Mixing and Segregation of Biomass Particles In a Bubbling Bed

Introduction

Fluidized beds are being widely considered for converting solid biomass into liquid fuels, syngas, and chemical products.     While fluidized beds have been extensively studied for many applications, the detailed hydrodynamics and mixing patterns in systems with disparate types of bed particles is still not fully understood.  A number of biomass gasification and pyrolysis processes now being considered employ bubbling beds of fine sand into which larger biomass particles are fed.  At steady state, the biomass particles usually constitute only a small fraction of the bed so that the dynamics involve a few coarser, lighter particles circulating in a finer, denser medium.  This disparity in particle sizes and densities creates the potential for segregation, especially at low fluidizing velocities.  The extent of segregation can be extremely important, because it determines the intimacy and duration of contacting between released gases and bed material.  In turn, the degree of gas-particle contacting can greatly affect product yield and composition, especially when the bed material has catalytic properties.  We report results from visual observation of bubbling laboratory beds combined with magnetic particle tracking to measure detailed mixing and segregation patterns for simulated biomass particles.  Our ultimate goal is to develop correlations and statistical models which can be used to predict biomass mixing and segregation patterns in bubbling bed reactors.

Experimental Materials and Methods

We use our previously reported (1) magnetic particle tracking method in a 55 mm three-dimensional bed fluidized with air at ambient conditions.  Safe, inexpensive magnetic tracers and detectors are used, thus avoiding the issues associated with past particle-tracking methods.  In this study, we used neodymium magnets embedded in wooden spherical and cylindrical particles and externally positioned magnetic field detectors to continuously locate the position of a tracer particle.  The method takes advantage of the tendency of the tracer particles to align their magnetic axes with the earth's magnetic field like a compass needle.  Proper positioning of the probes can take advantage of this tendency, which simplifies the data analysis.  We also collected high-definition slow-motion videos of the bed surface to observe top-segregated tracer motion.

Both beds of approximately 200 micron glass beads and 100 to 200 micron sand were studied.  Single biomass tracer particles were dropped into the beds and their positions determined over a 5-minute run time.  Tracers had diameters of 3 to 5 mm and densities from 0.55 to 1.2 g/cc.  Tracer positions were determined using the algorithm we previously reported on.  Statistical analysis on the time series position data was used to characterize the tracer behavior.  Fluidization velocity varied from 1.1 to 6 times minimum fluidization.  Slumped bed depth to diameter ratio was held constant at 1.

 

Results and Discussion

Frequency distributions of the vertical position for various conditions were determined.  In some cases these show segregation at the top or bottom and under other conditions a uniform distribution across the bed.  We found that these distributions could be reasonably well fitted with the two-parameter Weibull distribution function and that the two parameters correlated with tracer density and superficial velocity.  Vertical velocity distributions were also determined.  In addition, typical three-dimensional trajectories showed typical bubble-inducted solids circulation at low velocities, top and bottom segregation behaviors, and mixing at higher velocities.  It was found that a tracer with a density of 0.89 g/cc continued to circulate (appeared to be neutrally buoyant) at even very low fluidization velocities (1.1 times minimum fluidization velocity) while heavier tracers segregated to the bottom and lighter ones to the top at low velocities.  At high velocities, all tracers in the 0.55 to 1.2 g/cc density range mixed and circulated.  Higher-density tracers, however, did not always circulate, even at the highest velocities studied.  The implications for biomass processing is that appropriate processing conditions will likely depend on the type of biomass, degree of dryness, and whether it is processed or raw.

1. Patterson et al., Ind. Eng. Chem. Res. 2010, 49, 5037-5043.

Abstract submitted for presentation at the 2011 Annual A.I.Ch.E. meeting in Minneapolis, MN