Chemical

Engineering in
Practise 3
University of Edinburgh

School of Engineering and Electronics

Fluidised Beds
Yasmine Sweidan, Brian Armstrong, Lee
Norman, Nigel Simpson, Songwen Huang

Yasmine Sweidan
0829278

Monday 1st February 2010

This report focuses on the behaviour of two types of fluidised beds and how their performance
varies under different operating conditions. Bubbling in a fat bed and slugging in a narrower bed are
the basis of the experiment.

The relationship between the superficial velocity and the corresponding pressure drop and bed
height are explored, the main conclusion being that the heavier a particle is, the higher the minimum
superficial velocity must be in order for fluidisation to take place. Various references to theory are
made, such as the predicted minimum fluidisation heights and velocities, and successfully proved
through this experiment.

2
 PAGE №
Contents
....................... 2

1. Abstract ....................... 4

2. Introduction and Background ....................... 5

3. Theory and Experimental Procedure ....................... 6

4. Apparatus ....................... 7

5. Results ....................... 11

6. Discussion ....................... 11

7. Conclusions ..................... 12

8. Appendix:
Raw Data Tables

Graphs

Finding Uv

Narrow Bed Voidage

References

3
Introduction and Background:

Cracking, drying, mixing, waste treatment and catalyst regeneration are all process that play major
roles in industry. The common factor in of all these procedures is their incorporation of and reliance
on fluidised beds. Although available in many different forms and with capabilities to fulfil various
functions, fluidised beds all share the same underlying principal: the suspension of a solid in a
controllable fluid flow, allowing multiphase chemical reactions to take place.

At lower velocities, the bed behaves in a packed form, where the fluid passes through the voids in
the solid granules. Once the fluid has a high enough velocity to lift the solid particles, fluidisation
occurs and the reactor acts as a fluidised bed.

They have many advantages, such as their ability to achieve both uniform particle mixing and
temperature gradients: better fluid-solid contact is achieved with the absence of concentration
gradients, allowing for higher efficiency and quality, and degradation encountered from local hot or
cold spots generated in packed beds is no longer a problem. Another way in which they prove more
efficient than packed beds is the fact they can be operated in a continuous state.

In 1995, fluidised beds produced over $192 billion of the US GDA ₁, showing just how vital their
invention and implementation has been in industry. As a relatively new tool in the chemical
industry, much research especially into the manner in which the actual materials in a fluidised bed
behave is being undertaken₂. Other problem areas are also being tackled such as particle
entrainment, where finer particles need separated from the final product in what can be an
expensive process; pumping power and how it relates to pressure drops; managing the expensive
maintenance of the reactor vessel due to eroding nature of the fine fluid like solid particles. So
fluidised beds can only be expected to become more efficient, more effective, and even more
economical as time goes by. In this experiment we shall be investigating the effects velocity and
pressure have on a fat bed and narrower one.

Different flow regimes can be seen in the reactor upon changing the operating conditions and
properties of the solid phase. Of the four types of regimes found describing bed flow – slugging,
boiling, channelling and spouting – we shall be focussing on the first two.

4
Theory and Experimental Procedure:

We had different aims in the analysis of each fluidised bed.

In the fat bed, our goal was to find the relationship between the superficial velocity and the pressure
drop. We did this for when it was acting as a packed bed as well as a fluidised bed.

When acting as a fluidised bed, we expected the increase in supercritical velocity to cause a steep
increase in pressure drop just before the point of fluidisation. The fact the curve then levels off after
this point is due to a larger force required initially to move the particles from stagnant than maintain
them suspended. To create such a graph, we adjusted the rotameter [shown in Fig. 1.1 below] to a
certain flowrate. The pressure drop was then measured.

We then took note as to how the fat bed performed as a fixed beds of solid particles. Seeing as the
force is constant, in theory, a linear relationship between the pressure drop and superficial velocity
should be recorded. We followed the same procedure to collect data for the bed performing in this
form as in a fluidised form.

When experimenting with the narrow bed, we were more interested in the effects of different sized
particles and how their pressure drop – supercritical velocity curves varied with one another. As the
particle size increased, the velocity at which they become fluidised would also be expected to
increase. Considering the small diameter of the bed, it is likely that it falls under the slugging
regime, where bubbles formed due to the gas fill the cross section of the reactor.

The bed height given by H at points where the superficial velocity, U, is greater than the minimum
superficial velocity required for fluidisation, Umf, can be used in the below equation:

Where Hmf is the bed height at minimum fluidisation. This is of interest to us as k 2 is a constant that

can be found from plotting a graph of the left hand side of the equation on the y axis and

on the x axis. The intercept of this graph will be k 2 -1. The other unknown, Uv ,
which will either be Ub in the fat bed or Us for the narrow bed, can be found from the
gradient.

The values found from Uv can then be compared to the theoretical minimum superficial
velocities for the bubbling regime:

5
 And for the slugging regime in the narrow bed:

And the accuracy of our experiment tested.

Furthermore, the bed voidage, ε, of the narrow at minimum fluidisation was to be found and
compared to literature values. The equation below was used for this:

Where m is the mass of the particles, A is the bed cross sectional area, and the density of the
particles.

We then will look at how Torricelli’s law, as stated below:

Predicts the velocity of the flow of sand through an orifice of the fat bed. An estimate for Co can be
made.

Apparatus: rotameter

Rotameter

Manometers

manometers

Thin Bed

6 Fat Bed
 Fig. 1.1: Schematic Diagram of Apparatus

Results:

For the full set of results including values recorded for the pressure drop and superficial velocity for
the fat bed acting as a packed and fluidised bed, and the narrow bed with particles of three different
grades, please see Appendix (1).

As expected, we did obtain a linear relationship for the packed bed. In our first run however, we
overshot the minimum fluidisation velocity by too much, causing too big a change in pressure,
essentially ruining our results. This first set can be found in the appendix (Table 1.1). The graph
produced when we repeated the procedure for a second time is shown below in Fig 2.1:

The results for the fat bed in its fluidised form are shown below in a log graph for clarity in Fig. 2.2:

7
For the narrow bed analysis, Fig 2.4 below sums up the relationships between particle size, pressure
drop, and superficial velocity:

As seen from above, the larger the particle in the fluidised bed, the higher the velocity needed to
fluidise it.

A small experiment to find the rate at which sand was expelled through an orifice was carried out on
the fat bed when acting as a packed bed. Figure 2.6 shows the relationship between the mass of
sand expelled and the corresponding bed height. It shows their inverse proportionality. Exact
results and further calculations can be found in Table 3.1.

8
The relationship between the velocity at which the sand was expelled and the corresponding bed
height is displayed below. The discharge coefficient was found (see table 3.0) using Torricelli’s law
and the average velocity and bed heights. It came out as 0.000513, close to theoretical values.

The relationship between bed height and superficial velocity for all the beds acting in their fluidised
forms was also explored. See figures 2.6 and 2.7 below.

9
In plotting equation [1], only the values in which U>Umf were used. For the fat bed, the graph
produced (found in Appendix) gives K2 as 35.53 and Uv as 0.667 m/s.

The narrow bed graph is also in the Appendix, and the results are discussed in the Discussion.

10
Discussion:

Fig. 2.6 shows how the coarse particles in the narrow bed didn’t even get a chance to fluidise,
showing how larger particles need higher minimum fluidisation velocities.

The value of Ub predicted with the equation, 0.650 m/s , is a very similar result to the experimental
value which comes is read off the graph (see Appendix Fig 5.1), at only 0.03% higher. We can
therefore conclude its validity.

For a slugging bed, i.e. the narrow bed, the value of Uv can also be concluded for the graphs (see
Appendix Fig 5.2). For the fine particles, it gives a value of 0.347 m/s. This is quite a bit higher than
that predicted by the equation (see Appendix 6), which gives 0.021 m/s. Air trapped in the bed and
inaccurate measuring could be responsible. For the medium particles, a closer value is attained –
0.027 m/s predicted by the equation is closely matched by 0.015 m/s from the graph.

The k2 value for the fat bed is 36.525. For the narrow bed, the fine and medium particles have
similar k2 values – 0.02 and 0.01 respectively. This is what is predicted by theory, so is a good result.
It does not however offer any further explanation for the offset in velocities for the fine particles, as
they should be better matched.

There were issues in calculating the bed voidage, as we forgot to take the weights of each particle
grade. However, had we taken them, the way in which it would have been calculated can be found
in the Appendix. We would imagine the results to be similar to theoretical values found: bed
voidages of 0.4 for the smaller particles and 0.78 for the medium particles.

11
Conclusion:

The above discussion can lead us to the conclusion that this was a fairly successful experiment
producing results mostly concurrent with theoretical values.

It has shown how by manipulating the operating conditions and materials used in the reactors,
optimum conditions specific to the desired process can be achieved, thus proving the value of
fluidised beds to industry.

References:

₁ http://faculty.washington.edu/finlayso/Fluidized_Bed/FBR_Intro/dollars_scroll.htm

₂ http://en.wikipedia.org/wiki/Fluidized_bed_reactor#History_and_current_uses

 Davidson, J.F and Harrison, D., Fluidisation, Chapters 1 and 2

12