Particle Ejection Mechanisms Due to Bubble Eruptions in Fluidized Beds

Joseph C. Dille

A Thesis
Presented to the Graduate Committee
of Lehigh University
in Candidacy for the Degree of
Master of Science
Mechanical Engineering

Lehigh University


The mechanisms by which bubbles erupt at the free surface of a gas fluidized bed were studied experimentally. Two bed geometries were used. One was a two-dimensional bed with a rectangular cross section measuring 14.4 mm x 609.6 mm, and the other was a semicylindrical three-dimensional bed with an internal diameter at 288 mm. Five different types of particles are used with surface meandiameters ranging from 112 um to 912 um and densities ranging from 1,055 to 3,970 kg/m In some experiments bubbles were created by, injecting air into an incipiently fluidized bed; and in others, the bed was allowed to bubble freely. Fluidization velocities ranged from 1.05 to 2.44 Umf.

The bubble eruptions were recorded using a high speed video system. Four mechanisms of particle ejection were identified; and the results suggest that the relative importance of each mechanism depends on bed geometry and the nature of the bubble interactions. From studies to determine frequency of each mechanism, the most common type of eruption was found to be a single bubble breaking at the free surface, ejecting particles from the bulge or nose region of the bubble. No material from the bubble wake was ejected during the single bubble eruptions. This type of eruption was studied in detail and a universal non-dimensional curve for single bubbles is presented. The average maximum height of ejected material in single bubble eruptions was found to be 60 percent of the free surface diameter in the three-dimensional bed and 57 percent of the free surface diameter in the two-dimensional bed.


Particle elutriation is a very important consideration in the design and use of fluidized bed systems. For example, in the case of pressurized fluidized bed coal combustion for combined cycle power generation, the rate at which particles are entrained in the exhaust gas determines the type and capacity of dust controlling equipment. If the exhaust gas is to be used in a gas turbine, adequate dust control is especially critical.

Presently, information on elutriation is primarily limited to empirical correlations of experimental data, (for example, see Refs. 1-6.) These correlations are not useful for the design of fluidized bed systems if the operating conditions are much different from those used to generate the correlations. Presently, in industry, if exact information is needed on elutriation rates for a new process, scale model tests must be run.

Relatively little has been done to gain insight into the physical mechanisms which govern elutriation. The identification and understanding of these mechanisms are essential in developing a useful mechanistic model of the elutriation process. One thing which has been established is that the elutriation phenomenon can be divided into two distinct processes. The first is the eruption of bubbles from the free surface of the fluidized bed causing particles to be ejected into the freeboard. The second is the transport of the ejected particles by the fluidizing gas.

This investigation deals with the first process, bubble eruption and the ejection of particles above the free surface. It was stimulated by a controversy in the technical literature over the origin of the particles ejected by the bubble. Figure 1.1 shows typical bubbles rising in a two-dimensional bed. Some investigators have observed the bubble wake being ejected into the freeboard. One of these investigations was performed by George and Grace (7) who captured ejected particles in the freeboard region of a fluidized bed made from two types of bed materials. Another investigation which observed wake particles being ejected was conducted by Rowe and Partridge (8, 9) who obtained photographs of a bubble bursting on the surface of a cylindrical bed. On the other hand Do at. al. (10), using a two-dimensional fluidized bed and a high speed camera, observed the bulge material ahead of the bubble erupting into the freeboard.

The object of this work is to determine the mechanisms by which particles are ejected above the free surface of a bubbling bed. As is described in later chapters, the major ejection mechanisms change with particle parameters and fluidization conditions. In this study the effects of bubble size, particle size, fluidization velocity, particle density and bubble coalescense near the free surface are examined. Observations were also made to determine if any fundamental differences exist between the eruption mechanisms in two-and three-dimensional fluidized beds. The knowledge obtained in this investigation is intended to serve as a basis for the future development of particle ejection models.

Sections 2-6 are not included in this summary. If you are really interested in the entire content of my thesis, or have a severe sleep disorder, you can e-mail me and I will arrange a copy. You could also contact Lehigh University.

Here are some of the more interesting figures:


The conclusions drawn from this study are listed as follows:

  1. Four distinct particle ejection mechanisms which throw significant amounts of material into the freeboard exist in a bubbling fluidized bed. The mechanisms are bulge burst, middle-layer burst, jet spray and wake spike.
  2. The four mechanisms occur in all materials and for all bed depths. The bulge burst occurs in single bubbles in both two- and three-dimensional beds. The midlayer burst and jet spray occur in both two- and three-dimensional beds, but only when two or more bubbles coalesce near the free surface. The wake spike mechanism occurs only in three-dimensional geometries when two or more bubbles coalesce just below the free surface.
  3. In the bulge burst mechanism ejected particles originate only from the bulge. In the middle-layer burst mechanism ejected particles originate from both the bulge and middle layer. In the jet spray mechanism ejected particles originate from the bulge, middle layer and along the sides of the lower cavity. In the wake spike mechanism ejected particles originate from the bulge, middle layer and lower bubble wake with the wake particles rising the highest.
  4. Of all bubbles observed, about 10 percent are double bubbles.
  5. The trajectory of the three points during the single bubble, bulge burst eruption can be non-dimensionalized using the parameters H/D FS and -r.
  6. The average height reached by the bulge material during a single bubble eruption in the three-dimensional bed was 60 percent of DFS while the wake material only rose to a height of 0.33 DFS' In the two-dimensional bed, the average maximum bulge height was 57 percent of DFS' In every case with two-dimensional bubbles, the wake never went above the undisturbed bed free surface.

The following recommendations are offered :

  1. The two-dimensional fluidized bed is a useful tool for bubble visualization but since substantial differences were found between the two- and three-dimensional beds, and the three-dimensional bed more closely approximates industrial geometries, further studies should emphasize three-dimensional systems.
  2. In order to study larger bubbles a semi-cylindrical bed with a larger diameter should be constructed. This bed should be operated at larger bed depths to allow the formation of larger bubbles.
  3. The multiple bubble type of eruption ejects more particles to greater heights than a single bubble. For this reason, this mechanism should be studied further, and the double bubble event should be characterized.
  4. The above results show that the wake spike eruption is an important mechanism in particle ejection and that this mechanism does not appear with the two-dimensional geometry. In light of these facts, a system of baffles or internals should be studied which would have the effect of compartmentalizing or two-dimensionalizing the upper portion of the bed, thus preventing wake spikes but not greatly affecting the mixing characteristics.

The Dille's Home PageDille's Home Page Joe's PageJoe's Page Joe's Publications PagePublications