ELSEVIER Powder Technology 82 (1995) 153-167 POWDER TECHNOLOGY Lateral solid mixing measurements in circulating fluidized beds Detlef Westphalen, Leon Glicksman Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA Received 17 January 1994; revised 14 September 1994; accepted 4 October 1994 Abstract A thermal technique has been developed for measurement of solid mixing in fluidized beds. A batch of bed particles is preheated and then injected into the bed. An array of thermistor probes measures the spread of the tracer. A series of experiments has been done in the upper core region of a 7 m tall, 0.20 m diameter test circulating fluidized bed (CFB). The experiments involved one type of particle, having 180 /~m diameter and 2350 kg/m 3 density. Models for the spread of tracer particles, transfer of heat from tracer particles to surrounding gas and particles, and for heat transfer to the thermistors were used to make estimates of the radial particle ditfusivity in the core region. The diffusivity was on the order of 10 crn2/s for the range of velocities and solid concentrations tested. A second set of preliminary solid mixing experiments has been done in the bottom of a 1.8 m tall, 0.16 m square CFB using 55 /xm FCC (fluid cracking catalyst) particles. The mixing in this region, although not representative of a strictly diffusional process, appears to be somewhat more vigorous than the core mixing in the upper region of the bed. Keywords: Fluidized beds; Mixing; Circulating beds 1. Introduction Effectiveness of a circulating fluidized bed (CFB) combustor or reactor depends on the ability to mix adequately the incoming flows of reactants: fuel, sorbent and air. Emissions and combustion efficiency are strongly related to solid mixing. One example of this is the Nucla 110 NWe CFB combustor, in which the required limestone/sulfur feed ratio doubled when the number of limestone feed ports was reduced [1]. In the present investigation a thermal technique has been used to study local solids mixing in the upper core of a CFB. The technique requires models of the tracer spread, energy conservation and thermistor re- sponse to derive the particle diffusivity from the ex- perimental measurements. These models are presented along with a consideration of uncertainties introduced by the model assumptions. Results for the radial dif- fusivities are presented for a range of operating con- ditions. Preliminary mixing results for the lower dense region of the bed were obtained to contrast the con- ditions in the lower and upper portions of the bed. 2. Review of experimental work of other researchers The number of CFB solid mixing experiments reported in the literature is somewhat limited. The experiments usually involve injection of tracer particles into the bed and detection of the tracer particles at a distance from the injection point. Bader et al. [2] and Rhodes et al. [3] used salt particles as tracer. Samples of solids taken from the beds were analyzed by dissolving them in distilled water; the conductivity of the resulting solution is roughly proportional to the amount of salt tracers in the sample. Chesonis et al. [4] prepared tracer particles by impregnating the alumina bed solid with CaCI2; atomic absorption was used to determine tracer solid concentrations in samples of solid taken from the bed. A radioactive tracer technique was used by Ambler et al. [5]. A magnetic tracer technique was used by Avidan [6]. A technique using optical fibers and FCC treated with fluorescent dye was used by Kojima et al. [7]. These last three experiments involved sensors mak- ing real-time measurements rather than post-experiment analysis of bed samples. A problem inherent with all of these techniques is the difficulty of removing the tracer particles from the CFB system or of deactivating the trace. Occasional replacement of the entire bed solid inventory is required in order to minimize the effect of background tracer noise levels, For this reason, a relatively small number of tests was done in some of the studies. Experimental results from this study indicate a wide disparity in 0032-5910/95/$09.50 © 1995 Elsevier Science S.A. All rights reserved SSDI 0032-5910(94)02909-8