Lesson 6

Live your life. Create your destiny.

Department of Chemical, Metallurgical & Materials Engineering

Open Rubric
 Chapter 6: MECHANICAL-PHYSICAL
SEPARATION
Fluid Flow Through a Packed Bed of
Particles
Introduction

– Chemical engineering operations commonly involve

the use of packed and fluidized beds.
• These are devices in which a large surface area
for contact between a liquid and a gas (absorption,
distillation); or
• a solid and a gas or liquid (adsorption, catalysis) is
obtained for achieving rapid and heat transfer.
Objectives of this lesson:

– Derive the Carman-Kozeny equation with the help of

Darcy and Hagen Poiseuille’s equations.
– Use the Carman-Kozeny equation to determine the
pressure drop required for the liquid to flow through
the column at specified flow rates.
– The use of frictional factor in laminar and turbulent
flows.
– Fundamentals of filtration.
– Problem solving involving filtration.
Definition:

– A packed bed is a cylindrical column that is filled

with a suitable packing material.
– The treated liquid is distributed as uniformly as
possible at the bottom of the column and flows
upwards, wetting the packing material.
– The packed solids contain some species that are
soluble in the liquid, either absorbed or adsorbed in
the liquid. In the case of water treatment, the clean
water is taken out at the bottom.
Schematic diagram of fixed bed column bioreactor, packed with bgh designed to
function as a continuous flow system for biosorptionbiosorption of cadmium. (1) metal
reservoir; (2) peristaltic pump; (3) flow control; (4) glass column reactor; (5) bgh
biosorbent; (6) enlarged view of column packing; (7) extention for secondary column;
and (8) effluent storage.
– When a fluid passes through a porous material, it
flows through long thin passage of varying cross
section.

– The problem is how to calculate the flow rate based

on the nominal thickness of the layer. However, this
problem was tackled by Kozeny and later Carman.
Scrubbing/gas cleaning
Extraction Unit
Different types of packings
Distillation
Adsorption tower

Bed of
adsorbent
Ion exchange tower
Catalytic bed
6.1. Pressure Drop-Flow Relationship
6.1.1. Laminar Flow
– In the nineteenth century Darcy (1856) observed that the flow of
water through a packed bed of sand was governed by the
following expression:

(1)

Where U is the superficial fluid velocity through the bed and (-∆p) is
the frictional pressure drop across a bed depth H.
Q  fluid volumetri c flow rate 
Superficia l velocity   
A  cross - sectional area of bed 
– The flow of a fluid through a packed bed of solid particles may
be analysed in terms of the fluid flow through tubes. Therefore,
the starting point is the Hagen-Poiseuille equation for
laminar flow through a tube:

(2)

Where D is the tube diameter and µ is the fluid viscosity.

– Now, consider the packed bed to be equivalent to many tubes of

equivalent diameter De following tortuous paths of equivalent
length He and carrying fluid with a velocity Ui. Then, from
equation (2),

Ui is the actual velocity of fluid through the interstices of the packed

– Kozeny modelled the layer as many small capillary tubes of
diameter D making a layer of cross-sectional area A.

The actual cross-sectional area for the flow path is A’.

The
 A' 
 
ratio:  A  is called the porosity
( ) of the material. And the
volume flow rate through the layer is Q.

Kozeny used the notion that Q = AU, where U is the superficial

velocity. Q Q
Ui  ,U  (3)
Ai A

Dividing the two expressions in (3) yields,

A U
U i U  (4)
Aii 

The actual velocity is related to the superficial velocity and the

porosity through equation (4).
– Carman modified this formula (equation (4)) when he realised
that the actual velocity inside the tubes must be proportionally
larger because the actual length is greater than the layer
thickness.
ULi (5)
U 
It follows that: i L

Therefore, for equivalent diameters:

 p 32 U i 32   ULii  (6)
 2
 2  
He De De  L 

– Determining the porosity relationship with the volume of the

tube (Vi).
A AL V V (7)
 i  i  i  i
A AL VT Vi  VS

VSS (8)
Vi 
(1   )
To find equivalent diameters, De, divide both sides by the
surface area of the tube (S)

Vi VS (9)

S S (1   )

Vi (volume of the tube) is made of the diameter De and length Li.

De2 Lii
So, Vi  Ai Li  (10)
4
The surface area of the tube:

S De Li (11)
– Determining the diameter, divide the volume by the surface area.

Vi De2 Li De
  (12)
S 4(De Li ) 4

Equating equation (12)  S

Deto equation
V (9) yields
 (13)

4 S (1   )

4VS 4  VS  (14)

 De    
S (1   ) (1   )  S 
– Assuming that the solids are spherical particles of diameter
ds,
d s3 (15)
Sphere volume: VS 
6
(16)
Sphere surface area: S d s2

VS d s3 ds (17)
Volume/surface area:  2

S 6d s 6

Substituting equation (17) into equation (14) gives the equivalent-

diameter of the tube:
4  VS  4  d s  2d ss (18)
De     
(1   )  S  (1   )  6  3(1   )
Now, substituting the tube diameter De to Hagen-Poiseuille’s
equation (6) for equivalent diameters gives the:
Carman-Kozeny equation:

( p ) U (1   ) 2
180 2 3
or
He ds 
(19)
6.1.2. Turbulent Flow
– For turbulent flow through a randomly packed bed of mono-
ized spheres of diameter ds the equivalent equation is:

2
( p )  f U (1   )
1.75 2 3
He d 
(20)
or
6.1.3. General Equation for Turbulent and Laminar
Flow
– Based on extensive experimental data covering a wide range of
size and shape of particles, Ergun (1952) suggested the
following general equation for any flow conditions:

(21)

This is known as the Ergun equation for flow through a

randomly packed bed of spherical particles of diameter x or ds.
– Under laminar conditions, the first term dominates
and the equation reduces to the Carman-Kozeny
equation (150 instead of 180).

– Under turbulent flow conditions, the second term

dominates; the pressure gradient increases as
the square of superficial fluid velocity and is
independent of fluid viscosity.
• Reynolds number

* xU f
R 
(22)
e
 (1   )
*
Fully laminar conditions exist for R  10
e

Fully turbulent conditions exist for Re*  2000

– Ergun also expressed flow through a packed bed in terms of a
friction factor.

(23)

Equation (21) reduces to

with
6.1.4. Non-spherical Particles
– The Ergun and Carman-Kozeny equations also
accommodate non-spherical particles if x is replaced by xsv
the diameter of a sphere having the same surface to volume
ratio as the non-spherical particles in question.

Ergun equation:
(24)

Carman-Kozeny equation:
(25)
6.2. SOLID – LIQUID Separation:
Filtration
6.2.1. Introduction

Filtration is any of various mechanical, physical or biological

operations that separate solids from fluids (liquids or gases) by adding a
medium through which only the fluid can pass. The fluid that passes
through is called the filtrate. In physical filters oversize solids in the fluid
are retained. It involves the build-up of a bed or ‘cake’ of particles on a
porous surface known as the filter medium, which commonly takes the
form of a woven fabric.
– Pore size of the medium is less than the size of the particles to be
filtered.
– This can be analysed in terms of the flow of fluid through a packed bed
of particles, the depth of which is increasing with time.
– In practice, voidage, ɛ of the cake may also change with time, i.e.
compressible cake.
As an example of the
application for flow
through packed beds
of particles – cake
filtration is
considered.

Cake filtration is widely

used in industry to
separate solid particles
from suspension in
liquid.
Industrial applications: Different kinds of filters

• Filter press-brewery, slurries

• Membrane filters-water treatment
• Deep bed filtration- water treatment
• Drum filters- improving consistency of
pulp in paper mills
• (check other pulps such as juice industry).
Main factors to be considered when selecting
equipment and operating conditions are:

• The properties of the fluid, particularly its viscosity, density and

corrosive properties.
• The nature of the solid – its particle size and shape, size
distribution, and packing characteristics.
• The concentration of solids in suspension.
• The quantity of material to be handled, and its value.
• Whether the valuable product is the solid, the fluid, or both.
• Whether it is necessary to wash the filtered solids.
• Whether very slight contamination caused by contact of the
suspension or filtrate with the various components of the
equipment is detrimental to the product.
• Whether the feed liquor may be heated.
• Whether any form of pre-treatment might be help.
– The most important factors on which the
Rate of Filtration depends on are:

• The drop in pressure from the feed to the

far side of the filter medium.
• The area of the filtering surface.
• The viscosity of the filtrate.
• The resistance of the filter cake.
• The resistance of the filter medium and
initial layers of cake.
6.2.2. Incompressible Cake
– First case scenario, ignore the filter medium and consider only
the cake itself; hence the pressure drop versus liquid flow
relation is still described by the Ergun equation.
– Particle size and range of liquid flow and properties commonly
used in industry give rise to laminar flow and so the second term
(turbulent) vanishes.

Therefore, for a given slurry (with fixed particle properties) the

resulting cake resistance is defined as:
(26)

and so equation (21) becomes

(27)
– Given that V is the volume of filtrate (liquid) passed in
a time t and dV/dt is the instantaneous volumetric flow
rate of filtrate at time t, then:
– Particle size and range of liquid flow and properties
commonly used in industry give rise to laminar flow
and so the second term (turbulent) vanishes.

Therefore, for a given slurry (with fixed particle

properties) the resulting cake resistance is defined as:

(29)
– Each unit volume of filtrate is assumed to deposit a
certain mass of particles, which form a certain volume of
cake.

– This is expressed as , the volume of cake formed by the
passage of unit volume of filtrate.

(30)

and equation (27) becomes

(31)
• Constant Rate Filtration
– If the filtration rate dV/dt is constant, the pressure drop across
the filter cake will increase in direct proportion to the volume of
filtrate passed V.
• Constant Pressure Drop Filter:
– If (-∆p) is constant the,

or integrating equation (31)

(32)

where
(33)
6.2.3. Including the Resistance of Filter Medium
– Considering operation at constant pressure drop, which is the
most common case,

(34)

6.2.4. Washing the Cake

– SELF STUDY!!!
6.2.5. Compressible Cake

– In practice many materials give rise to compressible

filter cakes.
– In compressible cake rc increases with (-∆p).
– Change in rc is due mainly to the effect on the cake
voidage, ɛ.
– The force on the particles causes the particle packing
to become more dense, i.e. cake voidage decreases.
– For soft particles – size or shape may change, adding
to the increase in cake resistance.
– Consider an element of the filter cake of thickness dL across
which the pressure drop is dp. Applying Carman-Kozeny
equation for flow through this element:
 (35)

Where rc is the resistance of this element of the cake. For a

compressible cake, rc is a function of the pressure difference
between the upstream surface of the cake and the element (i.e.
referring to figure 6.3).
(36)

(37)

And equation (35) becomes

(38)

Relationship between rc and ps must be found from laboratory

experiments before equation (36) can be used in design.
EXAMPLE:
A leaf filter has an area of 0.5 m2 and operates at a constant pressure
drop of 500 kPa. The following test results were obtained for a slurry
in water which gave rise to a filter cake regarded as incompressible:

Volume of filtrate collected (m3) 0.1 0.2 0.3 0.4 0.5

Time (s) 140 360 660 1040 1500

Calculate:
a) The time needed to collect 0.8 m3 of filtrate at a constant pressure
drop of 700 kPa;
b) The time required to wash the resulting cake with 0.3 m3 of water at
a pressure drop of 400 kPa.
SOLUTION
For constant pressure drop:

t rc rc
 2 V 2 Veq
V 2 A (  p ) A (  p )
Plot of t/V versus V yields a straight line.

Gradient = rc
2 A 2 (  p )

Y interception = rc
2
Veq
A ( p )
SOLUTION cont...
Using the given data:
V (m3) 0.1 0.2 0.3 0.4 0.5
t/V (m3/s) 1400 1800 2200 2600 3000

∆Y

∆X
SOLUTION cont...
rc
From the plot: 2
2 A (  p )
Gradient: = 4000

rc
2
Veq
Y interception: A ( p ) = 1000

Remembering that:
A = 0.5 m2
(-∆p) = 500 x 103 Pa
SOLUTION cont...
Therefore: rc
2
4000
2 A (  p )
rc
2 3
4000
2(0.5) (500 10 )
 rc 1109 Pa s / m 2
and

rc
2
Veq 1000
A ( p )
1109
2 3
Veq 1000
(0.5) (500 10 )
Veq 0.125m3
SOLUTION cont...
a)

t rc rc
 2 V 2 Veq
V 2 A ( p ) A ( p )
t 1 109 1109
 2 3
(0.8)  2 3
(0.125)
(0.8) 2(0.5) (700 10 ) (0.5) (700 10 )
 t 2400 s(40 min)
SOLUTION cont...
b) Filtration rate at the end of the filtration period is:

1 dV ( p ) A

A dt rc (V  Veq )
1 dV (700 103 )(0.5)

(0.5) dt (1 109 )( 0.8  0.125)
dV
 1.89 10  4 m3 / s
dt
Assuming the wash water has same physical properties as the filtrate,
and noting that:

dV
 ( p)
dt
SOLUTION cont...
Therefore;

4  400 103 
Flow rate of wash water (at 400 kPa) 1.89 10  
3 
 700 10 
1.08 10  4 m 3 / s
Finally;

t V 0.3  4 2778s ( 46.3 min)

dV 1.08 10
dt
 CLASS EXERCISE
A leaf filter has an area of 2 m2 operates at a constant pressure drop of
250 kPa. The following results were obtained during a test with an
incompressible cake:

Volume of filtrate collected (litre) 280 430 540 680 800

Time (min) 10 20 30 45 60

Calculate:
a) The time required to collect 1200 litre of filtrate at a constant
pressure drop of 400 kPa;
b) The time required to wash the resulting cake with 500 litre of water
(assume same properties as the filtrate) at a pressure drop of 200
kPa.
SOLUTION
For constant pressure drop:

t r  r 
 2c V  2c Veq
V 2 A (  p ) A ( p )

Plot of t/V versus V yields a straight line.

rc
Gradient: 2
2 A ( p )

Y interception: rc
2
Veq
A ( p )
SOLUTION cont...
Using the given data:
V (m3) 0.28 0.43 0.54 0.68 0.8
t/V (m3/s) 2142.86 2790.7 3333.33 3970.59 4500

∆Y

∆X
SOLUTION cont...
From the plot:

rc
Gradient: 2
= 4566
2 A ( p )

Y interception: rc = 854.1

2
Veq
A ( p )
Remembering that:
A = 2 m2
(-∆p) = 250 x 103 Pa
SOLUTION cont...
Therefore:
rc
4566
2 A2 ( p)
rc
4566
2(2) 2 (250 103 )
 rc 9.132 10 9 Pa s / m 2

and
rc
Veq 854.1
A 2 (  p )
9.132 109
Veq 854.1
(2) 2 (250 103 )
Veq 0.094m 3
SOLUTION cont...
a)
t r  r 
 2c V  2c Veq
V 2 A ( p) A ( p )
t 9.132 10 9 9.132 109
 2 3
(1.2)  2 3
(0.094)
(1.2) 2(2) (400 10 ) (2) (400 10 )
 t 4753.2 s (79.2 min)
SOLUTION cont...
b) Filtration rate at the end of the filtration period is:

1 dV ( p ) A

A dt rc (V  Veq )
1 dV (400 103 )( 2)

(2) dt (9.132 109 )(1.2  0.094)
dV
 1.354 10  4 m 3 / s
dt
Assuming the wash water has same physical properties as the filtrate,
and noting that:

dV
 (  p )
dt
SOLUTION cont...
Therefore;

4  200 103 
Flow rate of wash water (at 200 kPa) 1.354 10  
3 
 400 10 
6.77 10  5 m 3 / s

Finally;

t V 0.5  5 7385.52 s (123.1 min)

dV 6.77 10
dt
HOMEWORK (EXERCISE 6.9)
A leaf filter has an area of 1.73 m2, operates at a constant pressure
drop of 300 kPa and produces the following results during a test on
filtration of a slurry:

Volume of filtrate collected (m3) 0.19 0.31 0.41 0.49 0.56 0.63

Time (s) 300 600 900 1200 1500 1800

a) Calculate the time needed to collect 1 m3 of filtrate during

filtration of the same slurry at a constant pressure drop of 400 kPa.
b) Calculate the time required to wash the resulting filter cake with
0.8 m3 of water at a constant pressure drop of 250 kPa.

(Assume the cake is incompressible and that the flow properties of the
filtrate are the same as those of the wash solution)