New Laboratory Flotation Rig

7/27/2019 New Laboratory Flotation Rig

1/16

1

A NEW LABORATORY FLOTATION RIG FOR ORE FLOATABILITYDETERMINATION

S.M. Vianna*, O.N. Savassi**, J-P. Franzidis* and E.V. Manlapig*

In flotation, the property that is exploited to effect the separation of a mineral from an ore is referred to as

its floatability. Assessment and quantification of ore floatability is of fundamental importance in modelling

and simulation of flotation circuits.

This paper describes the design, construction and operation of a new laboratory flotation rig that has been

developed at the Julius Kruttschnitt Mineral Research Centre (JKMRC) to determine the floatability of

different ores on a small sale. The rig is designed for continuous operation.

The major features of the rig are highlighted in this paper. An unique characteristic of the rig is a 9.1 L

conditioning vessel. This provides a mixing regime close to plug-flow which avoids complications such as

different collector coverages for similarly sized particles (which would considerably increase the

complexity of ore floatability determination).

The main advantages of this rig when compared to conventional batch flotation cells include better control

of water recovery, froth recovery and chemical environment; implying a reduction in the variability of

results caused by human error. Preliminary tests undertaken with a high-grade lead/zinc ore from BHP

Cannington are presented demonstrating the good level of reproducibility obtainable in operating the

equipment.

* Julius Kruttschnitt Mineral Research Centre (JKMRC)-University of Queensland, Isles Road

Indooroopilly QLD 4068-Australia. Phone: (07) 3365 5893

** Departamento de Engenharia Metalurgica e Materiais-Universidade Federal de Minas Gerais (UFMG)-

Rua Espirito Santo no35, Belo Horizonte, Minas Gerais- Brazil. CEP: 35400


2/16

2

1. INTRODUCTION

As a part of the Australian Mineral Industries Research Association (AMIRA)-P9 project, a comprehensive

flotation model (equation 1) has been developed (Savassi, 1998), which relates the overall recovery (Ri,j,b)

of particles of a given size-mineralogical-liberation class in a flotation cell to the ore floatability (P i,j,b),

bubble surface area flux (Sb), froth recovery (Rf,i), water recovery (Rw), degree of entrainment (ENTi) and

residence time ():

Techniques are available or have been developed to measure or estimate Sb,Rf,i, ENTiand ; however in

order to apply equation 1 for simulation of flotation circuits, a method is required to either measure or

determine the ore floatability (Pi,j,b). Floatability refers to the ability of the mineral particle in a flotation

pulp to be removed by attachment to an air bubble (true flotation) and reflects the mineral properties and

the pulp chemical environment (Bradshaw, 1997). The floatability of an ore is expected to be dependent on

particle size, mineralogy, liberation characteristics, density and shape, as well as on pulp chemistry (pH,

Eh, dissolved oxygen, reagent type and concentration, ions in solution, etc).

To date, no one in the literature has demonstrated a means of calculating the ore floatability (P i,j,b)

theoretically from the properties of a feed stream in a flotation process. To tackle this problem, a new

flotation rig has been developed and constructed which incorporates special characteristics to isolate, study

and subsequently model the ore floatability. Such a flotation rig is envisaged to perform similarly to the

JKMRC Drop Weight Tester; assessing ore floatability instead of breakage, and providing Pi,j,bvalues for

different size-mineralogical-liberation classes and collector dosages of flotation streams. In this paper the

features of the new rig are described and the results of a few exploratory studies are presented.

2. DESCRIPTION OF THE RIG

A photograph of the new flotation test rig is shown in Figure 1. The rig consists of a 110 L stainless steel

sump fitted with a pneumatic Lightnin mixer with 2 impellers, a centrifugal pump to recirculate solids and

improve particle suspension, a 9.1 L transparent perspex conditioning vessel, a 3 L bottom driven perspex

cell, a peristaltic pump with a variable speed drive to deliver the slurry (maximum capacity of 6 L/min) to

the cell, two Watson Marlow peristaltic pumps for collector and frother addition (flowrates from 0 to 0.5

( )1.)1)(...1(

.)1(...

,,,

,,,,,

wiwifbbji

wiwifbbjibji

RENTRRSP

RENTRRSPR

++

+=


3/16

3

mL/min), and a Watson Marlow peristaltic pump (flowrate 0 to 2 L/min) in the sand gate to minimise

settling in the tailings launder,

The design of the 3 L flotation cell is similar to the high Sb flotation cell developed by Vera (Vera,

Franzidis and Manlapig, 1999); however this is not a high Sb cell. The 3 L flotation cell is a laboratory

scale flotation cell made of transparent perspex and mounted on a steel frame for heavy-duty continuous

operation. The cell is equipped with a bottom driven impeller and air injection is carried out through the

stator on the top of the impeller. The froth depth can be varied without changing the pulp volume using a

movable lip (Figure 2) allowing better control and estimation of froth recovery using the Feteris technique

(Feteris, Frew and Jowett, 1987). A pressure regulator and a pressure gauge are installed in the air inlet

pipe to the flow meter. Filters are also installed in the air inlet pipe to prevent dirt and oil from entering the

flow meter and the cell. The pulp level and tailings discharge are controlled by a fixed weir. A T-valve

placed prior to the flotation cell allows feed samples to be collected. One of the main advantages of the rig

is that it allows full stream sampling. Frother and collector are added on line prior to the flotation cell and

conditioning vessel, respectively.

An unique characteristic of the rig is the 9.1 L perspex conditioning vessel. The main reasons for having

the conditioning vessel are:

to prevent the fast floating material from reporting to the top of the sump and forming a thin froth layer

if collector were added directly in the sump;

to provide a constant and sufficient conditioning time for collector adsorption (preferable to collector

being added in line or into the sump).

The design of the conditioning vessel has to fulfil 3 basic requirements:

to keep the mineral particles in suspension;

to provide enough conditioning time for collector adsorption to reach an equilibrium, and, most

importantly;


4/16

4

to provide a mixing regime as close as possible to plug-flow to prevent the extent of conditioning of

mineral particles before flotation being continuously distributed. This latter would occur if mineral

particles remained in the vessel for different times, in which case the degree of adsorption of collector

onto the valuable mineral delivered to the flotation cell would vary from zero to full saturation,

increasing the complexity of the assessment of floatability considerably.

The above requirements are satisfied using the design shown in Figures 3 and 4. The basic idea behind the

design is to have a number of perfect mixers or compartments in series that together provide a mixing

regime close to plug-flow as demonstrated by Levenspiel (Levenspiel, 1972). To this end, the conditioning

vessel is a cylindrical perspex vessel with effective volume 9100 cm3. The design includes 10 round

nylon horizontal mixers spaced equidistantly along a stainless steel shaft separated by 9 perspex circular

baffles. The circular baffles can be removed if required, changing the mixing regime in the conditioning

vessel from close to plug-flow to perfect mixing. A circular hole in the centre of the baffles allows the

slurry that is fed at the top of the vessel to pass through the mixing chambers in a zigzag pattern. O-rings

are placed between the compartments to avoid any leakage (Figures 3 and 4). A one way valve at the top

of the vessel is used to eliminate most of air entrapped in the slurry prior to a flotation test. A digital

tachometer installed on the top of the vessel (Figure 1) allows the control of the level of agitation (mixer-

speed) in the vessel. The conditioning vessel is mounted on a stainless frame to minimise vibration (Figure

1).

3. EXPERIMENTAL

3.1. Residence time studies

The simplest method of determining residence time distribution (RTD) is by the stimulus-response

technique which involves introducing a tracer to the vessel feed stream at time zero and either periodically

sampling or continuously monitoring the tracer concentration in the vessel discharge. The sampling period

is normally 3 to 4 times the mean residence time.

Residence time distribution studies of the flotation cell and the conditioning vessel were undertaken using a

highly concentrated solution (65% by weight) of lithium chloride as a tracer to estimate the mean residence

time and to determine the mixing regime. The results of these studies are shown in Figures 5 and 6.


5/16

5

From the experimental data, the mean residence time can be estimated using equation 2.

where ti is a discrete time value, Ciis the discrete normalised tracer concentration value at time t iand tiis

a time interval.

The mean residence time can also be calculated using equation 3.

where V is the effective volume of the vessel excluding the mechanism, gis the air hold-up by volume and

F is the feed volumetric flow rate.

The residence time distribution presented in Figure 5 (time against concentration of lithium chloride) for

the flotation cell is a typical distribution of a perfect mixing regime (Levenspiel, 1972), as expected. The

mean residence time of the cell calculated using equation 3 is close to the estimated value using the liquid

tracer.

As mentioned above, the conditioning vessel should provide a mixing regime close to plug-flow. Several

different designs were tried before the one described in the previous section. Figure 6 presents the results

of the residence time studies in the conditioning vessel. The graph is a typical distribution of a plug-flow

mixing regime( Levenspiel, 1972); therefore the mixing regime in the conditioning vessel is in fact close to

plug-flow. The calculated value of the residence time using the feed flowrate and vessel volume is very

close to the value obtained using lithium chloride. It should be noted that the residence time study was

carried out in the liquid phase and would closely approximate the behaviour of the solids phase only at very

fine sizes (below 38 m). Complementary residence time studies using a solid tracer (such as cassiterite)

will be undertaken to verify the mixing regime for coarse solids in the vessel.

)2(

=

ii

iii

tCtCt

)3()1(

F

V g

=


6/16

6

3.2. Continuous flotation tests

Six replicate continuous flotation tests in open circuit were undertaken to check the reproducibility of the

rig. The condition selected for the tests (starvation addition of collector) was specially chosen as being the

worst case scenario in terms of reproducibility (because of the very low concentrate flow rate).

Before each flotation test, 21 kg of a high grade lead-zinc ore (19.6 % Pb, 16.1 % Zn, 38 % SiO2 and 4.7 %

Fe) from the BHP-Cannington mine were stage crushed to below 3.35 mm in a roll crusher and wet ground

at 70 % solids to a target passing size (D80 of 155 m) using a stainless steel rod mill. The mill was

purged with high purity nitrogen at 14 L/min for 30 minutes prior to grinding to minimise oxidation in the

mill. After grinding the slurry was transferred to the sump and the percent solids was adjusted to 17.5 %

using tap water. Temperature and pH were monitored manually during the flotation tests with an ATC

probe and a glass electrode connected to an Orion SA-720 instrument. The redox potential (Eh) was

measured using a portable Metrohm instrument combined with a platinum electrode which was checked

prior to each experiment with a Fe3+/Fe2+ solution (Light, 1972). The reagents used in the tests were

diisobutyldithiophosphinate (50 % solution w/w) supplied by Cyanamid (reagent 3418A) and a

polypropylene glycol (DowFroth 250/technical grade) supplied by Cytec. The reagent solutions were

prepared daily prior to the flotation experiments.

A large number of preliminary tests were carried out to establish the suitable operating conditions, reagent

levels and test procedures. Fourteen minutes were required to reach steady state. The operating criteria for

steady state are constant physical and chemical conditions in the flotation feed, in the cell and in the tailings

discharge. This requires that the pulp level and flow rates to and from the cell are kept constant, that

reagents including air are delivered at a constant rate and that the impeller speed is held constant as well.

After 14 minutes from start-up, a tailings sample was taken and then, sequentially, concentrate and feed

samples with 10 second intervals between sampling the different streams.

Table 1 summarises the results which demonstrate good reproducibility of the rig under extreme conditions.

At nominal steady state, the replicate flotation tests showed that the rig possessed a 2% maximum

fluctuation in the feed flow rate and 6% in the percent solids. The values of mass recovery and waterrecovery at 95 % confidence intervals also suggest that the rig is performing well.

An added feature of this type of rig is that when compared to conventional batch flotation cells, it allows a

better control of water recovery and estimation of froth recovery using the Feteris technique (Feteris, Frew

and Jowett, 1987) which is very difficult in batch tests due to the non-steady state nature of the process.


7/16

7

Poor control of water recovery and the need to use shallow froth depths to minimise the effect of the froth

during batch flotation tests impact on the concentrate grade due to entrainment.

4. RESEARCH POTENTIAL

The main purpose of the rig is to isolate and study the effect of particle size, mineralogy, liberation and

density of adsorption of collector on the ore floatability P i,j,b. However, the rig can be used in other types of

projects such as:

entrainment studies using different types of minerals to investigate the effect of density of minerals on

the degree of entrainment; which could be incorporated in an entrainment model.

reagent testing and adsorption studies to investigate the effect of different types of reagents on ore

floatability on a size-mineralogical-liberation basis. These studies could be complemented using

sophisticated surface spectroscopic techniques such as X-ray photoelectron spectroscopy (XPS or

ESCA) and Time of Flight Secondary Ion Mass Spectroscopy (TOF-SIMS) to determine the

distribution of reagents, especially collectors, on the surface of the minerals. This would "close the

gap" by identifying and quantifying the actual surface species impacting on the particle's

hydrophobicity and hydrophilicity.

residence time studies using solid or liquid tracers to assess the impact of the mean residence time and

the mixing regime on the floatability Pi,j,b. The changeable configuration of the baffles in the

conditioning vessel allows these type of studies.


8/16

8

5. CONCLUSIONS

A new flotation laboratory test rig has been developed and built at the JKMRC. The main use of the rig

will be to assess the floatability of ores. The basic idea behind the development of the rig is to be able to

characterise and subsequently model the floatability of different ores on a small scale. The rig is expected

to perform similarly to the JKMRC Drop Weight Tester; assessing ore floatability instead of breakage.

Preliminary reproducibility tests suggest that the rig performs well under extreme conditions. The

flexibility of the rig design, in particular the conditioning vessel, presents new possibilities for research.

6. ACKNOWLEDGEMENTS

The authors would like to express their gratitude to Peter Runge who built the flotation cell, and JKMRC

pilot plant staff specially Bob Marshall, Mike Kilmartin, Andre Metzler and John Worth for their help

during the construction and commissioning of the rig. The authors would also like to thank Xiaofeng

Zheng for his assistance during the residence time studies. Finally, the authors would like to acknowledge

the AMIRA P9 project and its sponsors for funding which made this testwork possible.


9/16

9

7. REFERENCES

Bradshaw, D., 1997. Synergistic effects between thiol collectors used in flotation of pyrite, Ph.D. thesis,

University of Capetown, South Africa.

Feteris, S.M., Frew, J.A. and Jowett, A., 1987. Modelling the effect of froth depth in flotation,

International Journal of Mineral Processing, 20: 121-135.

Levenspiel, O., 1972. Chemical Reaction Engineering, John Wiley and Sons, Chapter 9, pp. 253-325.

Light, T.S., 1972. Standard solution for redox potential measurement. Analytical Chemistry,44 (6): 1038-

1039.

Savassi, O.N., 1998. Direct estimation of the degree of entrainment and the froth recovery of attached

particles in industrial flotation cells, Ph.D. thesis, JKMRC, University of Queensland, Australia.

Vera, M.A., Franzidis, J-P and Manlapig, E.V, (1999). The JKMRC high bubble surface area flux

flotation cell,Minerals Engineering, 12(5), pp. 477-484.


10/16

10

Figure 1- Flotation Test Rig

Sump

Feed Pump

Tachometer3.0- Litre FlotationCell

9.1-Litre Conditioning Vessel

Purge Pump

Pneumatic Mixer

T-valve

FrotherPump CollectorPump


11/16

11

Figure 2- Concentrate Launder (front view)

Figure 2- Flotation cell concentrate launder (front view)

Moveable Lip


12/16

12

Figure 3- 9.1 -Litre Conditioning Vessel

One way valve

(air discharge)

Rounded

Nylon Mixer

Circular Baffles

Pneumatic Mixer

Compartments


13/16

13

Figure 4- Conditioning Vessel Components

Compartments or

Chambers

Rubber SealCircular Baffles

Nylon MixersStainless

Steel Shaft


14/16

14

Figure 5- Residence Time Distribution in the 3-Litre Flotation Cell

0

50

100

150

200

250

300

0 50 100 150 200 250 300 350Time (s)

Concentration(ppm

)

Cell Effective Volume: 3.05 litre

Mean Residence Time Estimated using the Tracer: 1.61 min

Mean Residence Time Calculated using equation 3: 1.50 min

Air Hold-up: 8%


15/16

15

Figure 6-Residence Time Distribution in the Conditioning Vessel

0

50

100

150

200

250

300

350

0 200 400 600 800 1000 1200

Time (s)

Concentration(ppm)

Vessel Effective Volume: 9.1 litre

Mean Residence Time Estimated using the Tracer: 4.89 min

Mean Residence Time Calculated using equation 3: 4.87 min

Air hold-up: negligible


16/16

16

Table 1-Operating Conditions and Confidence Levels for the Replicate Flotation

Tests in Open Circuit

Variable Unit TargetLower

Limit

Best

Estimate

Upper

Limit

Confidence

Level

FeedFlow Rate

g/min 2143.7 2130.8 2147.4 2163.9 99 %

Percent of

Solids% 17.5 17.2 17.7 18.2 95 %

Massic

Recovery% no 9.3 10.6 11.9 95 %

Water

Recovery% no 1.7 1.9 2.2 95 %

ImpellerSpeed in the

Cell

rev/min 1150 1148 1153 1160 95 %

Impeller

Speed in the

ConditioningVessel

rev/min 220 218 225 240 95 %

pH - Alkaline 8.3 8.5 8.6 95 %

New Laboratory Flotation Rig

Documents

Transcript of New Laboratory Flotation Rig