|dc.description.abstract||Fluidized beds have been used extensively in chemical process industries for several years. The
success of fluidized operations is largely dependent on a well-defined and stable contact regime.
Hence, the ability to understand the flow regime behaviour plays a vital role in the design and
operation of fluidized bed units in order to achieve a particular stable fluid dynamic state.
Fluidization regime is dependent on factors such as particle properties (size, density and
geometry), column properties (size and geometry) as well as the fluidizing medium properties
(density, viscosity and velocity) (Fan, et al., 1981).
In order to expand the understanding of the hydrodynamics of a fluidized bed system, a
substantial amount of research has been committed to the measurement and analysis of pressure
fluctuations in a fluidized bed. This is due to the strong relationship between pressure
fluctuation and the hydrodynamic factors which include bubble size, bubble rising velocity and
the motion of the bed surface with time (Hartman, et al., 2009). Identification of each regime
could be accomplished through the time-series analysis of the pressure fluctuation in the time
domain, frequency domain and state space domain (van Ommen, et al., 2011).
The main focus of this dissertation was to apply time-series analysis in the frequency domain,
for the characterisation of the different fluidization regimes (particulate, bubbling, slugging and
turbulent) in a gas-solid fluidized bed. This was achieved through the use of spectral analysis
and the mathematical tool known as the Fast Fourier Transform (FFT) was used to analyse and
interpret the pressure fluctuation in the fluidized bed. Analysis was also performed in the time
domain by analysing the time-pressure behaviour as well as the change in the standard deviation
of pressure fluctuations. Experimental measurements were conducted in three different columns
with varying column height and diameter. Three different solid particles were used namely,
sand particles (Geldart Group B), plastic beads (Geldart Group D) and spent Fluid Cracking
Catalyst (Geldart Group A). The sampling frequency used for pressure measurements in this
work was fixed at 500 Hz with a sampling time of 30 minutes.
Results indicated that the pressure fluctuation signal is useful in providing information about
fluidized bed behaviour. Analysis in the time domain revealed that this technique could be used
primarily to identify whether fluidization has occurred or not. Analysis in the frequency domain
indicated a better representation of the fluidization behaviour and the different regimes could
clearly be identified and distinguished based on a dominant frequency. In the 5 cm diameter
column, the Geldart Group B particles displayed a distinct dominant frequency for the bubbling
regime while a dominant frequency could not be obtained for the Group A and D particles
respectively. In the 11 cm diameter column, the Geldart Group B materials were observed to
fluidize very easily, with three dominant frequencies corresponding to the bubble, slugging and
turbulent regimes, being identified. The Geldart Group D particles were found to fluidize at
high velocities in the 11 cm diameter column with the bubbling and slugging regimes being
identified. Geldart Group A particles were found to behave very differently from the other two
materials. A noticeable bed expansion was seen before fluidization actually occurred. The only
regime achieved with the Group A particles was the bubbling regime. Results for the 29 cm
diameter column indicated a dominant frequency for the bubbling regime for the Group B
particles while measurements could not be performed using other materials due to limitations on
the pressure transmitter. It was further observed that the dominant frequencies were much more
pronounced at higher bed heights. In addition, a change in the aspect ratio (bed height: column
diameter) had a significant influence on the dominant frequency as a visible shift was apparent.
An increase in the aspect ratio indicated a noticeable decrease in the dominant frequency
component. This was valid for all fluidization regimes investigated.||en_US