Magnetic Resonance Studies of Dead-Zones in Gas-Solid Fluidised Beds

Magnetic Resonance Imaging (MRI) was used to image non-invasively the entry of gas into a 3D, gas-solid fluidised bed through a drilled plate distributor. MRI has two main benefits for investigating gas entry hydrodynamics: firstly, voidage variations in the bed can be imaged at high spatial and temporal resolutions and secondly, dead-zones can be identified using direct measurements of particle velocities. Permanent jet heights, time-averaged voidage maps and dead-zone maps were acquired. The size of dead-zones was found to decrease with increasing orifice velocity. In addition, permanent jet stalks were observed both in the presence and absence of dead-zones. INTRODUCTION Gas distributors for fluidised beds, both at the laboratory and the industrial scale, often consist of drilled plates. However, the exact behaviour of the gas and solids in the region directly above the distributor (known as the grid zone, where much of the gas-solids contacting occurs) is still poorly understood. A better understanding of the effect of distributor design on mixing in the grid zone is therefore important because of its influence on the overall performance of a fluidised bed. The velocity of the gas through each of the orifices in a drilled plate distributor can be an order of magnitude greater than the superficial velocity of fluidisation, leading to the formation of jets of gas, largely free of solids, above the orifices. Rowe et al. (1) defined a jet as being a permanent region of high voidage and this definition is used in this paper. Jets can cause erosion if they impinge on the wall of the fluidised bed or on internal surfaces, and so it is desirable to understand the factors affecting their formation. In addition, heaps of unfluidised particles, known as deadzones (shown in dark grey in Figure 1a), can form between orifices. Dead-zones can lead to reduced rates of reaction because the rates of mass transfer between gas and solids in those regions will be less than in the fully-fluidised bed. In beds in which exothermic reactions are being undertaken, the poorer heat transfer in the dead-zones can give rise to localised heating. This in turn can cause the particles to agglomerate or sinter, thereby eventually reducing the rate of reaction, and potentially leading to blockages and equipment downtime. The literature suggests that the gas emerging from an orifice into a fluidised bed will appear in one of the following forms: (i) a stable jet, (ii) a stable jet “stalk” from which bubbles detach and pass into the main part of the bed, or (iii) a train of bubbles forming immediately at the distributor. When the superficial velocity of the gas, U, is less than that required for minimum fluidization, Umf, permanent, temporally-invariant jets are formed (2). However, above a critical superficial velocity, ~3 Umf, (1) streams of bubbles are formed immediately at the upper surface of the distributor, without any stable jet. Wen et al. (3) suggested that permanent jets of gas were formed only when dead-zones were present and conversely proposed that the presence of a dead-zone was a criterion for the formation of permanent jets of gas. It was predicted (3) that above a critical value of U the diameters of the bubbles produced would be equal to the pitch of the orifices and at this size would prevent the downflow of particles between orifices. As a result, dead-zones, and consequently permanent jets, would not be formed. Horio et al. (4) however found that bubbles detaching from the tips of jets did not affect the dead-zone adjacent to the jet stalk. Figure 1: Schematic diagrams of the three modes of gas entry reported in the literature. a) A permanent jet void (white) is surrounded by a slow-moving, fluidised annulus. Particles are entrained in the annulus and pass upwards through the jet void at high velocities. Dead-zones (regions of unfluidised particles, shown in dark grey) can be formed at the base of the jet. b) A jet ‘stalk’ from which bubbles detach. Two jet lengths are reported for such jets: lj,max, the maximum jet length seen immediately before a bubble detaches and lj, the height of the permanent jet stalk (also referred to as the minimum jet length). c) A train of bubbles forming immediately at the distributor. The optical opacity of fluidised systems makes observation of the transition between stable jets and bubble formation at the orifice difficult. Earlier studies of phenomena at distributors were limited by the experimental techniques available to observe particle motion in the grid zone. Two-dimensional (2D) bed studies have the advantage of allowing direct visual observation of flowing particles without intrusive measurements. However, Merry (5) and Wen et al. (3) have demonstrated that wall effects can lead to significant differences in the behaviour of jets between 2D and 3D beds. Experiments in 3D beds have largely relied on probe measurements (3, 6, 7) or slumping a bed containing tracer particles to observe time-averaged particle motion (4) and to determine the extent of mixing of particles located initially on the distributor so as to visualize dead-zones. However, recent advances in tomographic imaging, such as Magnetic Resonance Imaging (MRI) (8, 9), X-ray computed tomography (XCT) (10, 11) and electrical capacitance tomography (ECT) (12), now allow the non-invasive imaging of 3D beds. Given the unstable nature of the flow in bubbling beds, it is desirable to image voidage distributions within the bed at high temporal and spatial resolutions. In addition, a quantitative, non-invasive measurement technique is needed to identify dead-zones within the bed. MRI, in particular, has the unique capacity to measure directly and non-invasively both voidage and particle velocities. The present paper is concerned with demonstrating the feasibility of MRI for investigating the relationship between dead-zones and stable jets. Particulate phase Jet Annulus Jet Void Deadzone Distributor a) b) c)