Ž . Powder Technology 123 2002 147–165 www.elsevier.comrlocaterpowtec Computational investigation of slugging behaviour in gas-fluidised beds S.J. Zhang, A.B. Yu ) Centre for Computer Simulation and Modelling of Particulate Systems, School of Materials Science and Engineering, The UniÕersity of New South Wales, Sydney, NSW 2052, Australia Received 24 January 2001; received in revised form 30 July 2001; accepted 16 August 2001 Abstract A computational study has been carried out of the slugging behaviour in fluidised beds using a two-fluid continuum model. According to this approach, the two phases are treated as separate interpenetrating continuums, respectively described by the governing equations and coupled through an interfacial momentum exchange term. The computations are started from minimum fluidisation or bubbling flow conditions for two-dimensional and symmetrical three-dimensional fluidised beds. The well-known SIMPLE algorithm is employed to numerically solve the gas–solid two-phase governing equations. The validity of the approach is confirmed by comparing the model predictions with experimental measurements in the literature. The computational results are presented mainly in terms of porosity contours to illustrate the formation, growth, rising and bursting of slugs. With prescriptions of various wall conditions, round-nosed, square-nosed and wall slugs are predicted in the framework of the numerical simulations, where significant factors such as superficial gas velocity, wall or boundary condition and bed geometry are identified through a parametric study. It is also demonstrated that the slug rising velocity, length and frequency, and the bed expansion can be predicted by the proposed approach, and the particle–wall friction is a key factor responsible for the deficiency of the empirical correlations in the literature. q 2002 Elsevier Science B.V. All rights reserved. Keywords: Fluidisation; Slug flow; Two-fluid continuum model; Kinetic theory 1. Introduction A gas–particle bed cannot expand indefinitely into an ideal, uniform fluidised bed by passing an upward flow of gas through it. Instead, it forms a complex, heterogeneous structure traversed by voids or bubbles. A number of flow regimes have been identified to describe the complicated behaviour including the so-called smooth, bubbling, slug- w x ging, turbulent and lean phase fluidisation 1,2 . A reactor will tend to operate in the slugging regime if it is small and the gas velocity is sufficiently high so that the bubbles coalesce to become as large as the reactor. Small labora- tory reactors are particularly prone to operate in this regime. Slug flow is less likely with a large bed, but if a bed contains a bundle of tubes, slug flow can readily occur in the spaces among the tubes. Slugging is usually undesir- able since it increases the problems of entrainment and lowers the performance potential of a reactor. However, its study has been found to be useful in understanding the interactions among gas, particle and wall. ) Corresponding author. Fax: q 61-2-9385-5956. Ž . E-mail address: a.yu@unsw.edu.au A.B. Yu . wx Stewart and Davidson 3 describe two main solid flow patterns in the slug regime: round-nosed and square-nosed slugs. In a smoothly slugging bed, slugs are round-nosed and solids flow past the slug in an annular region on the wall. The square-nosed slugs fill completely the cross-sec- tion of a bed, with solids raining through the slug. In addition, asymmetric or wall slugs have also been de- wx scribed by Stewart and Davidson 3 and Kehoe and wx Davidson 4 . Obviously, the mechanism of instability that is responsible for these slugging phenomena is of great importance to understanding the mechanics of fluid–par- ticle suspensions. In the past, quite a few efforts, mainly experimental, have been made in this direction, leading to various empirical formulations for determining macro- scopic slug behaviour, such as onset slugging velocity, w x slug frequency, slug length and slug rising velocity 3–13 . However, better quantitative understanding of the hydrody- namics of the unstable fluidisation and detailed knowledge of gas–particle flow behaviour are highly needed for de- sign and scale-up of efficient fluidised reactors in the petroleum, chemical and electric power industries. As such, the prediction of the gas–particle flow field in fluidised bed systems based on the hydrodynamic principles is not only of theoretical, but also of practical, significance. 0032-5910r02r$ - see front matter q 2002 Elsevier Science B.V. All rights reserved. Ž . PII: S0032-5910 01 00445-4