ENCH3FS Fluid and Solids Transport

Section ‘B’
Flow through Porous Media:
Filtration

Dr S. Moodley1

1 Discipline
of Chemical Engineering
UKZN Howard College Campus

Semester 1, 2024

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 1 / 67
Outline
1 Introduction
2 Theory of filtration
Pressure drop of fluid through filter cake
Specific cake resistance
3 Constant-pressure filtration
4 Example 1
5 Constant-rate filtration
6 Constant-rate followed by constant pressure filtration
7 Example 2
8 Variable-rate + variable-pressure filtration
9 Example 3
10 Compressible cakes
11 Example 4
12 Washing of filter cakes
13 Example 5
14 Example 6
15 Filtration cycle and its optimization
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 2 / 67
Introduction

We shall consider separation processes

that are not due to differences among the
molecules; rather, the separations will be
realized through ‘mechanical-physical
forces’

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 3 / 67
Introduction

We shall consider separation processes

that are not due to differences among the
molecules; rather, the separations will be
realized through ‘mechanical-physical
forces’
Examples of such forces are gravitational
and centrifugal, amongst others

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 3 / 67
Introduction

We shall consider separation processes

that are not due to differences among the
molecules; rather, the separations will be
realized through ‘mechanical-physical
forces’
Examples of such forces are gravitational
and centrifugal, amongst others
Filtration is the separation of solid
particles from their liquid

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 3 / 67
Introduction

We shall consider separation processes

that are not due to differences among the
molecules; rather, the separations will be
realized through ‘mechanical-physical
forces’
Examples of such forces are gravitational
and centrifugal, amongst others
Filtration is the separation of solid
particles from their liquid
A key requirement for achieving filtration
is a pressure drop

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 3 / 67
Introduction (contd.)

Solids suspended in a liquid (which

together constitute the slurry ) or a gas
are physically/mechanically removed using
a porous medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 4 / 67
Introduction (contd.)

Solids suspended in a liquid (which

together constitute the slurry ) or a gas
are physically/mechanically removed using
a porous medium
The porous medium retains the solids as a
cake, and a clear filtrate passes through

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 4 / 67
Introduction (contd.)

Solids suspended in a liquid (which

together constitute the slurry ) or a gas
are physically/mechanically removed using
a porous medium
The porous medium retains the solids as a
cake, and a clear filtrate passes through
Either the solid cake or filtrate may be the
valuable product

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 4 / 67
Introduction (contd.)

Solids suspended in a liquid (which

together constitute the slurry ) or a gas
are physically/mechanically removed using
a porous medium
The porous medium retains the solids as a
cake, and a clear filtrate passes through
Either the solid cake or filtrate may be the
valuable product
The resistance to flow increases as the
cake builds up

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 4 / 67
Introduction (contd.)
Cake filtration

Solids are deposited as a “cake” on top of

a thin filter medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 5 / 67
Introduction (contd.)
Cake filtration

Solids are deposited as a “cake” on top of

a thin filter medium
Low pressure drop over filter medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 5 / 67
Introduction (contd.)
Cake filtration

Solids are deposited as a “cake” on top of

a thin filter medium
Low pressure drop over filter medium
Particles both big and small create
passages that remove even smaller
particles in the slurry/suspension

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 5 / 67
Introduction (contd.)
Cake filtration

Solids are deposited as a “cake” on top of

a thin filter medium
Low pressure drop over filter medium
Particles both big and small create
passages that remove even smaller
particles in the slurry/suspension
The filter cake that is formed acts like an
additional filter medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 5 / 67
Introduction (contd.)
Deep bed filtration

Solids are deposited inside the filter

medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 6 / 67
Introduction (contd.)
Deep bed filtration

Solids are deposited inside the filter

medium
This mechanism is favored when cake
deposition is undesirable

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 6 / 67
Introduction (contd.)
Deep bed filtration

Solids are deposited inside the filter

medium
This mechanism is favored when cake
deposition is undesirable
Particles are smaller than the filter
medium’s openings; they travel through
long pores and accumulate in the filter
medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 6 / 67
Introduction (contd.)
Deep bed filtration

Solids are deposited inside the filter

medium
This mechanism is favored when cake
deposition is undesirable
Particles are smaller than the filter
medium’s openings; they travel through
long pores and accumulate in the filter
medium
Pressure drop across a deep bed filter is
usually larger than that across a
surface/cake filter

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 6 / 67
Pressure drop of fluid through filter cake

We begin by recalling the Carman & Kozeny equation for laminar flow in a packed bed viz.

∆Pc k1 µ uc (1 − ε)2 S02

− = (1)
L ε3
where k1 is a constant (= 4.17 for random particles), µ is the filtrate viscosity (in Pa·s), ε is
the cake’s porosity (void fraction), L is the cake thickness (in m), S0 (= av ) is the specific
surface area of the particle (m−1 ), and −∆Pc is the pressure drop across the cake (in N/m2 ).
We note that the superficial velocity (originally defined as uc = Q ÷ A) may be re-written as

1 dV
uc = (2)
A dt
where A is the filter area (in m2 ) and V is the total amount of filtrate (in m3 ) collected up to
time t (in s).

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 7 / 67
Pressure drop of fluid through filter cake (contd.)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 8 / 67
Pressure drop of fluid through filter cake (contd.)

Let cs be the amount of solids per unit volume of filtrate (kg solids/m3 filtrate); we relate L to
V via a material balance:

LA (1 − ε) ρp = cs (V + εLA) (3)
where ρp is the particle density (in kg/m3 ). This is rearranged to

cs V
L= (4)
A (1 − ε) ρp
Now substitute (2) into (1), followed by replacing L in the resulting equation with (4).

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 9 / 67
Pressure drop of fluid through filter cake (contd.)

We obtain

1 dV −∆Pc −∆Pc
= 2
= (5)
A dt k1 (1 − ε) S0 µcs V µcs V
α
ρp ε 3 A A
where α = k1 (1 − ε) S02 ÷ ρp ε3 is the specific cake resistance (in m/kg). By analogy to
Equation (5) we define the filter medium resistance as

1 dV −∆Pf
= (6)
A dt µRm
where Rm is the filter medium resistance (in m−1 ) and −∆Pf is the corresponding pressure
drop.

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 10 / 67
Pressure drop of fluid through filter cake (contd.)

Both resistances are in series, which means Equations (5) and (6) can be combined to yield

1 dV −∆P
=   (7)
A dt αcs V
µ + Rm
A
where −∆P is the total pressure drop over the cake and filter medium.
Aside:
The amount of accumulated dry cake solids ‘W ’ (in kg) is related to V as
ρcx
W = cs V = V (8)
1 − mcx

where m is the mass ratio of wet-to-dry cake, ρ is the filtrate density (in kg/m3 ) and cx is the
mass fraction of solids in the slurry.

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 11 / 67
Specific cake resistance

We’ve seen that α = α (ε, S0 , P) (∵ P affects ε) =⇒ α (−∆P) can be determined at

constant P

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 12 / 67
Specific cake resistance

We’ve seen that α = α (ε, S0 , P) (∵ P affects ε) =⇒ α (−∆P) can be determined at

constant P
If α 6= α (−∆P) =⇒ incompressible sludge

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 12 / 67
Specific cake resistance

We’ve seen that α = α (ε, S0 , P) (∵ P affects ε) =⇒ α (−∆P) can be determined at

constant P
If α 6= α (−∆P) =⇒ incompressible sludge
Typically, though, α ↑ as −∆P ↑ i.e. cake is compressible

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 12 / 67
Specific cake resistance

We’ve seen that α = α (ε, S0 , P) (∵ P affects ε) =⇒ α (−∆P) can be determined at

constant P
If α 6= α (−∆P) =⇒ incompressible sludge
Typically, though, α ↑ as −∆P ↑ i.e. cake is compressible
Empirical observation: α = α0 × (−∆P)s (where α0 and s are constants)
Question: What does s = 0 imply?

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 12 / 67
Specific cake resistance

We’ve seen that α = α (ε, S0 , P) (∵ P affects ε) =⇒ α (−∆P) can be determined at

constant P
If α 6= α (−∆P) =⇒ incompressible sludge
Typically, though, α ↑ as −∆P ↑ i.e. cake is compressible
Empirical observation: α = α0 × (−∆P)s (where α0 and s are constants)
Question: What does s = 0 imply?
h 0
i
Alternatively, α = α00 1 + β (−∆P)s (where α00 , β and s 0 are constants)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 12 / 67
Constant-pressure filtration

Inverting Equation (7) results in

dt
= Kp V + B (9)
dV
where
µαcs
Kp = 2
(10)
A (−∆P)
and

µRm
B= (11)
A (−∆P)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 13 / 67
Constant-pressure filtration (contd.)

At conditions of constant (P, α) and an incompressible cake, t and V are variables in

Equation (9); we now separate these variables and integrate Equation (9) to obtain
Z t Z V
0
Kp V 0 + B dV 0


dt = (12)
0 0
whereafter a trivial rearrangement (...try it!) yields

t Kp
= V +B (13)
V 2
Question: How can Equation (13) be of practical use?

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 14 / 67
Worked example
Example 1

Example
Data for the laboratory filtration of CaCO3 slurry in water at 298.2 K at a constant pressure
(−∆P) of 338 kN/m2 are reported in the following table. The filter area was A = 0.0439 m2
and the slurry concentration was cs = 23.47 kg/m3 . Calculate α and Rm . At 298.2 K the
viscosity of water is 8.937 × 10−4 Pa·s
t (s) 4.4 9.5 16.3 24.6 34.7 46.1 59.0 73.6 89.4 107.3
V × 10 (m3 )
3
0.498 1.000 1.501 2.000 2.498 3.002 3.506 4.004 4.502 5.009
1 We begin by transforming the data to t ÷ V versus V

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 15 / 67
Worked example
Example 1

Example
Data for the laboratory filtration of CaCO3 slurry in water at 298.2 K at a constant pressure
(−∆P) of 338 kN/m2 are reported in the following table. The filter area was A = 0.0439 m2
and the slurry concentration was cs = 23.47 kg/m3 . Calculate α and Rm . At 298.2 K the
viscosity of water is 8.937 × 10−4 Pa·s
t (s) 4.4 9.5 16.3 24.6 34.7 46.1 59.0 73.6 89.4 107.3
V × 10 (m3 )
3
0.498 1.000 1.501 2.000 2.498 3.002 3.506 4.004 4.502 5.009
1 We begin by transforming the data to t ÷ V versus V
2 Following the transformation, we plot V (independent variable) against t ÷ V , and draw
a ‘best-fit’ line through the discrete data points

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 15 / 67
Worked example
Example 1

Example
Data for the laboratory filtration of CaCO3 slurry in water at 298.2 K at a constant pressure
(−∆P) of 338 kN/m2 are reported in the following table. The filter area was A = 0.0439 m2
and the slurry concentration was cs = 23.47 kg/m3 . Calculate α and Rm . At 298.2 K the
viscosity of water is 8.937 × 10−4 Pa·s
t (s) 4.4 9.5 16.3 24.6 34.7 46.1 59.0 73.6 89.4 107.3
V × 10 (m3 )
3
0.498 1.000 1.501 2.000 2.498 3.002 3.506 4.004 4.502 5.009
1 We begin by transforming the data to t ÷ V versus V
2 Following the transformation, we plot V (independent variable) against t ÷ V , and draw
a ‘best-fit’ line through the discrete data points
3 We note that from Equation (13), the slope of the line is Kp ÷ 2 and the t ÷ V -axis
intercept is B
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 15 / 67
Worked example (contd.)

Transformed data:
V (m3 ) 4.98E-04 1.00E-03 1.50E-03 2.00E-03 2.50E-03 3.00E-03 3.51E-03 etc.
t ÷ V (s/m3 ) 8835 9500 10859 12300 13891 15356 16828 etc.

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 16 / 67
Worked example (contd.)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 17 / 67
Worked example (contd.)

1 The grey line is fitted to all of the data points

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 18 / 67
Worked example (contd.)

1 The grey line is fitted to all of the data points

2 The purple line is fitted to all of the data points excluding the first – this is usually done
since the first few points are usually outliers as regards a straight line

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 18 / 67
Worked example (contd.)

1 The grey line is fitted to all of the data points

2 The purple line is fitted to all of the data points excluding the first – this is usually done
since the first few points are usually outliers as regards a straight line
3 Using the purple line, we compute the slope Kp ÷ 2 ≈ 3 × 106 s/m6 and the
(t ÷ V )-intercept B ≈ 6400 s/m3 . Thus, Kp = 6 × 106 s/m6

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 18 / 67
Worked example (contd.)

1 The grey line is fitted to all of the data points

2 The purple line is fitted to all of the data points excluding the first – this is usually done
since the first few points are usually outliers as regards a straight line
3 Using the purple line, we compute the slope Kp ÷ 2 ≈ 3 × 106 s/m6 and the
(t ÷ V )-intercept B ≈ 6400 s/m3 . Thus, Kp = 6 × 106 s/m6
4 Rearranging Equations (10) and (11) and solving for α and Rm , respectively, results in

Kp A2 (−∆P)
α= = 1.863 × 1011 m/kg
µcs
and
BA (−∆P)
Rm = = 10.63 × 1010 m−1
µ

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 18 / 67
Constant-rate filtration
We begin by recalling Equation (7), the general filtration equation, i.e.
1 dV −∆P
=  
A dt αcs V
µ + Rm
A

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 19 / 67
Constant-rate filtration
We begin by recalling Equation (7), the general filtration equation, i.e.
1 dV −∆P
=  
A dt αcs V
µ + Rm
A
“Constant-rate” filtration implies that
dV
= a constant
dt
and also that −∆P is no longer constant

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 19 / 67
Constant-rate filtration
We begin by recalling Equation (7), the general filtration equation, i.e.
1 dV −∆P
=  
A dt αcs V
µ + Rm
A
“Constant-rate” filtration implies that
dV
= a constant
dt
and also that −∆P is no longer constant
The general filtration equation can then be rearranged to
   
µαcs dV µRm dV
− ∆P = V+ = Kv V + C (14)
A2 dt A dt

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 19 / 67
Constant-rate filtration
We begin by recalling Equation (7), the general filtration equation, i.e.
1 dV −∆P
=  
A dt αcs V
µ + Rm
A
“Constant-rate” filtration implies that
dV
= a constant
dt
and also that −∆P is no longer constant
The general filtration equation can then be rearranged to
   
µαcs dV µRm dV
− ∆P = V+ = Kv V + C (14)
A2 dt A dt
If the cake is incompressible, then Kv and C are constants
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 19 / 67
Constant-rate filtration (contd.)

The pressure drop increases with increasing cake thickness and the filtrate volume that is
collected

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 20 / 67
Constant-rate filtration (contd.)

The pressure drop increases with increasing cake thickness and the filtrate volume that is
collected
We note that the total filtrate volume “V ” that is collected after time “t” is
dV
V =t
dt

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 20 / 67
Constant-rate filtration (contd.)

The pressure drop increases with increasing cake thickness and the filtrate volume that is
collected
We note that the total filtrate volume “V ” that is collected after time “t” is
dV
V =t
dt
Now we can re-write Equation (14) in a form where the pressure drop depends on time

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 20 / 67
Constant-rate filtration (contd.)

The pressure drop increases with increasing cake thickness and the filtrate volume that is
collected
We note that the total filtrate volume “V ” that is collected after time “t” is
dV
V =t
dt
Now we can re-write Equation (14) in a form where the pressure drop depends on time
The resulting equation is
"  2 #  
µαcs dV µRm dV
− ∆P = t+ (15)
A2 dt A dt

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 20 / 67
Constant-rate filtration (contd.)

For a compressible cake, we may express the pressure drop across the cake (based on
Equation (15)) as
µcs V 2
 
∆Pc
− = (16)
α t A
Question: What is the significance of placing α on the left-hand-side in the preceding
equation?

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 21 / 67
Constant-rate filtration (contd.)

For a compressible cake, we may express the pressure drop across the cake (based on
Equation (15)) as
µcs V 2
 
∆Pc
− = (16)
α t A
Question: What is the significance of placing α on the left-hand-side in the preceding
equation?
Recall that an empirically-based expression for α is

α = α0 × (−∆Pc )s

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 21 / 67
Constant-rate filtration (contd.)

For a compressible cake, we may express the pressure drop across the cake (based on
Equation (15)) as
µcs V 2
 
∆Pc
− = (16)
α t A
Question: What is the significance of placing α on the left-hand-side in the preceding
equation?
Recall that an empirically-based expression for α is

α = α0 × (−∆Pc )s

Substituting for α in Equation (16), and noting that (−∆P) = (−∆Pc ) + (−∆Pm ), we
obtain  2
V
[(−∆P) − (−∆Pm )]1−s = α0 µcs t (17)
A
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 21 / 67
Constant-rate filtration (contd.)

To correct the total/overall pressure drop for the not-that-easy-to-calculate filter

medium’s pressure drop, assume that the latter’s resistance is constant

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 22 / 67
Constant-rate filtration (contd.)

To correct the total/overall pressure drop for the not-that-easy-to-calculate filter

medium’s pressure drop, assume that the latter’s resistance is constant
From Equation (6), an easy rearrangement yields

(−∆Pm ) = Rm µuc

which implies that the pressure drop over the medium is also constant

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 22 / 67
Constant-rate filtration (contd.)

To correct the total/overall pressure drop for the not-that-easy-to-calculate filter

medium’s pressure drop, assume that the latter’s resistance is constant
From Equation (6), an easy rearrangement yields

(−∆Pm ) = Rm µuc

which implies that the pressure drop over the medium is also constant
Equation (17) now reduces to

[(−∆P) − (−∆Pm )]1−s = Kr t (18)

where Kr = µuc2 cs α0

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 22 / 67
Constant-rate filtration (contd.)

A simple logarithmic transformation of Equation (18) yields

ln (t) = (1 − s) ln [(−∆P) − (−∆Pm )] − ln (Kr ) (19)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 23 / 67
Constant-rate filtration (contd.)

A simple logarithmic transformation of Equation (18) yields

ln (t) = (1 − s) ln [(−∆P) − (−∆Pm )] − ln (Kr ) (19)

Plotting [(−∆P) − (−∆Pm )] against t as a log-log graph should result in a straight line
having slope ‘1 − s’

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 23 / 67
Constant-rate filtration (contd.)

A simple logarithmic transformation of Equation (18) yields

ln (t) = (1 − s) ln [(−∆P) − (−∆Pm )] − ln (Kr ) (19)

Plotting [(−∆P) − (−∆Pm )] against t as a log-log graph should result in a straight line
having slope ‘1 − s’
Typically, though, the experimental data is t vs (−∆P) i.e. (−∆Pm ) is not known a priori

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 23 / 67
Constant-rate filtration (contd.)

A simple logarithmic transformation of Equation (18) yields

ln (t) = (1 − s) ln [(−∆P) − (−∆Pm )] − ln (Kr ) (19)

Plotting [(−∆P) − (−∆Pm )] against t as a log-log graph should result in a straight line
having slope ‘1 − s’
Typically, though, the experimental data is t vs (−∆P) i.e. (−∆Pm ) is not known a priori
(−∆Pm ) can be estimated by plotting (−∆P) against t in Cartesian coordinates;
(−∆Pm ) is then estimated by extrapolation to the (−∆P) axis

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 23 / 67
Constant-rate filtration (contd.)

A simple logarithmic transformation of Equation (18) yields

ln (t) = (1 − s) ln [(−∆P) − (−∆Pm )] − ln (Kr ) (19)

Plotting [(−∆P) − (−∆Pm )] against t as a log-log graph should result in a straight line
having slope ‘1 − s’
Typically, though, the experimental data is t vs (−∆P) i.e. (−∆Pm ) is not known a priori
(−∆Pm ) can be estimated by plotting (−∆P) against t in Cartesian coordinates;
(−∆Pm ) is then estimated by extrapolation to the (−∆P) axis
If using the value of (−∆Pm ) obtained in the point above results in a straight line on the
log-log plot, then we can accept the estimated value of (−∆Pm )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 23 / 67
Constant-rate followed by constant pressure filtration
Incompressible cake

Typically, high initial flow rates through a clean filter medium are avoided; this prevents
the solids penetrating the clean filter medium

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 24 / 67
Constant-rate followed by constant pressure filtration
Incompressible cake

Typically, high initial flow rates through a clean filter medium are avoided; this prevents
the solids penetrating the clean filter medium
Contamination of the filtrate, and uneven deposition of the cake are thus avoided

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 24 / 67
Constant-rate followed by constant pressure filtration
Incompressible cake

Typically, high initial flow rates through a clean filter medium are avoided; this prevents
the solids penetrating the clean filter medium
Contamination of the filtrate, and uneven deposition of the cake are thus avoided
Now, if the constant-pressure mode begins only after an initial constant-rate duration of
ts (at which time Vs amount of filtrate has been collected), we re-integrate the
constant-pressure equation i.e. Equation (12) as follows:
Z t Z V
0
Kp V 0 + B dV 0


dt = (20)
ts Vs

which results in (you should be able to perform this simple exercise!)

t − ts Kp
= (V + Vs ) + B (21)
V − Vs 2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 24 / 67
Constant-rate followed by constant pressure filtration (contd.)

For t < ts ,
−∆P = µαcs uc2 t + (µRm uc )

which follows from Equation (15)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 25 / 67
Constant-rate followed by constant pressure filtration (contd.)

For V ≤ Vs ,
V = Qt
while for V > Vs , Equation (21) is applicable
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 26 / 67
Worked example
Example 2

Example
Filtration tests were carried out with a plate and frame filter press under the following
conditions: solids: ρp = 2710 kg/m3 ; liquid: µ = 0.001 Pa·s; suspension: c = 10 kg/m3 ; filter:
plate-and-frame press, 1 frame, dimensions 430 × 430 × 30 mm. The experimental data are in
the PDF file “Example 2.pdf”. Determine the specific cake resistance (α) and the medium
resistance (Rm ).

1 Based on the experimental data, we conclude this is a problem of constant rate followed
by constant pressure filtration

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 27 / 67
Worked example
Example 2

Example
Filtration tests were carried out with a plate and frame filter press under the following
conditions: solids: ρp = 2710 kg/m3 ; liquid: µ = 0.001 Pa·s; suspension: c = 10 kg/m3 ; filter:
plate-and-frame press, 1 frame, dimensions 430 × 430 × 30 mm. The experimental data are in
the PDF file “Example 2.pdf”. Determine the specific cake resistance (α) and the medium
resistance (Rm ).

1 Based on the experimental data, we conclude this is a problem of constant rate followed
by constant pressure filtration
2 Equation (21) is applicable

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 27 / 67
Worked example
Example 2

Example
Filtration tests were carried out with a plate and frame filter press under the following
conditions: solids: ρp = 2710 kg/m3 ; liquid: µ = 0.001 Pa·s; suspension: c = 10 kg/m3 ; filter:
plate-and-frame press, 1 frame, dimensions 430 × 430 × 30 mm. The experimental data are in
the PDF file “Example 2.pdf”. Determine the specific cake resistance (α) and the medium
resistance (Rm ).

1 Based on the experimental data, we conclude this is a problem of constant rate followed
by constant pressure filtration
2 Equation (21) is applicable
3 The data is transformed to V versus (t − ts ) ÷ (V − Vs )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 27 / 67
Worked example
Example 2

Example
Filtration tests were carried out with a plate and frame filter press under the following
conditions: solids: ρp = 2710 kg/m3 ; liquid: µ = 0.001 Pa·s; suspension: c = 10 kg/m3 ; filter:
plate-and-frame press, 1 frame, dimensions 430 × 430 × 30 mm. The experimental data are in
the PDF file “Example 2.pdf”. Determine the specific cake resistance (α) and the medium
resistance (Rm ).

1 Based on the experimental data, we conclude this is a problem of constant rate followed
by constant pressure filtration
2 Equation (21) is applicable
3 The data is transformed to V versus (t − ts ) ÷ (V − Vs )
4 ts = 3686 s and Vs = 0.3 m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 27 / 67
Worked example (contd.)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 28 / 67
Worked example (contd.)

The slope is Kp ÷ 2 = 26219 s/m6 and the vertical axis intercept is Kp Vs ÷ 2 + B = 9030
s/m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 29 / 67
Worked example (contd.)

The slope is Kp ÷ 2 = 26219 s/m6 and the vertical axis intercept is Kp Vs ÷ 2 + B = 9030
s/m3
Using the definitions of Kp and B (Equations (10) and (11), respectively), simple
rearrangements yield
Kp A2 (−∆P)
α=
µcs
and
A (−∆P)
Rm = [Kp Vs ÷ 2 + B]
µ

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 29 / 67
Worked example (contd.)

The slope is Kp ÷ 2 = 26219 s/m6 and the vertical axis intercept is Kp Vs ÷ 2 + B = 9030
s/m3
Using the definitions of Kp and B (Equations (10) and (11), respectively), simple
rearrangements yield
Kp A2 (−∆P)
α=
µcs
and
A (−∆P)
Rm = [Kp Vs ÷ 2 + B]
µ
The data is transformed to V versus (t − ts ) ÷ (V − Vs )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 29 / 67
Worked example (contd.)

A = 0.43 × 0.43 × 2 m2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 30 / 67
Worked example (contd.)

A = 0.43 × 0.43 × 2 m2
Given: c = 10 kg/m3 suspension; volume occupied by 10 kg solids
= 10 ÷ 2710 = 0.00369 m3 =⇒

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 30 / 67
Worked example (contd.)

A = 0.43 × 0.43 × 2 m2
Given: c = 10 kg/m3 suspension; volume occupied by 10 kg solids
= 10 ÷ 2710 = 0.00369 m3 =⇒
cs = 10 ÷ (1 − 0.00369) = 10.037 kg/m3 filtrate

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 30 / 67
Worked example (contd.)

A = 0.43 × 0.43 × 2 m2
Given: c = 10 kg/m3 suspension; volume occupied by 10 kg solids
= 10 ÷ 2710 = 0.00369 m3 =⇒
cs = 10 ÷ (1 − 0.00369) = 10.037 kg/m3 filtrate
The basic task of substitution yields α = 1.072 × 1011 m/kg and Rm = 6.458 × 1010 m−1

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 30 / 67
Variable-rate + variable-pressure filtration
Incompressible cake

dV
Using Equation (7), and noting that = Q, we can write
dt
 
A (−∆P) A
V = − µRm (22)
µαcs Q

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 31 / 67
Variable-rate + variable-pressure filtration
Incompressible cake

dV
Using Equation (7), and noting that = Q, we can write
dt
 
A (−∆P) A
V = − µRm (22)
µαcs Q
The pump characteristics relate Q and (−∆P)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 31 / 67
Variable-rate + variable-pressure filtration
Incompressible cake

dV
Using Equation (7), and noting that = Q, we can write
dt
 
A (−∆P) A
V = − µRm (22)
µαcs Q
The pump characteristics relate Q and (−∆P)
To filter a volume V of filtrate in time t, we integrate the expression
dV 0
dt 0 =
Q
i.e.
V
dV 0
Z
t= (23)
0 Q
Question: What is the simplest method for computing t as defined in Equation (23)?
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 31 / 67
Worked example
Example 3

Example
Determine the time required for the filtration
of 50 m3 of a slurry in a plate-and-frame filter
press with 25 frames of dimensions
1 × 1 × 0.035 m. The pump characteristic is
shown in the accompanying figure. Other data:
cake resistance (α) = 1.069 × 1011 m/kg;
medium resistance (Rm ) = 6.462 × 1010 1/m;
viscosity (µ) = 0.001 Pa·s; concentration (cs )
= 10.037 kg/m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 32 / 67
Worked example (contd.)

Using Equation (22), we first develop an expression for V = f (Q) viz.

 
1 × 1 × 2 × 25 (−∆P) −3 10
V = 50 − 10 × 6.462 × 10
1.069 × 1011 × 10−3 × 10.037 Q

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 33 / 67
Worked example (contd.)

Using Equation (22), we first develop an expression for V = f (Q) viz.

 
1 × 1 × 2 × 25 (−∆P) −3 10
V = 50 − 10 × 6.462 × 10
1.069 × 1011 × 10−3 × 10.037 Q

From the figure that is provided, V as a function of Q can be calculated; the results are:
Q (m3 /h) −∆P × 10−5 (N/m2 ) V (m3 ) 1/Q (s/m3 )

45 0.2 0.7 80
40 0.75 12.71 90
35 1.15 24.44 103
30 1.4 35.98 120
25 1.6 50.46 144
20 1.75 69.48 180
15 1.8 97.23 240

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 33 / 67
Worked example (contd.)
Now plot 1/Q vs V

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 34 / 67
Worked example (contd.)

We need to integrate the curve from V = 0 to V = 50 m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 35 / 67
Worked example (contd.)

We need to integrate the curve from V = 0 to V = 50 m3

The simplest numerical integration method at our disposal is a Riemann sum (which you
ought to know from elementary calculus!); one such discrete form is the so-called
mid-point rule i.e.
n 
X  −1   −1  
t= Q k
+ Q k−1
÷ 2 × ∆Vk
k=1

where ∆Vk = Vk − Vk−1

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 35 / 67
Worked example (contd.)

We need to integrate the curve from V = 0 to V = 50 m3

The simplest numerical integration method at our disposal is a Riemann sum (which you
ought to know from elementary calculus!); one such discrete form is the so-called
mid-point rule i.e.
n 
X  −1   −1  
t= Q k
+ Q k−1
÷ 2 × ∆Vk
k=1

where ∆Vk = Vk − Vk−1

Applying the formula from V = 0 to V ≈ 50.46 m3 , we obtain t = 5350 s

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 35 / 67
Compressible cakes
Basic equations

Recall that (−∆P) = (−∆Pm ) + (−∆Pc )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 36 / 67
Compressible cakes
Basic equations

Recall that (−∆P) = (−∆Pm ) + (−∆Pc )

From Equations (5) and (6), we note that
αµcs VQ
(−∆Pc ) =
A2
and
µRm Q
(−∆Pm ) =
A
where Q = dV /dt

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 36 / 67
Compressible cakes
Basic equations

Recall that (−∆P) = (−∆Pm ) + (−∆Pc )

From Equations (5) and (6), we note that
αµcs VQ
(−∆Pc ) =
A2
and
µRm Q
(−∆Pm ) =
A
where Q = dV /dt
Also recall: α = α0 × (−∆Pc )s , from which we may calculate αav as
Z −∆Pc
1 1 d (−∆Pc )
=
αav (−∆Pc ) 0 α

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 36 / 67
Compressible cakes
Basic equations

Recall that (−∆P) = (−∆Pm ) + (−∆Pc )

From Equations (5) and (6), we note that
αµcs VQ
(−∆Pc ) =
A2
and
µRm Q
(−∆Pm ) =
A
where Q = dV /dt
Also recall: α = α0 × (−∆Pc )s , from which we may calculate αav as
Z −∆Pc
1 1 d (−∆Pc )
=
αav (−∆Pc ) 0 α
It follows that
αav = (1 − s) α0 (−∆Pc ) (24)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 36 / 67
Compressible cakes
Basic equations (contd.)

Finally, we have
µcs VQ (−∆Pc )1−s
= (25)
A2 (1 − s) α0

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 37 / 67
Compressible cakes
Basic equations (contd.)

Finally, we have
µcs VQ (−∆Pc )1−s
= (25)
A2 (1 − s) α0
Equation (25) is the general equation for compressible cakes from which special cases i.e.
either constant-pressure, constant-rate or variable-pressure-variable rate can be derived

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 37 / 67
Compressible cakes
Constant-pressure filtration

The constant-pressure equations that were derived in the preceding sections are applicable

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 38 / 67
Compressible cakes
Constant-pressure filtration

The constant-pressure equations that were derived in the preceding sections are applicable
Naturally, we need to calculate α based on the operating pressure

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 38 / 67
Compressible cakes
Constant-rate filtration

V = Qt

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 39 / 67
Compressible cakes
Constant-rate filtration

V = Qt
Substitution into Equation (25) followed by rearrangement results in

Q2
(−∆Pc )1−s = α0 (1 − s) µcs t (26)
A2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 39 / 67
Compressible cakes
Variable-pressure-variable-rate filtration

In this case, (−∆Pc ), (−∆Pm ), t, V and Q vary

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 40 / 67
Compressible cakes
Variable-pressure-variable-rate filtration

In this case, (−∆Pc ), (−∆Pm ), t, V and Q vary

Following on from Equation (25), and by once again noting that
(−∆P) = (−∆Pm ) + (−∆Pc ), we arrive at
 " #
A2 (−∆Pc )1−s

V = × (27)
(1 − s) α0 µcs Q

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 40 / 67
Compressible cakes
Variable-pressure-variable-rate filtration

In this case, (−∆Pc ), (−∆Pm ), t, V and Q vary

Following on from Equation (25), and by once again noting that
(−∆P) = (−∆Pm ) + (−∆Pc ), we arrive at
 " #
A2 (−∆Pc )1−s

V = × (27)
(1 − s) α0 µcs Q

The overall pressure drop (−∆P) and Q are related by a pump characteristic curve

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 40 / 67
Compressible cakes
Variable-pressure-variable-rate filtration

In this case, (−∆Pc ), (−∆Pm ), t, V and Q vary

Following on from Equation (25), and by once again noting that
(−∆P) = (−∆Pm ) + (−∆Pc ), we arrive at
 " #
A2 (−∆Pc )1−s

V = × (27)
(1 − s) α0 µcs Q

The overall pressure drop (−∆P) and Q are related by a pump characteristic curve
Rm can be calculated from
µRm Q
(−∆Pm ) =
A

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 40 / 67
Compressible cakes
Variable-pressure-variable-rate filtration (contd.)

Equation (27) is used similarly to Equation (22)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 41 / 67
Compressible cakes
Variable-pressure-variable-rate filtration (contd.)

Equation (27) is used similarly to Equation (22)

Numerical integration then follows via

dV 0
dt 0 =
Q

for determining the total filtration time of V m3 of filtrate

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 41 / 67
Worked example
Example 4

Example
Determine the time necessary to filter 50 m3 of the same slurry in the same filter as in
Example 3, but this time for a compressible filter cake which obeys
α = 6.109 × 109 (−∆Pc )0.24 . The medium resistance is constant and the same as in Example
3. The same pump characteristics as in Example 3 are applicable.

Clearly, s = 0.24. By using Equation (27), an expression relating V and Q is formed i.e.
502 [(−∆P) − (−∆Pm )]0.76
V = ×
0.76 × 6.109 × 109 × 10−3 × 10.037 Q

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 42 / 67
Worked example
Example 4

Example
Determine the time necessary to filter 50 m3 of the same slurry in the same filter as in
Example 3, but this time for a compressible filter cake which obeys
α = 6.109 × 109 (−∆Pc )0.24 . The medium resistance is constant and the same as in Example
3. The same pump characteristics as in Example 3 are applicable.

Clearly, s = 0.24. By using Equation (27), an expression relating V and Q is formed i.e.
502 [(−∆P) − (−∆Pm )]0.76
V = ×
0.76 × 6.109 × 109 × 10−3 × 10.037 Q
Also,
10−3 × 6.462 × 1010 Q
(−∆Pm ) =
50

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 42 / 67
Worked example
Example 4

Example
Determine the time necessary to filter 50 m3 of the same slurry in the same filter as in
Example 3, but this time for a compressible filter cake which obeys
α = 6.109 × 109 (−∆Pc )0.24 . The medium resistance is constant and the same as in Example
3. The same pump characteristics as in Example 3 are applicable.

Clearly, s = 0.24. By using Equation (27), an expression relating V and Q is formed i.e.
502 [(−∆P) − (−∆Pm )]0.76
V = ×
0.76 × 6.109 × 109 × 10−3 × 10.037 Q
Also,
10−3 × 6.462 × 1010 Q
(−∆Pm ) =
50
Now we use the given pump characteristics and form a table that relates the values of V
to the corresponding 1 ÷ Q values (see next slide)
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 42 / 67
Worked example (contd.)

The table reads as follows:

Q (m3 /h) −∆P × 10−5 (N/m2 ) −∆Pm × 10−5 (N/m2 ) V (m3 ) 1/Q (s/m3 )

45 0.2 0.1616 2.3 80

40 0.75 0.1436 20.8 90
35 1.15 0.1257 35.4 103
30 1.4 0.1077 49.2 120
25 1.6 0.0898 66.5 144

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 43 / 67
Worked example (contd.)

The table reads as follows:

Q (m3 /h) −∆P × 10−5 (N/m2 ) −∆Pm × 10−5 (N/m2 ) V (m3 ) 1/Q (s/m3 )

45 0.2 0.1616 2.3 80

40 0.75 0.1436 20.8 90
35 1.15 0.1257 35.4 103
30 1.4 0.1077 49.2 120
25 1.6 0.0898 66.5 144
Now generate a plot of V (m3 ) versus 1/Q (s/m3 )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 43 / 67
Worked example (contd.)

All that remains for us to do is to (numerically) integrate the curve up to V = 50 m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 44 / 67
Worked example (contd.)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 45 / 67
Worked example (contd.)

The time t required for the filtration based on Equation (23) i.e.
Z 50 m3
1
t= dV 0
0 Q
is t ≈ 4750 s

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 46 / 67
Washing of filter cakes

Washing removes the mother liquor retained by the cake after filtration

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 47 / 67
Washing of filter cakes

Washing removes the mother liquor retained by the cake after filtration
The liquor is either a valuable component, or a contaminant

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 47 / 67
Washing of filter cakes

Washing removes the mother liquor retained by the cake after filtration
The liquor is either a valuable component, or a contaminant
The wash liquid flows through the filter cake, displacing the liquor, and a wash filtrate
that is a mixture of the liquor and wash liquid leaves the filter cake

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 47 / 67
Washing of filter cakes

Washing removes the mother liquor retained by the cake after filtration
The liquor is either a valuable component, or a contaminant
The wash liquid flows through the filter cake, displacing the liquor, and a wash filtrate
that is a mixture of the liquor and wash liquid leaves the filter cake
A washing curve describes the washing process

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 47 / 67
Washing of filter cakes

Washing removes the mother liquor retained by the cake after filtration
The liquor is either a valuable component, or a contaminant
The wash liquid flows through the filter cake, displacing the liquor, and a wash filtrate
that is a mixture of the liquor and wash liquid leaves the filter cake
A washing curve describes the washing process
The aforesaid curve plots
solute concentration in washings from filter cake
solute concentration in filter cake liquor before washing starts
against the “wash ratio” that is defined as
volume of wash liquid used
volume of liquor in filter cake at the start of washing

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 47 / 67
Washing of filter cakes (contd.)

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 48 / 67
Washing of filter cakes (contd.)

The purple curve represents ideal displacement washing, wherein just one void volume of
wash liquid is needed to recover 100% of the liquor in filter cake; this is akin to plug flow

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 49 / 67
Washing of filter cakes (contd.)

The purple curve represents ideal displacement washing, wherein just one void volume of
wash liquid is needed to recover 100% of the liquor in filter cake; this is akin to plug flow
Naturally, the situation described above is never observed in practice

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 49 / 67
Washing of filter cakes (contd.)

The purple curve represents ideal displacement washing, wherein just one void volume of
wash liquid is needed to recover 100% of the liquor in filter cake; this is akin to plug flow
Naturally, the situation described above is never observed in practice
The red curve is a typical washing curve, in which 30-86% of the residual liquor is
removed in a once-through wash

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 49 / 67
Washing of filter cakes (contd.)

The purple curve represents ideal displacement washing, wherein just one void volume of
wash liquid is needed to recover 100% of the liquor in filter cake; this is akin to plug flow
Naturally, the situation described above is never observed in practice
The red curve is a typical washing curve, in which 30-86% of the residual liquor is
removed in a once-through wash
Typical washing curves can be divided into two regions where different modes of solute
removal from the filter cake’s pores are observed

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 49 / 67
Washing of filter cakes (contd.)

The purple curve represents ideal displacement washing, wherein just one void volume of
wash liquid is needed to recover 100% of the liquor in filter cake; this is akin to plug flow
Naturally, the situation described above is never observed in practice
The red curve is a typical washing curve, in which 30-86% of the residual liquor is
removed in a once-through wash
Typical washing curves can be divided into two regions where different modes of solute
removal from the filter cake’s pores are observed
In the inital stage, direct displacement of the residual liquor by the wash liquor is observed

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 49 / 67
Washing of filter cakes (contd.)

The purple curve represents ideal displacement washing, wherein just one void volume of
wash liquid is needed to recover 100% of the liquor in filter cake; this is akin to plug flow
Naturally, the situation described above is never observed in practice
The red curve is a typical washing curve, in which 30-86% of the residual liquor is
removed in a once-through wash
Typical washing curves can be divided into two regions where different modes of solute
removal from the filter cake’s pores are observed
In the inital stage, direct displacement of the residual liquor by the wash liquor is observed
In the final stage, there are mass transfer mechanisms at play in the filter cake’s pores

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 49 / 67
Washing of filter cakes (contd.)
To determine the amount of wash liquid needed for a particular solute concentration
reduction, the wash ratio can be read directly from the washing curve

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 50 / 67
Washing of filter cakes (contd.)
To determine the amount of wash liquid needed for a particular solute concentration
reduction, the wash ratio can be read directly from the washing curve
To calculate washing rates, we assume that the washing conditions are equivalent to
those at the end of filtration

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 50 / 67
Washing of filter cakes (contd.)
To determine the amount of wash liquid needed for a particular solute concentration
reduction, the wash ratio can be read directly from the washing curve
To calculate washing rates, we assume that the washing conditions are equivalent to
those at the end of filtration
We also assume the cake structure remains unchanged when the slurry in the cake is
replaced by the wash liquid

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 50 / 67
Washing of filter cakes (contd.)
To determine the amount of wash liquid needed for a particular solute concentration
reduction, the wash ratio can be read directly from the washing curve
To calculate washing rates, we assume that the washing conditions are equivalent to
those at the end of filtration
We also assume the cake structure remains unchanged when the slurry in the cake is
replaced by the wash liquid
The washing rate is determined by the final filtering rate

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 50 / 67
Washing of filter cakes (contd.)
To determine the amount of wash liquid needed for a particular solute concentration
reduction, the wash ratio can be read directly from the washing curve
To calculate washing rates, we assume that the washing conditions are equivalent to
those at the end of filtration
We also assume the cake structure remains unchanged when the slurry in the cake is
replaced by the wash liquid
The washing rate is determined by the final filtering rate
For constant-pressure filtration, and using the same pressure during washing as was used
during filtering, we invert Equation (9), which results in
 
dV 1
= (28)
dt f Kp Vf + B
where the derivative on the left-hand-side is the rate of washing and Vf is the total
amount of filtrate collected up to the end of the filtration
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 50 / 67
Washing of filter cakes (contd.)

For plate-and-frame filter presses, the predicted washing rate is a quarter of the final
filtration rate i.e.  
dV 1 1
= (29)
dt f 4 Kp Vf + B

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 51 / 67
Washing of filter cakes (contd.)

For plate-and-frame filter presses, the predicted washing rate is a quarter of the final
filtration rate i.e.  
dV 1 1
= (29)
dt f 4 Kp Vf + B
The total filter cycle time is the sum of the filtration time, washing time, cleaning time

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 51 / 67
Washing of filter cakes (contd.)
Measuring washing curves

Experimental curves are used to determine the washing time and the amount of washing
liquid needed to remove a predetermined amount of solute from the filter cake

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 52 / 67
Washing of filter cakes (contd.)
Measuring washing curves

Experimental curves are used to determine the washing time and the amount of washing
liquid needed to remove a predetermined amount of solute from the filter cake
There are several ways in which the experimental results can be plotted

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 52 / 67
Washing of filter cakes (contd.)
Measuring washing curves

Experimental curves are used to determine the washing time and the amount of washing
liquid needed to remove a predetermined amount of solute from the filter cake
There are several ways in which the experimental results can be plotted
For brevity we shall very briefly study one type of plot viz. the washing curve defined by a
plot of solute concentration in the washing from the cake versus the washing time, both
in a dimensionless form

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 52 / 67
Washing of filter cakes (contd.)
Measuring washing curves

Experimental curves are used to determine the washing time and the amount of washing
liquid needed to remove a predetermined amount of solute from the filter cake
There are several ways in which the experimental results can be plotted
For brevity we shall very briefly study one type of plot viz. the washing curve defined by a
plot of solute concentration in the washing from the cake versus the washing time, both
in a dimensionless form
Symbolically, this is c/c0 vs W

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 52 / 67
Washing of filter cakes (contd.)
Measuring washing curves

Experimental curves are used to determine the washing time and the amount of washing
liquid needed to remove a predetermined amount of solute from the filter cake
There are several ways in which the experimental results can be plotted
For brevity we shall very briefly study one type of plot viz. the washing curve defined by a
plot of solute concentration in the washing from the cake versus the washing time, both
in a dimensionless form
Symbolically, this is c/c0 vs W
A precise expression for W (the wash ratio; cf. Slide 47) is
Qt
W =
εav ALS
(30)
volume of wash liquid used
=
void volume of the cake × cake saturation at start of washing

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 52 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake
A is the surface area of the cake normal to the flow direction

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake
A is the surface area of the cake normal to the flow direction
L is the cake thickness

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake
A is the surface area of the cake normal to the flow direction
L is the cake thickness
S is the cake saturation at the start of washing

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake
A is the surface area of the cake normal to the flow direction
L is the cake thickness
S is the cake saturation at the start of washing
Another representation of the experimental results is to plot the fraction of solute
removed from the filter cake, denoted F , against W

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake
A is the surface area of the cake normal to the flow direction
L is the cake thickness
S is the cake saturation at the start of washing
Another representation of the experimental results is to plot the fraction of solute
removed from the filter cake, denoted F , against W
Let cw be the concentration of solute in the wash liquid feed; we then define F as
Z W
c − cw
F = dW 0 (31)
0 c 0 − c w

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

In Equation (30):
Q is the volumetric flow rate of wash liquid that passes through the cake
t is the time of washing
εav is the average porosity of the cake
A is the surface area of the cake normal to the flow direction
L is the cake thickness
S is the cake saturation at the start of washing
Another representation of the experimental results is to plot the fraction of solute
removed from the filter cake, denoted F , against W
Let cw be the concentration of solute in the wash liquid feed; we then define F as
Z W
c − cw
F = dW 0 (31)
0 c 0 − c w
Equivalently we may define R as the fraction of solute remaining in the filter cake as
R =1−F (32)
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 53 / 67
Washing of filter cakes (contd.)
Measuring washing curves

F ∼ 1 as W −→ ∞

Application of filter cake washing theory

Experimental data from a washing test using solute-free washing liquid/liquor are presented
below.
c/c0 1.0 1.0 0.9 0.8 0.4 0.1 0.06 0.04 0.01
W 0 0.4 0.68 0.75 1.0 1.25 1.375 1.5 2.7
Generate the fractional recovery of solute curve for this dataset.

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 54 / 67
Washing of filter cakes (contd.)
Measuring washing curves

F ∼ 1 as W −→ ∞
Equivalently, R ∼ 0 as W −→ ∞

Application of filter cake washing theory

Experimental data from a washing test using solute-free washing liquid/liquor are presented
below.
c/c0 1.0 1.0 0.9 0.8 0.4 0.1 0.06 0.04 0.01
W 0 0.4 0.68 0.75 1.0 1.25 1.375 1.5 2.7
Generate the fractional recovery of solute curve for this dataset.

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 54 / 67
Washing of filter cakes (contd.)
Measuring washing curves

The washing curve (which is simply a plot of the given raw data!) is:

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 55 / 67
Washing of filter cakes (contd.)
Measuring washing curves

To generate the F -curve, calculate the area of each rectangle on the washing curve, and
form another column that is the cumulative sum of the rectangles’ areas

0
R Wk+1 R W0
W Wk0
(c/c0 ) dW 0 F = 0 (c/c0 ) dW 0
0 0 0
0.4 0.4 0.4
0.68 0.266 0.666
0.75 0.0595 0.725
1.0 0.15 0.875
1.25 0.0625 0.938
1.375 0.01 0.948
1.5 0.0062 0.954
2.7 0.03 0.984
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 56 / 67
Washing of filter cakes (contd.)
Measuring washing curves

To generate the F -curve, calculate the area of each rectangle on the washing curve, and
form another column that is the cumulative sum of the rectangles’ areas
Our table of treated data then reads:
0
R Wk+1 R W0
W W 0 (c/c0 ) dW 0 F = 0 (c/c0 ) dW 0
k
0 0 0
0.4 0.4 0.4
0.68 0.266 0.666
0.75 0.0595 0.725
1.0 0.15 0.875
1.25 0.0625 0.938
1.375 0.01 0.948
1.5 0.0062 0.954
2.7 0.03 0.984
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 56 / 67
Washing of filter cakes (contd.)
Measuring washing curves

Lastly, plot W against F :

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 57 / 67
Worked example
Example 5

Example
The same slurry used in Example 1 is now filtered in a plate-and-frame filter press having 20
frames and 0.873 m2 area per frame. The same pressure will be used in constant-pressure
filtration. Assuming the same filter cake properties and filter cloth, calculate the time to
recover 3.37 m3 of filtrate. Recall from Example 1 that A = 0.0439 m2 .

We calculated Kp (6 × 106 s/m6 ) and B (6400 s/m3 ) in Example 1

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 58 / 67
Worked example
Example 5

Example
The same slurry used in Example 1 is now filtered in a plate-and-frame filter press having 20
frames and 0.873 m2 area per frame. The same pressure will be used in constant-pressure
filtration. Assuming the same filter cake properties and filter cloth, calculate the time to
recover 3.37 m3 of filtrate. Recall from Example 1 that A = 0.0439 m2 .

We calculated Kp (6 × 106 s/m6 ) and B (6400 s/m3 ) in Example 1

α and Rm remain unchanged

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 58 / 67
Worked example
Example 5

Example
The same slurry used in Example 1 is now filtered in a plate-and-frame filter press having 20
frames and 0.873 m2 area per frame. The same pressure will be used in constant-pressure
filtration. Assuming the same filter cake properties and filter cloth, calculate the time to
recover 3.37 m3 of filtrate. Recall from Example 1 that A = 0.0439 m2 .

We calculated Kp (6 × 106 s/m6 ) and B (6400 s/m3 ) in Example 1

α and Rm remain unchanged
However, Kp and B need to be corrected; why?

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 58 / 67
Worked example
Example 5

Example
The same slurry used in Example 1 is now filtered in a plate-and-frame filter press having 20
frames and 0.873 m2 area per frame. The same pressure will be used in constant-pressure
filtration. Assuming the same filter cake properties and filter cloth, calculate the time to
recover 3.37 m3 of filtrate. Recall from Example 1 that A = 0.0439 m2 .

We calculated Kp (6 × 106 s/m6 ) and B (6400 s/m3 ) in Example 1

α and Rm remain unchanged
However, Kp and B need to be corrected; why?
From Equation (10),
1
Kp ∝
A2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport 2024 58 / 67

Worked example

The new area A = 20 × 0.873 = 17.46 m2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 59 / 67
Worked example

The new area A = 20 × 0.873 = 17.46 m2

∴ Kp = 6 × 106 (0.0439 ÷ 17.46)2 = 37.93 s/m6

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 59 / 67
Worked example

The new area A = 20 × 0.873 = 17.46 m2

∴ Kp = 6 × 106 (0.0439 ÷ 17.46)2 = 37.93 s/m6
Likewise from Equation (11),
1
B∝
A

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 59 / 67
Worked example

The new area A = 20 × 0.873 = 17.46 m2

∴ Kp = 6 × 106 (0.0439 ÷ 17.46)2 = 37.93 s/m6
Likewise from Equation (11),
1
B∝
A
∴ B = 6400 (0.0439 ÷ 17.46) = 16.1 s/m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 59 / 67
Worked example

The new area A = 20 × 0.873 = 17.46 m2

∴ Kp = 6 × 106 (0.0439 ÷ 17.46)2 = 37.93 s/m6
Likewise from Equation (11),
1
B∝
A
∴ B = 6400 (0.0439 ÷ 17.46) = 16.1 s/m3
Multiplying both sides of Equation (13) and substitution of the preceding values (and
noting that V = 3.37 m3 ) yields

Kp 2
t= V + BV = 269.7 s
2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 59 / 67
Worked example
Example 6

Example
Following on from Example 5, the filter cake is to be washed by through washing in a
plate-and-frame filter press using a volume of wash water equal to 10% of the filtrate volume.
Calculate the time of washing and the total filter cycle time (in minutes) if cleaning the filter
takes 20 minutes.

Equation (29) is applicable

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 60 / 67
Worked example
Example 6

Example
Following on from Example 5, the filter cake is to be washed by through washing in a
plate-and-frame filter press using a volume of wash water equal to 10% of the filtrate volume.
Calculate the time of washing and the total filter cycle time (in minutes) if cleaning the filter
takes 20 minutes.

Equation (29) is applicable

From Example 5, we know that Kp = 37.93 s/m6 , B = 16.1 s/m3 and Vf = 3.37 m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 60 / 67
Worked example
Example 6

Example
Following on from Example 5, the filter cake is to be washed by through washing in a
plate-and-frame filter press using a volume of wash water equal to 10% of the filtrate volume.
Calculate the time of washing and the total filter cycle time (in minutes) if cleaning the filter
takes 20 minutes.

Equation (29) is applicable

From Example 5, we know that Kp = 37.93 s/m6 , B = 16.1 s/m3 and Vf = 3.37 m3
The washing rate is  
dV
= 1.737 × 10−3 m3 /s
dt f

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 60 / 67
Worked example
Example 6

Example
Following on from Example 5, the filter cake is to be washed by through washing in a
plate-and-frame filter press using a volume of wash water equal to 10% of the filtrate volume.
Calculate the time of washing and the total filter cycle time (in minutes) if cleaning the filter
takes 20 minutes.

Equation (29) is applicable

From Example 5, we know that Kp = 37.93 s/m6 , B = 16.1 s/m3 and Vf = 3.37 m3
The washing rate is  
dV
= 1.737 × 10−3 m3 /s
dt f

The volume of wash water is 0.1 × 3.37 = 0.337 m3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 60 / 67
Worked example (contd.)

The washing time is thus

0.337
t= = 194 s
1.737 × 10−3

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 61 / 67
Worked example (contd.)

The washing time is thus

0.337
t= = 194 s
1.737 × 10−3
Finally, the total cycle time is
194.0 269.7
+ + 20 = 27.7 minutes
60 60

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 61 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering
Cake washing

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering
Cake washing
(Perhaps) Secondary partial dewatering

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering
Cake washing
(Perhaps) Secondary partial dewatering
Cake discharge

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering
Cake washing
(Perhaps) Secondary partial dewatering
Cake discharge
This is a cyclic process, and lends itself to
continuous operation

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering
Cake washing
(Perhaps) Secondary partial dewatering
Cake discharge
This is a cyclic process, and lends itself to
continuous operation
Denote the total cycle time as tc

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization
Rotary vacuum drum (RVD) filter

One revolution of an RVD filter consists

of
Cake formation
Cake filtration
Primary partial dewatering
Cake washing
(Perhaps) Secondary partial dewatering
Cake discharge
This is a cyclic process, and lends itself to
continuous operation
Denote the total cycle time as tc
Three basic operations can be identified

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 62 / 67
Filtration cycle and its optimization (contd.)

These operations are

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)

These operations are

Filtration (tf )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)

These operations are

Filtration (tf )
Washing (tw )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)

These operations are

Filtration (tf )
Washing (tw )
Cake discharge + dewatering (td )

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)

These operations are

Filtration (tf )
Washing (tw )
Cake discharge + dewatering (td )
Hence, tc = tf + tw + td

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)

These operations are

Filtration (tf )
Washing (tw )
Cake discharge + dewatering (td )
Hence, tc = tf + tw + td
For a batch filter, td is not necessarily constant

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)

These operations are

Filtration (tf )
Washing (tw )
Cake discharge + dewatering (td )
Hence, tc = tf + tw + td
For a batch filter, td is not necessarily constant
We now proceed with developing equations for the case of constant-pressure filtration

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 63 / 67
Filtration cycle and its optimization (contd.)
RVD

We recall Equation (13) viz.

t Kp
= V +B
V 2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 64 / 67
Filtration cycle and its optimization (contd.)
RVD

We recall Equation (13) viz.

t Kp
= V +B
V 2
The filtering time that results in V m3 of filtrate being collected is
Kp 2
t= V + BV
2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 64 / 67
Filtration cycle and its optimization (contd.)
RVD

We recall Equation (13) viz.

t Kp
= V +B
V 2
The filtering time that results in V m3 of filtrate being collected is
Kp 2
t= V + BV
2
Hence, by inspection, the amount of filtrate collected is
p
−B + B 2 + 2Kp t
V = (33)
Kp

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 64 / 67
Filtration cycle and its optimization (contd.)
RVD

Substituting for Kp and B from Equations (10) and (11), and dividing both sides by tA,
results in q
V [2 (−∆P) cs α ÷ µt] + (Rm /t)2 − Rm /t
= (34)
tA cs α
where A is the submerged area of the filter and V ÷ t is the rate of filtrate collection

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 65 / 67
Filtration cycle and its optimization (contd.)
RVD

Substituting for Kp and B from Equations (10) and (11), and dividing both sides by tA,
results in q
V [2 (−∆P) cs α ÷ µt] + (Rm /t)2 − Rm /t
= (34)
tA cs α
where A is the submerged area of the filter and V ÷ t is the rate of filtrate collection
We note that in a rotary drum filter, the actual filter time tf is related to the total cycle
time tc by tf = ftc

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 65 / 67
Filtration cycle and its optimization (contd.)
RVD

Substituting for Kp and B from Equations (10) and (11), and dividing both sides by tA,
results in q
V [2 (−∆P) cs α ÷ µt] + (Rm /t)2 − Rm /t
= (34)
tA cs α
where A is the submerged area of the filter and V ÷ t is the rate of filtrate collection
We note that in a rotary drum filter, the actual filter time tf is related to the total cycle
time tc by tf = ftc
f is the fraction submergence of the drum in the slurry; hence,
q
V [2 (−∆P) cs αf ÷ µtc ] + (Rm /tc )2 − Rm /tc
= (35)
tc A cs α

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 65 / 67
Filtration cycle and its optimization (contd.)
Case study: plate-and-frame filter press

We already know that for constant-pressure filtration, the filtration time tf is

Kp 2
t= V + BV
2

1
This is termed the “down time”
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 66 / 67
Filtration cycle and its optimization (contd.)
Case study: plate-and-frame filter press

We already know that for constant-pressure filtration, the filtration time tf is

Kp 2
t= V + BV
2
Let t 0 be the time required for dismantling and re-assemblying the filter press1 . The total
cycle time in which V m3 of filtrate is collected is tf + t 0

1
This is termed the “down time”
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 66 / 67
Filtration cycle and its optimization (contd.)
Case study: plate-and-frame filter press

We already know that for constant-pressure filtration, the filtration time tf is

Kp 2
t= V + BV
2
Let t 0 be the time required for dismantling and re-assemblying the filter press1 . The total
cycle time in which V m3 of filtrate is collected is tf + t 0
The overall filtration rate is simply
V
W =
Kp 2
V + BV + t 0
2

1
This is termed the “down time”
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 66 / 67
Filtration cycle and its optimization (contd.)
Case study: plate-and-frame filter press

We already know that for constant-pressure filtration, the filtration time tf is

Kp 2
t= V + BV
2
Let t 0 be the time required for dismantling and re-assemblying the filter press1 . The total
cycle time in which V m3 of filtrate is collected is tf + t 0
The overall filtration rate is simply
V
W =
Kp 2
V + BV + t 0
2
We want to maximize W ; hence, set dW /dV = 0
1
This is termed the “down time”
Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 66 / 67
Filtration cycle and its optimization (contd.)
Case study: plate-and-frame filter press

Simple differentiation leads to

 
Kp
V 2 + BV + t 0 − V (Kp V + B) = 0
2

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 67 / 67
Filtration cycle and its optimization (contd.)
Case study: plate-and-frame filter press

Simple differentiation leads to

 
Kp
V 2 + BV + t 0 − V (Kp V + B) = 0
2

Solving for Vopt. results in s

2t 0
Vopt. =
Kp

Dr S. Moodley (UKZN HC) ENCH3FS Fluid and Solids Transport Semester 1, 2024 67 / 67