Thomas J. Fitzgerald

Plume model for large particle fluidized-bed combustors

Abstract A model is proposed to represent the underfeed, fluidized coal combustor. Its assumptions are described and its mathematical representation developed. This model accounts for the possible rapid release of volatiles near the feed ports, particle reaction and shrinkage in the bed, elutriation of unburned fines and afterburning of volatiles above the bed. Design charts are then presented to predict the carbon efficiency of the combustor as well as the temperature jump above the bed in terms of the type of coal used, the number of feed ports in the bed, percentage excess air, gas velocity and the amount of secondary air needed to introduce the coal. As a special case, this model also represents the fluidized combustor with an overfeed of large coal particles or any combination of overfeed of large particles and underfeed of fines.

The uses of magnetic fields in the processing of solids

Abstract This paper shows how properly applied magnetic fields can advantageously be used to control the behavior of streams of solids which contain some magnetizable material. In particular, we describe the development of a new magnetic flow meter for solids, a quick acting on-off valve with no mechanical action, a sliding solids reactor which may be used for particle filtration and related operations, a positively controlled multistage moving bed contactor, fluidized beds with no gross bubbling or gas bypassing, and especially, a reliably operating distributor-downcomer for fluidized beds. The prospects are considered that these devices may open the way to new types of gas/solids contactors including non-plugging trouble-free multistage fluidized beds.

37 A comparison of the plume model with currently used models for atmospheric fluidized bed combustion

Abstract Models in the literature for fluidized bed coal combustors are tabulated and examined with regard to their critical assumptions. These models are shown to not represent, even qualitatively, the expected behavior of the large scale atmospheric fluidized bed combustor (AFBC) with its large-particle tube-filled beds operated at high gas velocity. This paper focuses on the plume model which has been specifically developed to represent the AFBC. The assumptions, main features and simplification of this model are presented, as well as its predictions for carbon efficiency and temperature jump in the freeboard.

A Model for Heat Transfer to Horizontal Tubes Immersed in a Fluidized Bed of Large Particles

Clues from experiment are used to formulate a model for heat transfer between horizontal tubes and bed. The model assumes that when the emulsion is present at the tube surface, heat is transferred by particles and by gas percolating between particles and the surface; when bubbles are present heat is transferred across the gas film. This model is developed for large particle beds, (dp > 1 mm), and agrees with reported results over a wide range of conditions.

Solid circulation system with shrinking core kinetics in both reactor and regenerator

Abstract This paper treats the behavior of solids which circulate between two environments, both in mixed flow. In the reactor the solid react and remove a component from the gas stream, in the regenator another reaction takes place releasing the captured material and restoring the solid to its initial state. Shrinking core kinetics is assumed in which case the particles will have a varying number of concentric rings of product. We related the mean composition of the solid stream with the system parameters (flow rate of solid and size of units) and the particle kinetics, and present the results in the form design charts.

The Mixing of Tracer Gas in Fluidized Beds of Large Particles

Gas mixing is studied in fluidized beds of large sand and dolomite particles with mean diameters of 1.3 mm to 4.0 mm. A tube array of.050 m cylinders in 0.10 m equilateral triangular pitch was used to study the effects of tubes.

Theory of blending in single inlet flow systems

Abstract Good design requires that the fluctuations in concentration which occur at the output of a blending device be small. Using the criterion of minimum variance of the output concentration, it is shown that blender performance depends only on the residence time distribution of the blender and the autocovariance function of the input concentration. The optimal blender design is found for the case of rapid uncorrelated fluctuations at the inlet, and nearly ohtimal blender designs are found for input fluctuation which are autocorrelated over a moderate time interval. A general method for blender design is given using discrete time increments.