See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/265596677

Developments in Mud Filtration Technology in the Sugarcane Industry

Chapter · October 2014

CITATIONS READS
5 9,374

3 authors:

Thomas J. Rainey Omkar P Thaval

Queensland University of Technology Queensland University of Technology
88 PUBLICATIONS   1,063 CITATIONS    6 PUBLICATIONS   12 CITATIONS   

SEE PROFILE SEE PROFILE

Darryn Rackemann
Queensland University of Technology
51 PUBLICATIONS   799 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Improving the dewatering of bagasse paper pulp View project

Environmental Impacts of Shipping Emissions View project

All content following this page was uploaded by Thomas J. Rainey on 28 May 2015.

The user has requested enhancement of the downloaded file.

 Please cite as

Rainey, T.J., Thaval., O.P., Rackemann, D.W. 2014. “Devlopments in Mud

Filtration Technology in the Sugarcane Industry” In Sugarcane: Production,
Consumption and Agricultural Management Systems. Edited by Eleanor
Webb, 263-292, New York.

Chapter

DEVELOPMENTS IN MUD FILTRATION

TECHNOLOGY IN THE SUGARCANE
INDUSTRY

T. J. Rainey, O. P. Thaval and D. W. Rackemann

Queensland University of Technology, Brisbane, Australia

ABSTRACT
Dirt collected with sugarcane is processed and separated from the
juice in the sugar factory by filtration equipment for return to the cane
fields. New technologies over the past decade have enabled performance
improvements to be obtained for this key unit operation. Filter mud
product still contains a reasonable amount of sugar and the transportation
of high moisture mud product has considerable cost. Australia’s
traditional approach has been to use Rotary Vacuum Filters for
processing and separating mud and other impurities from juice, but in
recent years there has been interest in reducing sugar losses and
transportation costs through utilisation of new technologies such as


Corresponding author: T. J. Rainey. Queensland University of Technology, 2 George St,
Brisbane 4000, Australia. E-mail: [email protected].
2 T. J. Rainey, O. P. Thaval and D. W. Rackemann

Horizontal Bed Filters, Vacuum Belt Press Filters, Membrane Press

Filters and Centrifuges. Increasingly, these alternative equipment are
being installed in new factories. This chapter describes the general
principles of mud filtration theory and mud conditioning followed by a
detailed description and review of the various filtration technologies and
analysis of the relative merits associated with the equipment.

INTRODUCTION
The removal of insoluble non-sugar impurities from sugarcane juice is an
essential part of raw sugar production and this is commonly achieved through
juice clarification. Lime saccharate and polymeric flocculants remove
suspended matter which precipitate and leave the clarifier in the underflow and
the clarified juice is sent to the evaporators for concentration [1].
The mud underflow from the clarifier typically contains 5–10 %
precipitate in suspension and 10–15 % sugar in the juice. Dirt collected with
sugarcane is separated from the underflow juice by filtration equipment for
return to the cane fields. Historically, Australian factories predominantly use
traditional Rotary Vacuum Filters (RVFs) to separate >85 % of the mud solids
in juice to produce a filter mud product of 75–80 % moisture [2], however the
technology landscape is changing.
A schematic of mud processing typically undertaken in Australian sugar
factories with RVFs and their associated equipment is shown in Figure 1.

LC Level controller
CLARIFIER VC Pick-up vacuum controller
Bagacillo

Saccharate BAGACILLO CYCLONE

Flocculant
Primary mud Injection
Wash water
MUD water
LC MIXER
FILTER CONDENSER
Filter
Cake
feed

LC

VC
HIGH VACUUM
LOW VACUUM RECEIVER
RECEIVER
Filtrate

LC LC
filtrate dilution

FILTRATE PUMP TORRI

WELL VACUUM
PUMP
 Developments in Mud Filtration Technology in the Sugarcane Industry 3

Figure 1. Schematic of sugarcane mud processing with RVFs.

Mud filtration is the last opportunity for the remaining sugar in mill mud
to be recovered. The goals of mud filtration are to: (i) recover sugar and return
it as part of the filtrate to the process; and (ii) maximise mud solids retention,
minimising the amount of mud recycled back into the process. These two
goals are achieved by efficient washing with water.
More recently there has been impetus for improving mud distribution back
to the cane farms in terms of the transport costs and also the wide spread
application of this nutrient-rich product.
As moisture is the major component of filter mud product, processing
options and technologies that allow more effective dewatering (while
maintaining acceptable levels of sugar recovery) are of most interest to sugar
factories.
Although factory mud has nutrient value, the high transport costs to return
it to the cane fields impacts on the distances it can be economically transported
which also can depend on farm application rates. When taking into account
nutrient value, a recent case study found application rates of 100–150 t/ha was
more economic than fertiliser application for distances of up to 20 km from the
factory [3]. This was based on fertiliser-replacement value although reduced
application rates were required to economically extend the treated area.
Transportation costs which limit the widespread distribution of factory mud
has led to adverse secondary issues such as the accumulation of heavy metals
on farms in closer proximity to the factory. This can have a negative
environmental impact due to water runoff following high rainfall events.
Increasing the mud solids density in the filter mud product (i.e., reducing
moisture content) can extend the economic distribution and application of
filter mud over a wider area to reduce heavy metal accumulation
(environmental benefit) and fertiliser requirements (economic benefit).
New technologies over the past decade have enabled performance
improvements to be obtained for this key unit operation. Such technologies
have the potential to improve sugar recovery and/or transport distance whilst
improving wash water efficiency. Australia’s traditional approach has been to
use Rotary Vacuum Filters which are in almost universal use in Australian
factories. Internationally, there has been considerable interest in Vacuum Belt
Filters (VBFs), Membrane Press Filters (MPFs) and Centrifuges.
Vacuum Belt Filters in particular, are increasingly being installed in new
factories, especially in Brazil. For most of these technologies, the mud product
4 T. J. Rainey, O. P. Thaval and D. W. Rackemann

can be described as cake, except for Centrifuges which produces granulated

material.

FILTRATION THEORY AND ITS

APPLICATION TO SUGAR FACTORY MUD
Filter feed conditioning has a significant effect on the filtration properties,
such as cake permeability, cake resistance and cake porosity [4]. Commonly,
these terms are used interchangeably but incorrectly; they are more closely
related for unit operations which are close to steady-state such as in the
operation of RVFs. Permeability is related to cake resistance but both are in
fact distinct properties from porosity. This distinction becomes important
when comparing performance parameters of different technologies.
Permeability is a measure of the ease in which a fluid can flow through a
porous medium and is quantified by Darcy’s Law. The filtration and washing
of mud by RVFs generally obeys Darcy’s Law as limited compression of the
mud occurs. Darcy’s Law relates to steady-state laminar flow through a
homogeneous porous media. A common simplified description of Darcy’s
Law is:

∆
(1)
∆

where Q is the volumetric flow rate (cm3/s) through a bed of porous material
with cross-sectional area A (cm2), ΔP is the frictional pressure drop (mPa)
across the length ΔL (mm), μ is the filtrate viscosity and K is the permeability
constant (cm2). However, the sugar industry generally uses a more functional
form of Darcy’s Law for quantifying RVF performance. The functional form
includes specific cake resistance which is inversely related to permeability [4]
as follows:

∆
(2)

where v is filtrate volume, t is time, α is the average specific cake resistance, C

is the mass of dry solids in the cake and Rm is the hydraulic resistance of the
screen.
With alternative separation technologies (Press Filters and Centrifuges),
the porous media compresses and so permeability and hence cake resistance
 Developments in Mud Filtration Technology in the Sugarcane Industry 5

changes with porosity according to the Kozeny-Carman relation for capillary

flow. The unsteady-state filtration still obeys the principles of capillary theory,
namely Kozeny-Carman’s equation:
(3)

where ε is the porosity (fraction of void volume), Sv is the specific surface area
and k is most commonly assumed to be a constant. The underlying theory for
the Kozeny Carman relation relates to the assumption of the flow of water
through capillaries. The Kozeny-Carman relation can be derived from
Poiseulle’s Law for flow through long cylindrical pipes [5].
The traditional approach to quantifying cake resistance vis a vis Darcy’s
Law increases theoretical complexity for new mud filtration technologies that
are not operating under steady-state conditions. Dynamic filtration of porous
media involves interaction between the compression and permeability of the
media; as a medium compresses, the porosity, ε, reduces and the permeability
decreases. The variability of permeability (and cake resistance) with porosity
is shown in Figure 2 which assumes other factors are constant.
Steady-state compressibility behaviour can be determined by:

Ps=MφN (4)

where Ps is the pressure on the solid phase, φ is the solid fraction (i.e., 1-ε), M
and N are constants.
6 T. J. Rainey, O. P. Thaval and D. W. Rackemann

Figure 2. Relation between permeability, K, and porosity, ε, for capillaric porous

media with a constant Kozeny factor (k) and specific surface area (Sv).

There is a wide variety of dynamic compressibility models in both soil

science literature and bagasse. A general dynamic filtration model proposed by
Banks [6] was applied to bagasse by Owen and co-workers [7] and Kent and
McKenzie [8]. However, for mud it is possible to quantify and characterise the
steady-state permeability and compressibility behaviours separately and
combine them into a dynamic model. The theoretical framework for
generalised constant rate and constant pressure dynamic filtrations for similar
porous media is provided by Landman and co-workers [9, 10].
Mud filter cake contains a significant amount of sugar that is lost from the
factory. The primary method of reducing the amount of lost sugar is through
the application of sufficient wash water during filtration. However, the
effectiveness of the water application depends on the porosity of the cake and
the means of application. Provided the wash water is added effectively,
increasing the wash water rate will decrease the amount of sugar lost. Most
mud filtration technologies rely on displacement washing whereby wash water
pushes the juice through capillaries in the mud cake with limited mixing (the
basis for the Kozeny Carman relation) rather than diffusing and diluting the
juice in the cake. Compressing the cake too much reduces its porosity
(resulting in high cake resistance) such that added wash water does not easily
penetrate the cake to displace any sugar, while a very porous cake will
encourage channelling of the water and a poor washing efficiency will result.
Thus, in both cases, sugar recovery would be limited. For RVF operation, if
the cake resistance is high, excess water will run off the surface of the cake
into the mud boot where less effective dilution washing takes place.

FEED CONDITIONING
Regardless of the technology used, feed conditioning is essential for good
separation performance. Mud filter cake resistance is most affected by the
quality and quantity of bagacillo. The term bagacillo is given to the very fine
fibre particles of bagasse which is the fibrous residue of sugarcane after the
juice has been squeezed out. Fine bagacillo improves mud solids retention and
its addition improves porosity. In adequate quantities it can increase filtrate
rate up to 300 %. Low quality bagacillo can incorporate large pieces of
sugarcane rind but these provide no benefit to sugar recovery or porosity (i.e.,
 Developments in Mud Filtration Technology in the Sugarcane Industry 7

washing efficiency). Following bagacillo addition, lime saccharate has a major

beneficial effect on increasing cake porosity, improving sugar recovery and
providing good mud solids retention.
Good RVF performance is achieved when adequate saccharate is added to
maintain filtrate at a pH of 8.5. However, adding lime is essentially adding
more scale-forming components (i.e., calcium) to the process and can also
have the effect of blinding filter screens. This can increase maintenance, and
reduce throughput and recovery of the filter station.
Finally, flocculant also has a significant effect on the dewatering
capability of filter feed. If used effectively, it reduces cake resistance and
improves sugar recovery and mud solids retention. Around 5–10 ppm is
required in filter feed and the type of flocculant required varies from factory to
factory, depending on the unique soil characteristics of the region. Too much
or too little flocculant can be detrimental to the goals of mud solids retention
and sugar recovery. To avoid disrupting he flocculated mud particles, the feed
to the filter is optimally gravity fed rather than pumped.
Mud density can also impact on filter performance and cake thickness.
Lighter mud (i.e., 3–5 % mud solids) which is achieved through dilution with
either water or recycling a portion of the filtrate produces a thinner filter feed
of lower consistency. This enables a more even and uniform distribution of
feed and cake to improve filter performance and reduce feed pipe blockages
[11]. The current knowledge on sugarcane mud filtration is based on studies of
the operation of RVFs under factory conditions [12-16] or laboratory studies
conducted using a batch filter or similar apparatus [4, 17, 18].
There are relatively few recent studies on sugarcane mud filtration
reflecting the maturity of the technology. Recent research on sugarcane mud
filtration has focused on alternative filtration technologies such as Horizontal
Bed Filters [19] and Centrifuges [2]. Current attention has focussed on VBFs,
as a dominant technology in the Brazilian sugar industry, although there are
surprisingly few publications on this technology for sugarcane mud filtration.
Vacuum Belt Filters have traditionally been used for dewatering in
mining, wastewater treatment, food production as well as pulp and paper
factory sludge processing due to their large capacity per unit size, high levels
of automation and superior filtration performance.

FILTRATION TECHNOLOGIES
8 T. J. Rainey, O. P. Thaval and D. W. Rackemann

Rotary vacuum filters have changed little over the past 20 years. While
they are the predominant filtration technology employed around the world,
there is a number of alternative filtration technologies used in the sugarcane
industry and these are described in the following.
New technologies may have the potential to decrease the cake moisture
while also increasing the sugar recovery and mud solids retention and hence
increase profitability for Sugar Mills. This section introduces traditional
Rotary Vacuum Filters, and then Vacuum Belt Filters are discussed which are
becoming increasingly popular. Other alternative filter technologies follow,
namely Membrane Press Filters and Centrifuges.

Rotary Vacuum Filters

Oliver Campbell RVFs are the most widely used filter in sugar factories.
The filter is composed of a hollow drum rotating about a horizontal axis and
party submerged in the mud feed (i.e., the filter boot) to be filtered (Figure 3).
Filter capacity and operation is improved by feeding the mud at multiple
points along the length of the filter boot. To avoid settling and stagnation of
the mud in the filter boot, an oscillating agitator (2–4 rpm) which breaks the
liquid surface is installed, pivoted on the axis of the filter and driven by a
separate motor. The filter drum is driven by a small electric motor to provide a
drum speed of 4–10 rpm.
The periphery of the drum serves as the filtering surface, divided into 24
independent sections, and extending along the full length of the drum. Each of
these sections is connected individually to a vacuum system by a small metal
pipe terminating in a distributing valve situated at one end of the drum and
applying three different functions:

1 One with connection to atmosphere

2 The second connecting with a chamber where a low vacuum, of the
order of 20 40 kPa (abs) is maintained
3 The third connecting to a chamber where higher vacuum is
maintained, of the order of >60 kPa (abs)

The filter screens are copper, brass or stainless steel and perforated with
~100–120 holes/cm2 which are 0.5 mm in diameter [20]. With due care, the
screens suffer very little wear and may be used for several seasons without
being replaced. They are, however, fragile and delicate.
 Developments in Mud Filtration Technology in the Sugarcane Industry 9

Operation
As the filter rotates, the section which first enters the mud boot is
immediately connected with low vacuum.

Figure 3. Diagrammatic operation of continuous rotary vacuum filter.

The liquid (filtrate) passes through the perforations which become coated
with fine bagacillo and suspended matter. The first filtrate is sent to the low
vacuum receiver (Figure 1). The cake continues to build, until the filter section
emerges from the mud in the filter boot.
After the screen rotates and emerges from the filter boot, it encounters
high vacuum. The juice passing through the cake is more effectively filtered
by its own impurities and by the fine bagacillo which it contains, forming the
necessary filtering surface which builds the cake on the surface of the drum.
This filtrate is sent to the high vacuum receiver.
10 T. J. Rainey, O. P. Thaval and D. W. Rackemann

The filter section is then sprayed with hot wash water. High temperature
wash water is essential since the viscosity of juice is inversely proportional to
temperature and lower viscosity favours filtration [21].

The vacuum draws the water through slowly, and is designed in such a
way that the water has just sufficient time, to pass through the cake and to
displace the juice.
Drying commences in the final stage of rotation as the water is drawn
towards the inside of the cake without further addition of water applied.
Finally the filter screens contact a scraper to remove the cake from the
drum. When the filter section is about to reach the scraper, the distributor
valve breaks the vacuum which has held the cake against the filter surface. The
slightest contact of the scraper detaches the cake dropping into a screw or belt
conveyor transporting the mud product to a storage bin prior to distribution
back onto cane fields.
The thickness of the cake is variable, generally 5–20 mm and depends on
the rotational speed of the drum and mud feed conditioning. The final filter
cake typically contains 0.5–3 % sugar, corresponding to a pol loss of 0.2–1.2
%.
The mud feed for filtration and wash water should be >80 °C to avoid the
risk of waxes blocking the filter screens and to prevent microbial growth.

Filter Sizing and Performance

The required filter area should be related to the mud solids entering the
factory. Generally for sugar factories with milling trains, two-thirds of the mud
solids entering the factory remain in the juice with the remainder leaving the
factory with bagasse. Historical RVF filter area ratings of 0.5–0.8 m2 per t/h
cane were standard for milling factories with lower ratings for diffuser
factories [11]. However, improved filter station capacity ratings provide 85–
100 m2 of filter area per 1 t/h of mud solids loading Pol loss in cake at this
value is expected to be around 0.3 % for wash water % cake of 200. The filter
mud solids loading ( , t/h/100 m2) is expressed in terms of (dry) mud solids
output per unit of filter area [16]:

(5)

 Developments in Mud Filtration Technology in the Sugarcane Industry 11

A similar term is used to rate the wash water application to the filters
expressed as wash water loading ( , t/h/100 m2), which is analogous to
mud solids loading and is defined as:

(6)

The sugar recovery achieved on the filter stations is linked heavily with
filter station capacity. Wright and co-workers [16] proposed a correlation for
pol loss in cake versus mud solids loading using data from a pilot filter.
The work was conducted using a constant wash water % mud solids value
of around 1200. The regression expression for these data was found to be:

.
0.15 0.1334 (7)

where

MSL is mud solids loading, t/h/m2; and

% %
(8)
%

where

% is the pol % mud solids in cake.

% is mud solids in cake % cane

The relationship between pol % MS and the wash water % mud solids in
cake ( % ) is given by the expression:

.
% 176.9 % 34.04 (9)

Using the empirical relationship based on pilot data predicts a pol % MS

value of 5–6 for typical industry wash water rates ( % of 1200). Both
expressions can be combined to provide an economic model to estimate the
value of pol % MS. The combined expression is given below:

. %
% 0.15 0.1334 100
%
.
32.17 % 6.19 (10)
12 T. J. Rainey, O. P. Thaval and D. W. Rackemann

Wright and co-workers [16] further developed their performance

relationship to fit factory data for estimating pol loss based on operating
parameters. As the % is dependent on , the regression was
simplified to determine the pol % MS as a function of mud solids loading and
wash water loading on the filters as follows:

. .
% 1.265 30.78 12.43 (11)

For data collected from a typical sugar factory, the estimated pol % MS
was plotted using the above equation for various mud solids and wash water
loadings to yield the trends shown in Figure 4.
With increasing wash water application rates there is a diminishing return
on improvements in sugar recovery. Increased amounts of wash water also
increases the amount of filtrate which when recycled to the process dilutes the
juice and increases evaporation loads.

Feed Conditioning
Steindl [4] undertook trials to measure the cake resistance for various
levels of feed conditioning. Filtration tests were performed using an apparatus
consisting of a piece of filter screen of approximately 100 mm diameter
connected through a condenser to a measuring cylinder. During a test, filtrate
passes through the screen and is cooled in the small, water cooled condenser
before collecting in the measuring cylinder. The whole apparatus is connected
to the wash filtrate receiver as the vacuum source.
 Developments in Mud Filtration Technology in the Sugarcane Industry 13

Figure 4. Prediction of RVF performance for various application rates and filter station
performance.

Figure 5. Filtration rate data showing the effects of added flocculant followed by the
addition of extra bagacillo [4].
14 T. J. Rainey, O. P. Thaval and D. W. Rackemann

The apparatus allows filtrate collection rates to be measured over short

time intervals to provide a gauge of cake resistance and filter performance.
The effect of bagacillo and flocculant addition on filtration performance is
shown in Figure 5.

Vacuum Belt Filters

Vacuum Belt Filters have been used in other industries to produce very
dry filter cakes and have been used in Australian sugar mills for ash
dewatering. In the 1980’s VBFs were assessed by Crees and Willersdorf [19]
and Kruger [22] for mud filtration and showed that reasonable filtration
performance could be achieved with sufficient feed conditioning. The addition
of flocculant was essential in reducing pol loss.
Crees and Willersdorf [19] found the pol loss was minimised with
flocculants at dosage rates of 300–350 ppm on mud solids.
Crees and Willersdorf [19] were able to produce cakes of 6–7 pol % mud
solids with wash application rates of 150 % on cake (~1000 % on mud solids)
when operating with vacuum levels of ~ 40 kPa (g).
Kruger [22] found the optimum operation of the VBFs was operation with
thin cakes of ~6 mm that allowed more effective cake washing at lower
vacuum levels (20–25 kPa (g)) to achieve cakes of 0.5 % pol and cake
moistures of 70–75 %. These performance figures are slightly better than
typical RVF operation. Further work on VBFs was discontinued in the mid
1980s due to higher capital and maintenance costs relative to RVFs at the time.
In the 1990’s, Technopulp Industrial (Brazil) developed a Vacuum Belt
Press Filters (VBPF) which incorporates an additional press section to the
VBF. Technopulp currently have over 600 units installed in South America.
The Technopulp VBPF is shown in Figure 6 and Figure 7. Following feed
distribution, a filter unit is comprised of three distinct sections:

● Gravity drainage section. Filtrate is allowed to drain from the fresh

cake. This filtrate is the cleanest of the three separate filtrates from the
filter. About 70 % of the filtrate is recovered from this section. Wash
water is added to this section.
● Vacuum drainage section. A low vacuum of up to 20 kPa (g) is
applied to draw off water prior to the press section. About 15–25 % of
the filtrate is recovered from this section. The last of the wash water is
 Developments in Mud Filtration Technology in the Sugarcane Industry 15

added at the start of the vacuum section. It is more a case of needing

air flow through the cake for drying rather than vacuum.
● Cake press section. The filtrate is squeezed between two belts as it
passes over a series of rollers. The filtrate from this section is about
2–3 brix and is of the lowest quality.

Operational and performance data provided by Technopulp include:

● Values of pol loss % pol in cane of 0.35–0.5 (cake pols of 1.25–1.5

%).
● Moisture of the cake is typically 60–65 %.
● Wash water application rates of 150 % on cake (~1200 % on mud
solids).
● Recommended cake thickness of 8–12 mm.
● Retention is typically about 96 %.
● Flocculant dosage at around 3 to 6 ppm on cane (400–450 ppm on
mud solids).
● Addition of lime required to maintain filtrate pH of 7.8–8.0.
● No bagacillo required if the level of bagacillo in juice is above 0.2 %.

www.technopulp.com.br.

Figure 6. Schematic of the Technopulp VBPF showing the different processing

sections.
16 T. J. Rainey, O. P. Thaval and D. W. Rackemann

www.technopulp.com.br.

Figure 7. Photograph of an installed Technopulp VBPF installed in Brazil.

VBPFs can process higher solids loadings per unit of filter area than RVF
(approximately 3 x higher). The largest Technopulp VBPF unit has a
processing capacity of 1.2–1.5 t/h of mud solids (37 m2 of filter area). Current
designs provide for multistage counter current washing for reduced water
usage and higher pol recovery. The VBPFs require slightly similar smaller
footprint area (and unit weight) compared to equivalent RVF processing
capacity and would require ancillary equipment such as a larger filter drive,
wash water and flocculant pumps.

Membrane Chamber Press Filters

Historically, plate and frame presses were the first technology applied for
mud filtration but lost favour in the 1960’s due to high footprint area required,
high operating costs (labour and maintenance) and unsatisfactory washing
efficiency [22]. However, the technology has been revisited in recent years.
A modification of simple plate and frame filter press is the Membrane
chamber Press Filter (MPF) system.
The chamber press consists of a frame and plates that are held in place
during operation by a hydraulic ram. The plates are recessed, forming
chambers that fill with mud (or wash water) during operation with one inner
membrane (diaphragm) forming the end of each chamber. The inner
 Developments in Mud Filtration Technology in the Sugarcane Industry 17

membrane is able to expand to squeeze the mud cake at higher pressures using
compressed air. The membrane squeezing step differentiates the MPF from
conventional plate and frame filter presses that only employ the filtration and
washing steps. The membrane squeezing step reduces the required feed
pressure (from 15 bar down to 6 bar), can produce cake of lower moisture and
allows shorter cycle times. The chamber plates including the membrane are
covered with porous cloths that are used to filter the mud. The area behind the
cloth collects the filtrate and channels it out of the plate for removal.
Allen and Wimmler [23] reported on a Netzsch MPF that was able to
produce cakes with around 10 pol % mud solids which is slightly higher sugar
loss than typical “good practice” RVF operation. The high pol losses resulted
from poor washing efficiency. The high levels of compaction of the mud
within the membrane chamber would help to trap pol in pockets and small
pores by increasing the specific cake resistance and hence limit displacement
washing mechanisms. Early designs also incorporated a feed directed into one
quadrant of the chamber rather than uniformly across all quadrants with
similar addition for wash water (but a different quadrant to the mud feed).
This sub-optimal design leads to non-uniform cake production over the
entire chamber and poorly distributed contact between cake and water and
coupled with the non-uniform cake thickness, non-uniform washing results.
Modifications to equipment design have helped to improve washing
efficiency. While the MPF operates in batch cycles, the process can be
automated and multiple units allow a measure of continuous operation to avoid
build up of mud stocks within the factory. One of the advantages of MPFs
other than their ability to produce low moisture cakes are their cleaner
operation with less vapour and dust produced compared to RVFs that are open
to the environment.
For equivalent processing capacity to RVFs, the MPF would require
almost double the filter area and weight but this is accommodated in a much
smaller and compact footprint area.
Additional ancillary equipment required includes a small hydraulic device,
membrane inflation compressor, large feed pump and cloth washing plant.
The MPF operates under the following steps which are illustrated in
Figure 8:

1 Filling. The plate pack of the press is closed and sealed by a hydraulic
cylinder. The membrane chambers are then filled with filter feed
(between the two cloths) by pumping the feed to the unit.
18 T. J. Rainey, O. P. Thaval and D. W. Rackemann

2 Filtration. The filter feed is pumped into the membrane chambers

under gradually increasing pressure (to a maximum of 5–6 bar).
Filtrate is forced through the end and inner membrane cloths and is
collected. Over time, the formed cake builds up and the feed pressure
rises to overcome the increasing resistance of the thicker cake. The
cake grows from the filter cloth inwards until it fills the entire
chamber.
3 Washing. Water is added under pressure to the inner membrane. The
water then passes through the mud cake displacing sugar as it is
directed through only the end cloths where the filtrate is collected.
4 Squeezing. Compressed air (15–20 bar) is added to the inner
membrane which expands to squeeze the mud cakes between the
cloths. Filtrate is forced through the end cloths and the inner
membrane cloths where it is collected.
5 Drying/flushing. Compressed air is added under pressure into the
membrane chambers to flush out any remaining fluid inside the mud
cake and through the end cloths where it is collected. It also removes
any remaining filtrate. A drying/flushing step may also be performed
both before and after the washing step to flush out the water and feed
from the inlet and filtrate pipes.
6 Cake discharge and cleaning. The pressure is released and the
chambers are opened. This allows the cake to be discharged by
gravity or mechanical means. The membrane cloths can then be
washed if necessary. The cake discharge step is fully automated in
commercial installations.

The washing, squeezing and drying steps are optional but help produce
low moisture cakes in shorter cycles.
Full scale MPF units have been installed and operated in China and Asia.
Reported operational conditions and performance data include:

● Cycle times of two hours with filtration to a maximum of 4 bar

pressure and cake washing at pressures of 4–6 bar. Membrane
squeezing was performed at 7–8 bar. Lower pressures reduce the wear
and tear on the filter and membrane cloths.
● Cake pols were ~4 % and cake moistures of ~ 50 % were produced.
● Wash water rates were similar to RVFs.

A typical industrial MPF is shown in Figure 9.

 Developments in Mud Filtration Technology in the Sugarcane Industry 19

www.andritz.com.

Figure 8. Diagram of the MPF operation.

www.andritz.com.

Figure 9. Typical industrial MPF manufactured by Andritz.

Centrifuge

Centrifuge technology has not received the same widespread acceptance

as RVFs and VBFs but this technology has merit as it can produce a product
with less than 50 % moisture and can efficiently recover sugar.
There have been numerous investigations by centrifuges in the past 40
years by two research organisations SRI/QUT and BSES [2, 24-27].
The first series of trials were conducted in 1973-1976 and the centrifuge
was installed at a factory over a number of seasons.
20 T. J. Rainey, O. P. Thaval and D. W. Rackemann

The technology was later reviewed in 2003-2005. The centrifuges used in

the 1970s investigations had limited capacity and the final economic
assessment favoured RVFs. Sugar recovery was aided by adding dilution
water to the feed or by adding wash water to the bowl, although adding wash
water had a negative impact on solids retention.

Previous Studies
Early investigations identified that although good sugar recovery was
possible, mud solids retention and erosion were key issues.

 Mud solids: Two separate early investigations found that increasing

the fibre ratio to 0.4 and using flocculant improved mud solids
retention to 65 %; some trials operating at a pH of 10 resulted in mud
solids retention of 95 % [24]. The quality of the bagacillo also plays a
major role with finer bagacillo greatly improving mud solids
retention.
 Erosion: The erosion was controlled by removing the coarsest
components (i.e., sand) using hydro-cyclones prior to the centrifuge
[25-27].

Although technically competitive at the time, centrifuges were noted to be

less economical than RVFs and Stewart and co-workers [25-27] foresaw that
the technology should be reviewed as their capacity increased. To this end, the
technology was reviewed in 2003-2005.
Operational Principle
The decanter centrifuge consists of a rotating inner screw contained within
a rotating casing. A cross-sectional view of a centrifuge during operation is
shown in Figure 10. The mud is fed through the inner screw and is discharged
into a horizontal zone called the centrate pool. The insoluble mud solids are
separated from the liquid (centrate) by centrifugal forces.
The mud solids are transported through to the discharge by a screw
conveyor which rotates at a slight speed differential to the bowl. The mud
solids concentrates through an incline to a region known as the ‘beach’ drying
region and the centrate overflows adjustable weirs at the opposite end.
The speed differential between the bowl and the scroll plays a vital role in
adjusting the final product moisture. The weir depth changes the amount of
solids in the centrate and also affects final product moisture.
Other parameters affecting performance include the feed rate, the level of
flocculate addition and how the feed mud has been conditioned.
 Developments in Mud Filtration Technology in the Sugarcane Industry 21

Internal baffles can also be installed to press the dry product and further
reduce moisture content.

Performance Characteristics
Two Alfa Laval centrifuges with a capacity of 10–12 t/h were
investigated. These were a P3400 and a G2 40 unit (Figure 11).

Courtesy of Alfa Laval.

Figure 10. Cross-sectional view of a decanter centrifuge.

22 T. J. Rainey, O. P. Thaval and D. W. Rackemann

Figure 11. Pilot G2 40 decanter centrifuge [2].

Trials with the G2 40 unit were more successful mainly because of the
higher level of instrumentation, particularly torque monitoring instrumentation
which was useful in providing information about the degree of separation
between solids and liquid within a centrifuge. Final cake moistures of 50 %
were possible. The investigation ultimately found that the centrifuge was
slightly more expensive than an equivalently sized RVF and the technology
was not implemented in Australia. Since the 2003-2005 study, Alfa Laval have
developed a G3 unit which it claims is capable of 10 % higher throughput (or
dryer cake) and 40 % lower energy costs [28].
Trials showed that with no flocculant addition, increasing the fibre ratio
increases mud solids retention (Figure 12) but reduces the final moisture
(Figure 13). However using relatively high flocculant levels (~1000 ppm on
mud solids), the centrifuge was able to achieve virtually 100 % mud solids
retention and product of 53–58 % moisture for a feed fibre ratio of 0.3–0.4. At
high flocculant rates, the feed rate to the centrifuge had little impact on mud
solids retention.

Distribution of Mud Product

One of the interesting features of the VBFs and Centrifuge is the

spreadability of the final mud product.
The cake produced from VBFs was easy to spread with existing trucks as
the cake granulated and crumbled readily based on its low moisture. Spreading
the drier mud product from a Centrifuge was investigated using two methods:
(i) a truck which is conventionally used to spread gypsum (these trials
generated a lot of dust, see Figure 14, which could be potentially reduced by
using more moist mud); and (ii) a standard mud truck which discharges by
tipping the mud from the back.
Figure 15 shows the auger distributor at the back and base of a truck bed
that helps distribute the mud product during spreading. The mud is evenly
distributed onto cane fields, see Figure 16. Little dust was generated during
unloading from standard mud trucks although as the mud dried a small amount
of dust was created. Other mud transport trucks such as fertiliser trucks with
rotating discs or slat conveyors rather than the auger arrangement shown in
Figure 15 are utilised in Brazil for distributing drier mud product.
 Developments in Mud Filtration Technology in the Sugarcane Industry 23

Centrifuges produce a product which has relatively low bulk density.

Compared to RVFs, trucks carrying centrifuged mud would need to be slightly
larger (~10 %) to take the same number amount of mud solids whilst still
carrying the same weight. On the other hand, the mud cake produced by VBFs
was of similar density to RVF cake and due to lower moisture would require
less number of trucks to transport the same amount of mud solids.

Figure 12. Effect of fibre ratio on mud solids retention (no flocculant).
24 T. J. Rainey, O. P. Thaval and D. W. Rackemann

Figure 13. Effect of fibre ratio on the final moisture of mud product (no flocculant).

Figure 14. Centrifuge mud spread by gypsum truck during trials [2].

The distribution of mud product onto cane fields is not the only transport
issue. The conveyors, storage and hopper systems are one area that currently
restricts the transport of drier mud products, at least in Australia due to current
equipment configurations.
Some mills add additional water to the conveyors and hopper systems to
improve mud transportability and flowability into the mud transport trucks.

Figure 15. Auger distributor at the back base in the bed of the mud truck.
 Developments in Mud Filtration Technology in the Sugarcane Industry 25

Figure 16. Centrifuge mud spread by a conventional mud truck.

FURTHER DEVELOPMENTS
IMPACTING ON MUD SEPARATION
Mud Solids Recycling

The juice clarification system employed in the sugar factory is limited by

its capacity to process mud solids. Mud filtration aims to retain as much of the
mud solids in the cake but often the filtrate is recycled back to the process
adding 15–20 % of mud solids to the clarification system.
In some overseas factories a separate clarifier for filtrate is utilised to
produce filtrate that is forward processed within the factory to improve juice
clarification capacity and performance [29]. This processing strategy is largely
based on the need to improve overall factory sugar quality and recovery that
results from the detrimental effects of recycling filtrate backwards to the
process. Filtrate can also be added to the mud to condition the filter feed for
improved filtration and reduced sugar losses [11].
Diluted filter feed can produce cakes of high porosity enabling more
effective washing but typically 10–20 % filtrate is recycled in this manner.

Automation
26 T. J. Rainey, O. P. Thaval and D. W. Rackemann

The filter station is one of the last remaining sections of the factory that
relies heavily on operators to maintain optimum performance. While the
monitoring of mud conditioning characteristics such as mud solids loading and
fibre ratio prior to the filter would provide pre-emptive advantages for
optimising control of the filter stations, such instruments are not currently
commercially available. An alternative is the monitoring of operating
parameters after the clarifier to provide an early indication of filter processing
problems. Key parameters requiring continual operator monitoring and manual
adjustment include the mud-juice interface in the clarifiers and the
conditioning of the feed to the rotary vacuum filters. The mud level can
change dramatically in minutes depending on the quality of the cane supply.
Smoothing out the flow of mud solids within the factory with increased
levels of automation is a good strategy for improving the performance of the
clarifier and filter station. The performance of the clarifier is directly linked to
the performance of the filters through the recycle of soluble and insoluble
impurities in the filtrate and the operation of the clarifier is a critical factor in
achieving good quality sugar. In recent years ultrasonic sonar, fibre optics,
guided wire radar and penetrating pulse devices have been considered for the
measurement of the mud-juice interface level in clarifiers.
However these instruments need to overcome the harsh and specific
operating conditions with a requirement for the mud level transducer to
withstand temperatures in excess of 100 °C while fully submerged in cane
juice, the system must be capable of identifying a mud-juice interface with a
density differential of no more than 10 kg/m3 and the system must providing a
signal that is suitable for control, without interference from suspended mud
particles, fibre and moving or stationary clarifier components.
Similarly microwave technology has been considered for indirectly
measuring density and concentration as a gauge of mud conditioning [30].
These two examples show how if these technologies can be utilised,
control procedures could be developed to provide optimal performance of the
clarifier and filter stations.

COMPARISON OF MUD FILTRATION TECHNOLOGIES

Table 1 presents a brief comparison between the technologies. The list is
presented in the order of the most prominent technology in the industry
(RVFs) to the least prominent (i.e., Centrifuges).
 Developments in Mud Filtration Technology in the Sugarcane Industry 27

The filtration performance of RVFs is very well established and so are the
issues, such as large footprint, high mud moisture content, high initial
investment costs and known maintenance issues. RVFs will continue to be
utilised in new factories based on risk minimisation although the main interest
in RVFs is in improving performance of existing units.
Vacuum Belt Filters are becoming increasingly popular due to their lower
initial investment cost, lower final moisture content and excellent mud solids
retention but they have higher maintenance costs.
The authors expect the proliferation of VBFs to continue in their various
embodiments (with and without vacuum and additional press sections) and that
performance will continue to improve. Membrane Press Filters produce cake
with excellent final moisture content and they perform with good mud solids
retention and a slight cost advantage over RVFs.
However, this is evaluated against higher sugar losses and maintenance
costs. It is a little unclear as to what their future uptake will be, but recent
trends suggest further reduction in capital costs and improved designs to
reduce sugar losses will increase the installation of MPF technology in the
sugar industry.
Finally Centrifuges also produce excellent final moisture content, produce
an easily distributed mud and have a small footprint, although electricity costs
are high and maintenance requirements are uncertain.
Centrifuges are becoming more cost-competitive with RVFs but may not
become mainstream due to the recent proliferation of VBFs as the main
alternate technology. They may find a niche in factories where footprint is a
primary concern or the company has a certain strategy for distributing mud.
 Table 1. Comparison of mud separation technologies

Final
Mud solids Distribution of
Technology Sugar recovery moisture Cost Ancillaries Footprint Maintenance
retention mud
content
Medium Long life expectancy.
Rotary Reasonable (0.3– electricity Issues with screen
Can be
Vacuum 0.4 pol loss % 75–80 % >90 % consumption Largest maintenance which is Limiting
expensive
Filters pol in cane) -vacuum pump dependent on feed
-wash sprays conditioning.
Reasonable if Higher electricity Slightly more
Relatively cheap
high flocculant consumption maintenance than
Vacuum Belt especially if Slightly less than
used (0.3–0.35 60–70 % >95 % -vacuum pump RVFs. Need to Reasonable
Filters manufactured in RVFs
pol loss % pol in - floc pump replace belts every 2-
low cost country
cane) -wash facilities 4 years
Medium
electricity Medium but heavy
Fair (0.6–0.7 pol
Membrane Slightly cheaper consumption weight (structural Higher costs than
loss % pol in 55–60 % >95 % Good
Press Filters than RVFs -feed pump requirements may RVFs.
cane)
- compressor be needed)
-wash facilities
Reasonable with
Comparable to Very high
More expensive sufficient feed Excellent
RVF. For good electricity
than RVFs but conditioning. Long although dust
Centrifuges recovery, require 50–60 % >85 % consumption Smallest
continually term costs not known potentially an
internal basket (approx double
improving due to limited issue
wash RVF usage)
application
 T. J. Rainey, O. P. Thaval and D. W. Rackemann 29

When comparing mud separation technologies, a factor for consideration

is the availability of expertise for the selected technology within the sugar
company and at cooperating nearby sugar factories.

ACKNOWLEDGMENTS
The authors acknowledge the work of Mr Rod Steindl and Dr. Floren
Plaza for their contributions on the various filtration technologies.
The authors are not associated with distributors of any filtration
technology and there is no conflict of interest.

REFERENCES
[1] Doherty, W. O. S., et al., Flocculation and sedimentation of cane sugar
juice particles with cationic homo- and co-polymers. Journal of Applied
Polymer Science, 2003. 90(1): p. 316-325.
[2] Steindl, R. J., T. J. Rainey and F. Plaza, Low moisture mud from solid
bowl decanters, In: Proceedings of the International Society of
Sugarcane Technologists, 2010: Veracruz, Mexico.
[3] Qureshi, M. E., M. K. Wegener and G. M. Mason, Mill mud case study
in Mackay. 2000, Cooperative Research Centre for Sustainable Sugar
Production: Townsville.
[4] Steindl, R. J., Reducing the specific cake resistance to reduce the pol
loss in filter cake, In: Proceedings of the International Society of
Sugarcane Technologists, 2001. p. 392-394.
[5] Rainey, T. J., A study into the permeability and compressibility
properties of Australian bagasse pulp, In: Faculty of Built Environment
and Engineering. 2009, PhD Thesis, Queensland University of
Technology: Brisbane.
[6] Banks, P. J., Theory of constant-rate expression and subsequent
relaxation, in Chemeca, 1984. p. 813-822.
[7] Owen, D. R. J., S. Y. Zhao and J. G. Loughran. An overview of crushing
theory investigations at Swansea. In: Proceedings of the Australian
Society of Sugarcane Technologists, 1994. p. 264-270.
30 T. J. Rainey, O. P. Thaval and D. W. Rackemann

[8] Kent, G. A. and N. J. McKenzie, A simple test for determining some

important material properties for bagasse. Proceedings of the Australian
Society of Sugarcane Technologists, 2003. cdrom.
[9] Landman, K. A., C. Sirakoff and L. R. White, Dewatering of flocculated
suspensions by pressure filtration. Physics of fluids A, 1991. 3(6): p.
1495-1509.
[10] Landman, K. A., L. R. White and M. Eberl, Pressure Filtration of
Flocculated Suspensions. AIChE Journal, 1995. 41(7): p. 1687-1700.
[11] Rein, P., Cane sugar engineering. 2007, Berlin: Bartens.
[12] McGrath, G. J., Filtration and washing aspects of filter mud, In:
Proceedings of the Queensland Society of Sugarcane Technologists,
1971. p. 261-267.
[13] Noble, A. G., The use of polyelectrolytes in mud filtration, In:
Proceedings of the Queensland Society of Sugarcane Technologists,
1971. p. 251-259.
[14] Noble, A. G., The effect of flocculants on filter performance.
Proceedings of the Queensland Society of Sugarcane Technologists,
1972. p. 303-314.
[15] Wright, P. G., Modelling the effects of soil and other factors on mud
filtration, In: Proceedings of the Australian Society of Sugarcane
Technologists, 2006. p. 507-517.
[16] Wright, P. G., C. C. Steggles and R. J. Steindl, The balance between
capacity and performance of rotary mud filters. Proceedings of the
Australian Society of Sugarcane Technologists, 1997. p. 434-440.
[17] Hale, D. J., E. Whayman and A. L. Willersdorf, Mud filtration-
preliminary studies, In: Proceedings of the Queensland Society of
Sugarcane Technologists, 1971. p. 269-279.
[18] Lionnet, G. R. E., Mud conditioning for good filter operation, In:
Proceedings of the South African Sugar Technologists Association,
1984. p. 39-41.
[19] Crees, O. L. and A. L. Willersdorf, An evaluation of a horizontal
vacuum filter, In: Proceedings of the Australian Society of Sugarcane
Technologists, 1983. p. 215-218.
[20] Hugot, E. and G. H. Jenkins, Handbook of cane sugar engineering.
Sugar Series, ed. G. H. Jenkins. Vol. 7. 1986: Elsevier.
[21] Kulkarni, D. P., Cane sugar manufacturing in India. Vol. 1. 2009: The
Sugar Technologist Association of India (STAI).
 Developments in Mud Filtration Technology in the Sugarcane Industry 31

[22] Kruger, G. P. N., The application of a horizontal vacuum belt filter to

smuts dewatering and cane mud filtration, In: Proceedings of the South
African Sugar Technologists Association, 1984. p. 34-38.
[23] Allen, C. J. and W. Wimmler, The membrane chamber filter press- an
alternative to rotary vacuum filters, In: Proceedings of the Australian
Society of Sugar Cane Technologists, 1994. p. 340-343.
[24] Hale, D. J., et al., Mud Centrifugation- Preliminary Trials, In:
Proceedings of the Queensland Society of Sugarcane Technologists,
1974. p. 243-248.
[25] Stewart, P., et al., Evaluation of a solid bowl centrifuge, In: Proceedings
of the Queensland Society of Sugar Cane Technologists, 1974. p. 249-
259.
[26] Stewart, P., A. G. Noble and G. A. Brotherton, Centrifuge perofmance in
the treatment of cane mud, In: Proceedings of the Queensland Society of
Sugar Cane Technologists, 1975. p. 319-329.
[27] Stewart, P., A. G. Noble and G. A. Brotherton, Improved mud
centrifugation at Mossman, In: Proceedings of the Queensland Society of
Sugar Cane Technologists, 1975. p. 235-240.
[28] Alfa Laval. Aldec G3. 2014 [cited 2014 10/06/2014]; Available from:
http://www.alfalaval.com/products-and-solutions/pages/products-and-
solutions.aspx.
[29] Singhal, S., et al., Treatment of filtrate from rotary vacuum filters to
avoid recycling, In: Proceedings Of the International Society of Sugar
Cane Technologists, 2005. p. 170-171.
[30] Theisen, K. H. and I. Geyer, Microwave technology: An ubiquitous in
line measurement instrument for density, total solids, concenrtation,
especially brix in today's sugar factory. International Sugar Journal,
2007. 1301(109): p. 310-319.

N.G.

View publication stats