NEW THICKENER WITH THICK PASTE REMOVAL SYSTEM

Boivin A, Dufour R, Peloquin G* and St-Laurent M

Rio Tinto Alcan, Arvida Research and Development Centre, Québec, Canada
Abstract
The objective of most thickeners is to produce mud at the highest possible solids fraction in order to improve the separation
efficiency. Extracting this dense mud from a thickener can be challenging and operation at high underflow rate is often
causing channelling where liquor is short circuiting to the underflow. A new paste removal system that allows extraction
of very thick mud at high flow rates was designed and tested in our laboratory. It consists of a device that mechanically
extracts the mud out of the thickener and brings it to the underflow pump. Consistent operation with red mud at up to 54%
solids was obtained in our laboratory. These results led to the development of a 3-m diameter flat-bottom pilot thickener
fitted with the new paste removal system. The system is described and typical results are presented.

1. Introduction the underflow although high solids fraction mud is present in

the thickener. It is affected by mud residence time, mud bed
Over the last few years, significant work has been carried out
height, underflow pump rate fluctuations, flocculant dosage and
in our research centre to study red mud flocculation (Peloquin
underflow configuration.
et al., 2005a) and compaction in thickeners (Peloquin et al.,
2005b). Some of this work indicated that operational problems To evaluate the impact of the thick paste removal system on these
are often related to the difficulty of extracting red mud at high two issues, the first step was to select operating parameters that
solids fraction out of thickeners. To allow pumping of such high would display such behaviour. This was done by running tests
solids fraction mud, the concept of a screw discharge underflow from a combination of mud levels, mud residence times and feed
system was proposed (Peloquin et al., 2007) and tested in our mass fluxes. The solids fraction at the underflow was measured
laboratory. Following preliminary good results obtained with a when it reached equilibrium. Note that during each test, the
small flat-bottom pilot thickener fitted with this system, it was conditions were kept constant for about three times longer than
determined that a larger pilot thickener was needed to validate the the target mud residence time.
performance of this thick paste removal system at a scale more
comparable to that of industrial thickeners. This paper presents
results from the preliminary evaluation of that performance. The
evaluation of the system operating with high solids fraction mud
and its ability to control ratholing are presented. The impact on
mud rheology is discussed. The technical advantages are later
presented.

2. Experimental
Tests were carried out in one of the Rio Tinto Alcan alumina
refineries. It is installed in parallel to the last washer of the
mud circuit. The pilot thickener is a flat-bottom cylindrical tank Figure 1. Diagram of the screw thickener with the thick paste removal system
with a 3-m diameter and is described in Figure 1. The thick and auxiliary equipment
paste removal system is composed of two concentric helicoidal
screws of different diameters. It is installed at the bottom of Since red mud is shear thinning and thixotropic, its rheology is
the thickener. The system has the possibility to operate the affected by the amount of shear and by its history. By extracting
two screws simultaneously. Tests can also be run with the two high solids fraction mud without relying on a centrifugal pump,
screws inactive. The screw system removes the mud from the the thick paste removal system allows to obtain mud at high
bottom of the thickener and brings it to a holding box from where yield stress. Samples of red mud extracted with the system were
it is pumped, using a centrifugal pump, back to the underflow collected and yield stress measured.
pipe of the plant last washer. The mud feed to the pilot thickener 3.1 Removing high solids fraction mud
consists of two lines that take dense mud from the underflow
Tests to determine the ability of the screw to efficiently remove high
of the plant last washer and liquor from its overflow. The two
solids fraction mud were performed with a high mud bed height.
streams are mixed together through an inline mixer. The target
Feed mass flux was 6.7 t/m2/day. The set point of the underflow
feed solids fraction and feed flow rate are 10% and 34 m3/h.
rate was 2.2 m3/h. These conditions easily generated high solids
They are controlled by two automatic valves coupled with a flow
fraction mud at the underflow. In order to characterize the effect
meter and a density meter. The flocculant is pumped from the
of the screw, its rotation rate was progressively decreased once
plant by a positive displacement pump and is added directly to
equilibrium was reached. Data are presented in Figure 2.
the off-centred feed well of the pilot thickener at a dosage of
~60g/t. The underflow solids fraction is measured with a density For the first few hours, the system was run with stable conditions
meter. until equilibrium was reached. For these conditions, the underflow
rate, the screw velocity and the underflow pump velocity were
3. Results
all stable. The velocity of the screw was first increased at 7:30.
Two commonly encountered issues in thickener operation The underflow pump and the underflow rate were not affected.
are the removal of very high solids fraction mud and control The screw velocity was then progressively decreased to zero over
of the ratholing effect. For tested mud, the first one can be a period of 90 minutes. The underflow rate was not affected
observed when the solids fraction of the mud reaches around until ~8:30 when it started to decrease. This occurred at 18%
52%. Ratholing is characterized by low solids fraction mud at of the screw rate velocity. From this point on, the underflow

Proceedings of the 8th International Alumina Quality Workshop • 2008 17

AQW2008_Proceedings_FINAL-1-120.17 17 4/08/2008 2:54:54 AM

 rate decreased progressively. In an effort to keep the underflow
rate constant to the target set point, the pump velocity was
increased by the control system but the underflow rate continued
to decrease in function of the screw rate until it reached almost
0 m3/h when the velocity of the screw was 0%. The pump was
unable to remove the high solids fraction mud from the thickener
without the action of the screw. Throughout the test, the
underflow solids fraction remained constant at ~54%. From the
previous result, it is clear that the thick paste removal system
helps removing high solids fraction mud out of the pilot thickener.
The hypothesis explaining the action of the screw is based on the
amount of mud moved by the screw compared to the amount of
mud required by the underflow pump to achieve its target flow
rate. As long as the screw brings enough mud to the pump, the
target flow rate is achieved with the pump running at moderate
velocity. As the velocity of the screw is reduced, the pump is Figure 3. Underflow solids with the screw on and off
starving and increases its velocity to try to maintain the target
flow rate. However, since the mud is very viscous, the flow rate
cannot be obtained only with the suction of the pump.

Figure 4. Impact of the underflow rate on underflow solids fraction with the
screw off and on

Figure 2. Impact of the screw velocity on the underflow of the thickener 3.3 Mud rheology
Several samples of red mud were collected in an effort to
3.2 Control of ratholing measure their yield stresses. Figure 5 presents a close-up
The screw performance was evaluated under two conditions view of the underflow system. The locations of sampling points
that generate ratholing. For the first condition, operational are indicated in the figure. Sampling point 1 is located at the
parameters were kept constant and tests were carried out with outlet of the holding box just before the underflow centrifugal
the screw system both in operation and off operation. Tests were pump. Sampling point 2 is located at the end of the line, about
performed with a low mud bed height and a low mud residence 15 m downstream of the underflow pump. Yield stresses were
time. The feed mass flux was 24 t/m2/day. Part of the screw measured with a Haake VT550 rheometer and a vane FL100
was blinded to simulate operation of a shorter screw. The tests sensor. Measurements were done at room temperature. Data
were carried out during three days. All operational parameters are reported in Table 1. The removal of mud with the action of
were kept constant except the operation of the screw. During the screw occurs at a much lower shear rate than the pumping
the first and the third day, the screw was in operation. During of mud with a centrifugal pump. Hence, the yield stress of the
the second day, the screw was not in operation. Underflow solids mud taken before the pump is significantly higher than that of
fraction against time is presented in Figure 3. The average solids the samples taken after the pump. This behaviour is valid for
fractions obtained for the tests run with the screw in operation are moderate solids fraction and for higher solids fraction mud.
higher than for the test run with the screw stop. The difference is The thick paste removal system allows removal of mud out of a
approximately 6% and is significant. We can also observe that for thickener while keeping the yield stress significantly higher. This
the period where the screw was not in operation, the underflow can have a large impact in terms of the stacking capability of
solids fraction was significantly less stable. the mud.

For the second evaluation, the impact of a rapid change in the

underflow rate was studied for conditions with the screw in
operation and off operation. Tests were performed with a low
mud bed height. The feed mass flux was 6 t/m2/day. Part of
the screw was also blinded. After equilibrium was reached, the
underflow rate was quickly increased from 2 m3/h to 10 m3/h
for about two hours and brought back to 2 m3/h afterwards.
Results are presented in Figure 4. For both conditions, there is Figure 5. Sampling points at the underflow of the pilot thickener
a decrease in the underflow solids fraction when the underflow
rate is suddenly increased. The decrease is significantly larger
(-6.8%) for the condition when the screw is not in operation than
for the condition when the screw is in operation (-3%). This shows
a better control of ratholing with the operation of a screw.

18 Proceedings of the 8th International Alumina Quality Workshop • 2008

AQW2008_Proceedings_FINAL-1-120.18 18 4/08/2008 2:54:57 AM

 Table 1. Yield stress of mud at the underflow of the pilot 5. Conclusions
thickener
The thick paste removal system allows taking high solids fraction
Yield Stress Yield Stress mud out of a flat bottom thickener without relying on the suction of
Solids Fraction at Sampling at Sampling
Sampling Date a centrifugal pump. In doing so, it reduces any liquor channelling
Point 1 Point 2
to the underflow point and thus allows maintaining high underflow
% Pa Pa
solids. This advantage is also maintained at high underflow rate.
13 February 2008 50.5 158 45
The thick paste removal system allows production of mud at a
18 February 2008 53.0 247 95
very high yield stress. Continuous operation for ten days was
3.3 Longer term performance achieved and showed an improvement of about 4% solids on
the underflow solids fraction compared to the last washer of the
Longer term performance was measured for ten days and plant.
compared to the performance of the plant last washer. Both were
operating with the same mud, liquor and flocculant. The plant Our analysis shows that there are significant advantages to install
last washer is a deep cone washer. Normal plant constraints, a thick paste removal system in the plant last washer. A detailed
like volume control, availability of wash water, etc. were effective evaluation is currently underway to examine the potential of such
for the operation of the plant last washer. Mass fluxes were a system in other refineries.
10 t/m2/day for the plant washer and 19 t/m2/day for the pilot
thickener. Figure 6 presents the underflow solids fraction for
both thickeners. The average underflow solids fraction is higher
by 4% for the pilot thickener although it is operating at about
twice the mass flux compared to the plant washer.

4. Technical advantages
The impact of the installation of a thick paste removal system
in the last washer of the alumina refinery was studied. Solely
the impact of the increase in solids fraction at the underflow
of the washer was considered. Based on data in Figure 6, the
installation of a screw discharge underflow system would increase
the average underflow solids fraction from 48% to 52%. The
increase in underflow solids fraction translates to a significant
reduction in soda and alumina losses.

60

58

56

54
Solids fraction (%)

52 51.9%

50

48 47.8%

46

44
P ilot T hickener
42 P lant Last W asher
40
0 1 2 3 4 5 6 7 8 9 10 11 12
T im e (day)

Figure 6. Comparison of underflow solids for the pilot thickener and for the
plant last washer

The increase in the last washer underflow solids fraction also

corresponds to a reduction of the amount of liquor being sent to
the mud disposal site. This, in turn, reduces the amount of water
that needs to be evaporated in order to consolidate the mud. It
also reduces the amount of mud farming work that needs to be
done in order to speed up consolidation. The monetary impact
on the mud disposal site is also significant.

A mud with a higher solids fraction exhibits a higher yield

stress. This allows stacking of that mud at a steeper angle, thus
maximizing the usage of the mud disposal site.

References
Peloquin, G., Simard, G., Boivin, A., Dufour, R., Lemyre R. and Carruthers, A. 2005a, ‘Apparatus for and method of settling of mineral slurries’, US Patent
6936178.
Peloquin, G., Bui, R.T., Kocaefe, D. and Simard, G. 2005b, ‘Modélisation mathématique de la décantation de la boue rouge’, Canadian Journal of Chemical
Engineering, vol. 83, no. 3, pp. 458-465.
Peloquin, G., Dufour, R., Carruthers, A. and Puxley, D. 2007, ‘Apparatus for production, withdrawal and transfer of highly viscous slurry’, Patent application
WO 2007/022622A1.

Proceedings of the 8th International Alumina Quality Workshop • 2008 19

AQW2008_Proceedings_FINAL-1-120.19 19 4/08/2008 2:54:57 AM