86
3 A Structuring Regime to Control Bubbling Beds
confinement of bubbles induced solely by pulsation also contributes to the calculated
intensity. Under a slowly pulsed flow, bubbles form at each pulsation and concentrate
as a horizontal array without clear separation. This type of flow, nevertheless, cannot
be recognised as a structured flow according to human perception. Moreover, the
reference flow is constructed by redistributing extracted bubbles of each frame in a
randomisation manner, whereas, in fact, experimental bubbles in chaotic flows are
correlated.
Apart from manipulation of bubbles, the emerged patterned flows manifest the
advantages to tightly control over the distributions of size, residence time, and spatial
separation of bubbles, to a certain extent. From an engineering perspective, these
flexibilities allow one operating in structured flows to decouple conflicting design
objectives, such as spatial mixing and bubble size. In addition, the unique appearance of bubbles array itself allows bypassing the flow maldistribution, non-uniform
gas-solid contact as well as geometry dependence, which normally challenge operation and design of conventional fluidised beds. Nevertheless, overall mixing in
structured flows could be less pronounced than in conventional fluidisation, but the
arrays of bubble transform macroscopic mixing into a highly controlled, homogenous
and local micromixing, constituting a series of laterally compartmented units. Each
micromixing unit is therefore governed by two characteristic length scales: bubble
size and the pattern wavelength. Given relatively tiny bubbles and low flow rates for
inducing structured flow, one would expect to mitigate the barriers of interphase heat
and mass transport, with transport rates that lie in between those in a fixed and a
vigorous bubbling bed.
In terms of implementation, pattern formation induced directly via pulsating the
inlet gas flow, and then it is easily applicable to existing setups in a non-invasive
way. Nevertheless, expansion in the axial direction is limited as per experimental
observation. The bottleneck is not the formation of alternatively nucleated bubbles
from the distribution plate, but the propagation process as bubbles rise. Due to the
necessity of a large λ/D b to prohibit disturbance and sustain structured flows, either
solely increasing separation or reducing size is supposed to improve the propagation.
For example, one could try to extend the bed thickness but still maintain its quasi-2D
nature, in order to alter the growth rate of bubbles. Besides, internals in the third
direction could be also placed at different levels to constrain the passing bubbles.
3.4 Conclusions
This chapter demonstrates the recognition and utilisation of a macroscopic dynamically structured flow of bubbles in fluidisation created under modulated oscillations.
A model-based pattern recognition algorithm has been developed to quantify the
degree of order induced via pulsation by examining the cross-correlation of bubble
motions. As per the computed pattern intensity, two different classes of regimes,
structured and unstructured flows, in the full operating regime for pulsed fluidisation
in quasi-2D beds can be distinguished. In the unstructured regime, bubble dynamics
3 A Structuring Regime to Control Bubbling Beds
confinement of bubbles induced solely by pulsation also contributes to the calculated
intensity. Under a slowly pulsed flow, bubbles form at each pulsation and concentrate
as a horizontal array without clear separation. This type of flow, nevertheless, cannot
be recognised as a structured flow according to human perception. Moreover, the
reference flow is constructed by redistributing extracted bubbles of each frame in a
randomisation manner, whereas, in fact, experimental bubbles in chaotic flows are
correlated.
Apart from manipulation of bubbles, the emerged patterned flows manifest the
advantages to tightly control over the distributions of size, residence time, and spatial
separation of bubbles, to a certain extent. From an engineering perspective, these
flexibilities allow one operating in structured flows to decouple conflicting design
objectives, such as spatial mixing and bubble size. In addition, the unique appearance of bubbles array itself allows bypassing the flow maldistribution, non-uniform
gas-solid contact as well as geometry dependence, which normally challenge operation and design of conventional fluidised beds. Nevertheless, overall mixing in
structured flows could be less pronounced than in conventional fluidisation, but the
arrays of bubble transform macroscopic mixing into a highly controlled, homogenous
and local micromixing, constituting a series of laterally compartmented units. Each
micromixing unit is therefore governed by two characteristic length scales: bubble
size and the pattern wavelength. Given relatively tiny bubbles and low flow rates for
inducing structured flow, one would expect to mitigate the barriers of interphase heat
and mass transport, with transport rates that lie in between those in a fixed and a
vigorous bubbling bed.
In terms of implementation, pattern formation induced directly via pulsating the
inlet gas flow, and then it is easily applicable to existing setups in a non-invasive
way. Nevertheless, expansion in the axial direction is limited as per experimental
observation. The bottleneck is not the formation of alternatively nucleated bubbles
from the distribution plate, but the propagation process as bubbles rise. Due to the
necessity of a large λ/D b to prohibit disturbance and sustain structured flows, either
solely increasing separation or reducing size is supposed to improve the propagation.
For example, one could try to extend the bed thickness but still maintain its quasi-2D
nature, in order to alter the growth rate of bubbles. Besides, internals in the third
direction could be also placed at different levels to constrain the passing bubbles.
3.4 Conclusions
This chapter demonstrates the recognition and utilisation of a macroscopic dynamically structured flow of bubbles in fluidisation created under modulated oscillations.
A model-based pattern recognition algorithm has been developed to quantify the
degree of order induced via pulsation by examining the cross-correlation of bubble
motions. As per the computed pattern intensity, two different classes of regimes,
structured and unstructured flows, in the full operating regime for pulsed fluidisation
in quasi-2D beds can be distinguished. In the unstructured regime, bubble dynamics