Proc. Int. Soc. Sugar Cane Technol.

, 24: 80-85 2001

DEVELOPMENT OF THE NEW GENERATION SRI CLARIFIER DESIGN

R.J. STEINDL
Sugar Research Institute, Mackay, Queensland, Australia

Abstract
This paper reports on the development of the new generation SRI clarifier design through three
phases. The first phase involved investigations to model the juice flows in a SRI clarifier using the
CFD package FIDAP. Secondly, the model predictions were validated with factory trials on a
modified clarifier incorporating design modifications developed through application of the models.
The third phase involved the ongoing development of modified clarifier designs to eliminate
previously identified shortcomings with the standard SRI clarifier. Design studies have included an
examination of the feed distribution system and short-circuiting characteristics, of the existing feed
launder system. Several new generation clarifiers have been operational since the first prototype was
commissioned in late 1996. The benefits include significantly higher throughput rates, improved
standards of clarification and reduced sugar losses. The results represent a significant gain to the
Australian sugar industry, not only in clarification performance but also in reducing the cost for mill
throughput expansions.

Introduction in understanding the fluid flows within the clarifier

The SRI clarifier, first installed at Gin Gin sugar and to assess the impact of design variations on the
mill in 1969 (Hale and Whayman, 1970), represented fluid flows and mud separation. At the time, FIDAP
a major advance in clarification technology for the was considered to be the best computer package for
Australian cane sugar industry. The SRI clarifier has the analysis of two phase flows in a constrained flow
served the industry well since then, providing high space. It also provided a user-friendly interface to aid
quality clarified juice at throughput rates not envis- in the development of models involving complex
aged when the unit was first developed. However, the geometries.
SRI clarifier does have minor operational deficien- Steindl (1995) reported on initial investigations to
cies. These include its sensitivity to upstream process examine the juice and mud particle flows in a standard
fluctuations, possible short-circuiting of juice and 12 m diameter SRI clarifier processing 600 m3/h of
mud, and difficulties in obtaining an even juice distri- juice. A number of aspects of the flow pattern were
bution. The latter is most likely to occur at lower identified that could contribute to poor clarification
juice flow rates. performance under certain circumstances including:
In the sugar industries of other countries, the SRI strong horizontal flows exiting the gap between
clarifier and its variants, often referred to as short the feedwell and the deflector which then
residence time (SRT) clarifiers have been widely flowed upward and directly to the three take-off
adopted either through purchasing via previously launders;
licensed agents, in some cases copying it or, more a relatively small proportion of the available
recently, through direct purchase of designs from area was used for upflowing juice; and
SRI. The incorporation of SRT clarifiers into some significant recirculation regions were present
overseas factories has not been successful because of and these affected the settling trajectories of
a lack of appreciation of the critical factors influ- mud particles.
encing the clarification performance in the SRI Two options that demonstrated some potential for
clarifier. However, recent installations undertaken improved flow conditions were to increase the gap
through SRI have been very successful. between the bottom of the feedwell and the deflector,
Feedback from installations in several countries and to alter the location of the feedwell relative to the
indicate: wall of the vessel.
improved clarification standards; CFD modelling was also applied by Steindl
reduced usage of lime and flocculant; (1996) to investigate the effects on juice flows and
reduction in ESJ colour; and mud settling trajectories by changing the floor shape
more consistent mud to the filters. and relative dimensions of the internal components of
A number of these installations are in sulphitation a 10.7 m clarifier that was being installed to clarify
factories producing white sugar. juice from a cane diffuser.
Investigations commenced in 1994 to address the More recently, Peacock et al. (2000) reported on
issues affecting the clarification performance in a initial work to investigate juice flows in a Rapidorr
standard SRI clarifier. The computational fluid 444 multitray clarifier using the FLUENT CFD
dynamics (CFD) package FIDAP was used to assist package. He reported some success in changing the

I KEYWORDS: SRI Clarifier, Design, Mud, CFD, Multitray Clarifier, Juice.

Steindl, R.J., Proc. Int. Soc. Sugar Cane Technol., 24: 80-85 2001

fluid dynamics by using small baffles located at A major project was commenced in 1996 to
various sites within the compartment. validate the CFD model predictions with data from
factory trials on a modified clarifier. Bundaberg
Development of the new clarifier design Sugar Ltd made available a 5.5 m clarifier that was
Initial factory trials used only during periods when the cane supply con-
CFD models were used to examine the potential tained high levels of dirt and when the two 7.3 m
of a number of different and innovative approaches to clarifiers at the factory struggled to cope with the
changing the juice flows to use more effectively the high mud loading in the clarifier feed. Using current
available cross-sectional area within the clarifier. A sizing criteria, a 5.5 m clarifier in standard SRI con-
standard 10.4 m design processing 700 m3/h of juice figuration would have a notional maximum rating of
was modelled as a two-phase flow problem in an axi- 150 m3/h of secondary juice. When the design con-
symmetric geometry. The equations of the juice cept was incorporated into a CFD model of the 5.5 m
phase were the standard conservation equations of clarifier processing 150 m3/h of juice, the prediction,
fluid dynamics for momentum, mass, energy, turbu- as shown in Figure 3, indicated an opportunity for
lent kinetic energy, and turbulent dissipation. The improved clarification standards and the potential for
flow was assumed to be isothermal. A view of the a significant increase in throughput capacity. The
clarifier design and the section that was modelled are prediction illustrates the elimination of the high
given in Figure 1. velocity flow from the deflector towards the clarified
The model predictions are shown in the velocity juice launder (i.e. there are no clear areas indicating
vector plot of Figure 2. The maximum velocity flows greater than 10 d s ) and the absence of
shown in Figure 2 has been limited to 10 mmls to strong recirculating flows. The low flow velocity as
avoid congestion. This is twice the industry design the juice leaves the deflector will allow the floccu-
settling velocity of the mud particles (-5 d s ) . Thus lated mud particles to separate from the juice more
the clear regions in the feedwell and the two streams quickly than is the case in the standard design.
moving away from the bottom of the feedwell repre- When the modifications were incorporated into
sent juice flows with velocities higher than 10 mmls. the clarifier and validation trials undertaken, clarifi-
The dark areas either side of the clear areas show cation performance was at least equivalent and gener-
velocities close to but less than 10 d s . The short ally superior to the performance being achieved in an
lines (vectors) of which the picture is composed show adjacent clarifier of twice the cross-sectional area
the direction of the flow. handling the same juice throughput rate. Juice rates in
Concepts that were investigated using CFD excess of 300 m3/h (more than twice the notional
models included centre, intermediate and peripheral maximum design rate) were achieved through the
feed arrangements as well as options to remove clari- modified clarifier without any detectable deteriora-
fied juice. Fluid dynamic models of another concept tion in clarification performance.
demonstrated the potential to have a significant This new generation prototype performed
impact on the juice flows. extremely well throughout the 1997 crushing season

JUICE LAUNDER 7 LAUNDER^

MIDDLE CLARIFIED INNER CLARIFIED
JUKE
FEED LAUNDER OUTER CLARIFIED
FEEDWELL [JUICE LAUNDER

Fig. 1-A sectional view of a standard SRI clarifier with a single feed entry point. The section used in the CFD models is lightly shaded.
Steindl, R.J., Proc. Int. Soc. Sugar Cane Technol., 24: 80-85 2001
I

Mud-juice interface
Fig. 2-Velocity vector plot for juice flow in a standard 10.4 m SRI clarifier processing 700 m3/h. Vectors are not shown in regions where
the velocity exceeds 10 mmls.

t Clarified juice Feed 4

Mud-juice interface
Fig. 3-Velocity vector plot for the 5.5 m clarifier processing 150 m3/h when the design concept is incorporated. Vectors are not shown
in regions where the velocity exceeds 10 mmls.

despite the adverse processing conditions experi- Further development

enced by all mills in the region during the early part The design of the single point feed system for the
of the season. Because of the effects of cyclone SRI clarifier is based on achieving constant velocity
Justin, all cane in the region was heavily sprawled of juice flow in the feed launder with the objective of
and cane supplies delivered to the mills contained obtaining uniform distribution of feed around the full
very high levels of dirt. The clarifier consistently circumference of the feedwell. The feed launder
procqssed juice at 50%, on average, above the design design incorporates a number of slots of uniform size
loading and achieved very high clarification around the wall separating the feed launder and the
standards. feedwell. It has been observed that, at flow rates
Full details of the validation trials are provided by substantially lower than design, the amount of juice
Steindl et al. (1998). reaching the shallow section of the feed launder can
Steindl. R.J., Proc. Int. Soc. Sugar Cane Technol., 24: 80-85 2001

be significantly lower than the mean flow, thus Analysis of the trajectories of the flocculated mud
resulting in poor juice distribution. particles in the CFD models indicate that most mud
It was determined through preliminary factory flocs separate from the juice flow streams as the juice
trials that, to achieve a uniform feed distribution, it leaves the bottom of the feedwell. The magnitude of
was necessary to reduce the depth of the first slots the juice velocity in this region determines the final
either side of the feed chamber and then to succes- path of the settling flocs. At high feed rates, such as
sively position the base of the following slots in a pro- those shown in Figure 2, the high juice velocities can
gression down to the original level of the last slots. cause significant carryover of small (<2 mm) par-
A program was developed using open channel ticles with the clarified juice leaving the clarifier.
flow and weir calculations to predict the flow through Various deflector arrangements were modelled to
each slot of any feed launder design. To verify the determine the most effective method of dissipating
program design and the equations selected, data from the momentum. The constraint applying to the design
the preliminary factory trials were compared with the of any internal components is that all horizontal
calculated dimensions from the program. The calcu- surfaces have to be scraped to prevent accumulation
lated slot dimensions agreed to within k 1 mm of the of settled mud. A simple two-stage deflector arrange-
slats that provided the closest approach to uniform ment achieved the desired result.
feed distribution in the factory trials. Model predictions for a new generation clarifier
Subsequently options were considered for the design are shown in Figure 4 for the same clarifier
design of the feed slots that would achieve a closer size and feed conditions as applied to the predictions
approach to uniform juice feed distribution. The shown in Figure 2. The reductions in juice velocity
options included: and the improved use of the available cross-sectional
fixed width, variable depth slots; and area are demonstrated.
fixed depth, variable width slots.
Simulations for a range of flow rate conditions Implementation into Australian factories
indicated that the option of variable width, constant The new generation clarifier design was first incor-
depth produces a more uniform feed distribution. porated into existing clarifiers in 1998. Since the first
' It has also been noted that the flow of juice through release to the Australian sugar industry, 19 design
the feed slots can result in a high level of localised tur- specifications have been prepared for factories. In all
bulence as the juice cascades onto the surface of the cases where the modifications have been imple-
juke in the feedwell. This can cause break-up of mud mented, factory staff have reported the following
flocs, producing small fragments and releasing changes in clarification performance:
trapped bagacillo particles. As a consequence, the feed improved clarity of the clarified juice;
launder has been incorporated into the feedwell. This * higher throughputs coupled with higher stand-
has; eliminated both the problems with floc break-up ards of performance;
and other problems that have been reported for the improvements in those sugar quality parameters
s a ~ slots
d such as blockages of the return chute. directly affected by clarification performance;

Mud-juice interface
Figl. 4-Velocity vector plot for juice flow in a new generation 10.4 m SRI clarifier processing 700 m3/h. Vectors are not shown in regions
where the velocity exceeds 10 mm/s.
Steindl, R.J., Proc. Int. Soc. Sugar Cane Technol., 24: 80-88001

higher mud densities; Summary

reduced levels of bagacillo carryover; and Techniques have been developed that allow the
reduction in flocculant usage by 10-15%. juice and mud particle flows in clarifiers to be
One of the few remaining multitray clarifiers was simulated using computational fluid dynamics. The
converted to the new generation design for the 1999 use of the CFD package FIDAP has enabled an
crushing season. The clarifier is 6.7 m in diameter analysis of the flows within a standard SRI clarifier
and the internal components were completely design. This analysis identified a number of issues
replaced as part of the conversion process. The feed that could adversely affect clarification efficiency
juice at the factory usually contains a relatively high and performance.
dirt loading. For the two years 1998 and 1999, the
mud solids loading in the mixed juice averaged 2.6%. Alternative designs have been investigated to
The performance comparison as provided by the address those issues and, in particular, to make better
Production Superintendent follows: use of the available cross-sectional area. One design
concept that, according to the CFD model predic-
Parameter Multitray Converted tions, offered significant potential was incorporated
(1998 season) to SRI into an existing clarifier for assessment. Results from
(1999 season)
the factory trials validated the improvements
Juice feed rate, m3/h 220 295
predicted by the models.
Turbidity of clarified 45 10
juice ( a 9 0 0 nm) Following extensive investigations of all aspects
of the standard SRI clarifier design, a new generation
Phosphate removal, % 93 97
design has emerged that represents a significant leap
Primary mud density "Light" "Well forward in improving clarification performance.
thickened
Experiences gained in the three years 1998-2000 in
Ease of operation "Difficult" "Easy"
Australian sugar factories with clarifiers incor-
porating the modifications have demonstrated that
It is expected that this converted clarifier will higher standards of throughput and performance are
now have a maximum juice throughput exceeding readily achievable.
380 m3/h. This project has highlighted the benefits to be
gained from computer modelling. Access to the CFD
Other applications modelling package FIDAP has provided an oppor-
The design principles have also been applied suc- tunity to investigate juice flows for a number of
cessfully to an ash clarifier. A small ash clarifier clarifier designs before the best design is selected
featuring components of the new generation juice and before any plant modifications are made.
clarifier was installed in a factory in 1998 to process
wet scrubber water from a bagasse boiler. The clari- Acknowledgments
fier has 'demonstrated significant improvements in
performance and throughput capacity when compared The contribution and assistance provided by
with an adjacent existing clarifier which is 15% Bundaberg Sugar Ltd and, in particular, Messrs
larger but limited to about 40% of the throughput Alan Fitzmaurice and Craig Alman are gratefully
rate. A major advantage is the significant reduction in acknowledged.
the size of the dewatering plant (rotary drum or belt Part funding for the projects was provided by the
filter) necessary to treat the ash underflow from the Sugar Research and Development Corporation. They
clarifier. also funded the initial purchase of FIDAP.

REFERENCES
Hale, D.J. and Whayman, E. (1970). The Sugar Research high capacity subsider. Proc. Qld Soc. Sugar Cane Technol., 37: 243-248.
Peacock, S.D., Davis, S.B., Govender, K.A., Moodley, K. and Brouckaert, C.J. (2000). Computational fluid dynamics modelling of a
Rapidorr 444 clarifier. Proc. S. Afr. Sugar Technol. Assoc., 74: in press.
Steindl, R.J. (1995). Optimum performance through CFD modelling of clarifier designs. Proc. Aust. Soc. Sugar Cane Technol.,
17: 207-215.
Steindl, R.J. (1996). Optimum performance through CFD modelling of the SRI clarifier design. Proc. S. Afr. Sugar Technol. Assoc.,
70: 145-148.
Steindl, R.J., Fitzmaurice, A.L. and Alman, C.W. (1998). Recent developments in clarifier design. Proc. Aust. Soc. Sugar Cane Technol.,
Steindl. R.J.. Proc. Int. Soc. Sugar Cane Technol., 24: 80-85 2001

DEVELOPPENT DU DECANTEUR SRI: NOUVELLE CONCEPTION

R.J. STEINDL
Sugar Research Institute, Mackay, Queensland, Australia

On a dCveloppC le nouveau dCcanteur SRI, en trois phases. La premikre comprend la modClisation de

la circulation du jus a travers du logiciel FIDAP. Deuxihmement, on a valid6 les prkdictions du
modkle en appliquant les rCsultats a un dCcanteur industriel modifie pour reprCsenter les rCsultats du
modkle. La troisihme phase comprend le dCveloppent continue1 de modifications pour corriger les
dksavantage du SRI standard. Les Ctudes ont inclus l'examen du systkme d'alimentation et de
distribution du jus. Le premier prototype a CtC installe en 1996, et depuis plusieurs dkcanteurs sont
en service. Les avantages comprennent des dCbits de jus plus forts, une meilleure clarification et des
pertes de sucre rCduites. Ce dCcanteur reprCsente un sQieux avantage pour l'industrie Australienne
non seulement pour la clarification elle meme mais aussi pour les coats d'expansion de l'industrie.

Mots clefs: dCcanteur SRI, conception, boue, CFD, jus.

EL DESARROLLO DE LA NUEVA GENERACION DISERO DEL SRI CLARIFICADOR

R.J. STEINDL
Sugar Research Institute, Mackay, Queensland, Australia

Resumen
Este papel informa en el desarrollo de la nueva generaci6n el SRI clarificador planeado a travCs de
tres fases. La primera fase involucra las investigaciones para modelar 10s flujos de jug0 en un
clarificador de SRI que usa el CFD software FIDAP. En la segunda fase, las predicciones modeladas
se validaron con ensayos de fibrica en un clarificador modificado que incorporo las modificaciones
desarrolladas a travCs de la aplicaci6n de 10s modelos disefiados en la primera fase. La tercera fase
involucr6 el continuado desarrollo de clarificadoras modificadas, disefiadas para eliminar las
limitaciones previamente identificadas con el clarificador de SRI estindar. Los estudios para el
disefio han incluido un examen del sistema de distribuci6n de suministro de jug0 y el bypass del
sistema de clarificado existente. Varias nuevas generaciones de clarificadores han sido operacionales
desde que el primer prototipo fue comisionado a finales de 1996. Los beneficios incluyen un
significativo aumento en el ritmo de producci6n, normas mejoradas de clarificaci6n y las pCrdidas de
azdcar reducidas. Los resultados representan una ganancia significante a la industria de azdcar
australiana, no s610 en el rendimiento de la clarificaci6n peso tambiCn reduciendo el costo para las
expansiones de producci6n del ingenio.