories of both Hiby [3] and Verloop and Heertjes [4] assumed homo- genous fluidization, i.e. there are no bubbles in the bed, so these theories only apply to very shallow beds. Davidson [5] treated the bed as a piston acting on the gas in the plenum chamber and consequently his theory only applies to distributors with a low resistance to flow. Verloop and Heertjes [4], Kehoe and David- son [6] and Baeyens and Geldart [7] all used very similar models for the frequency of slugs in fairly deep beds. Their principal assumption was that the pressure fluctuates with the same frequency as the passage of the slugs. Baskakov et al. [8] proposed a theory for the frequency of pressure fluctuations assuming that bubbles erupt one at a time at the surface. Sada- sivan et al.'s [9] correlation was derived by a linear regression of experimental results. A quite different approach was taken by Roy et al. [10], who treated the fluidized bed as an organ pipe: they assumed the wavelength to be λ =4H, where H is the height of the expanded bed. Thus, the frequency of pressure fluctuations is f = u s / λ, with u s being the velocity of sound in the fluidized bed. Powder Technology 177 (2007) 87 – 98 www.elsevier.com/locate/powtec ⁎ Corresponding author. Tel.: +44 1223334787; fax: +44 1223762962. E-mail address: cm433@cam.ac.uk (C.R. Müller). 0032-5910/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.powtec.2007.02.010