Abstract—The study provides optimization data on the work

carried out on the Knelson Concentrator KC XD-20 at the High Pressure Leach (HPL) plant at Kansanshi Mining PLC (KMP),

Zambia. The concentrate sent to HPL plant contains approximately 8 g/t of gold. This study involved the optimization of various parameters and performance determination of KC XD-20. The parameters investigated include G-force, fluidization water flowrate, cycle time and particle size distribution. The feed flow rate and pulp density were kept constant. A series of optimization tests were carried out at KMP. The results of the tests showed that there was a significant increase in free gold recovery. The optimization tests

indicated an improvement of 12 – 25% of gold recovery under the conditions investigated. The optimum parameters obtained were 150 G, fluidization water flowrate of 13m3/h and cycle time of 30 min.

Keywords—Gold, Gravity recovery, Knelson Concentrator,

Optimization.

I. INTRODUCTION

A. Kansanshi Mining PLC

ANSANSHI Mining PLC (KMP), a subsidiary of First

Quantum, owns 80% and operates the copper and gold

mines while ZCCM‘S subsidiary owns the remaining

20%. Gold at KMP copper and gold mine exits in two

different forms namely gravity gold (nuggets) and refractory

gold, locked up in the pyrite and chalcopyrite matrix. The nugget gold is collected by the use of falcons with feed

coming directly from the mills. Gravity Gold (Nuggets) is

recovered from course material before any further processing

of the ore. The refractory gold is collected using Knelson

concentrators at the high pressure leach (HPL) plant after the

autoclave leaching liberates it from its complex lock up [1].

J. Siame is with the Copperbelt University, School of Mines & Mineral

Sciences, Department of Chemical Engineering, P.O. Box 21962, Kitwe,

Zambia. (Phone: +260 96 920 6022; fax: +260 21 222 8212; E-mail:

john.siame@cbu.ac.zm).

K. Muchima was with the Copperbelt University, School of Mines &

Mineral Sciences, Department of Chemical Engineering, P.O. Box 21962,

Kitwe, Zambia. (E-mail: kashelamuchima@gmail.com).

D. Chirwa is with Kansanshi Mining PLC, High Pressure Leaching Plant

(HPL) Solwezi, Zambia.

P. C. Magawa was with Kansanshi Mining PLC, High Pressure Leaching

Plant (HPL) Solwezi, Zambia.

B. Knelson Concentrator and its Basic Separation

Mechanisms

The first unit of batch Knelson Concentrator was introduced

into the mineral processing industry in 1978. Nearly 3 decades

of development and modification, different series Knelson

Concentrators for different application purposes have been

well manufactured and used in precious metals recovery

industries all over the world [2], [3], [4], [5]. The following is

a summary of the Knelson concentrator XD-20 presently used

in industry [5].

B.1. Extended duty series (XD)

The XD series, with its futures of compact design, stainless

steel construction and high quality components, is for hard

rock milling circuits to withstand severe operating conditions.

These features of the design have made the XD series the most

demanding application for precious metals recovery industry.

Centrifugal fields can be generated in two different ways.

Firstly, a fluid is introduced at a high tangential velocity into a

cylinder or conical vessel such as a hydro cyclone. Generally, the larger and heavier particles will be collected near the wall

of the separator, while the smaller and lighter ones will be

taken off through an outlet near the axis of the vessel. The

second way is the use of a centrifuge. In this case, a fluid is

introduced into a rotating bowl and it is rapidly accelerated.

All the fluid tends to rotate at a constant angular velocity,

and a forced vortex is established. The tangential velocity is

directly proportional to the radius at which the fluid is rotating

[6].

In most practical cases, when a particle is moving in a fluid

under a centrifugal field, gravitational effects will be comparatively small, and can therefore be neglected. The

equation for the particles in the centrifugal field will be similar

to that for the motion in the gravitational field, except that the

gravitational acceleration ‗g‘ must be replaced by the

centrifugal acceleration [6]. The centrifugal acceleration is given by:

(1)

The main forces acting on a particle inside the Knelson

Concentrator are centrifugal force and axial drag force. For the

drag force, it is assumed that under laminar flow conditions,

the drag force on a spherical particle was entirely due to viscous forces within the fluid (Stock's equation), and can be

described as [7], [8]:

Optimization of Gravity Recovery of Gold at

High Pressure Leach Plant of Kansanshi Mining

PLC, Zambia

J. Siame, K. Muchima, D. Chirwa, and P. C. Magawa

K

Intl' Conf. on Chemical, Integrated Waste Management & Environmental Engineering (ICCIWEE'2014) April 15-16, 2014 Johannesburg

165

(2)

where Fd: inward drag force, g.cm.s-2; Vr: velocity at radial distance r, cm.s-1; μ: the viscosity of the fluid medium, 0.01

g.cm-1.s-1 for water at 20 °C.

If the effects of the fluidization water and other forces on

the particle are not considered, the particle reaches its terminal

settling velocity when Fc equals Fd. Because we mainly

consider the behaviour of fine particles, Stokes equation could

be used to approximate the terminal settling velocity in a

centrifugal field by substituting for ‗g‘, as shown in Eq. (3) [6], [9].

( )

(3)

Why, in the centrifugal force field, can very fine particles

be more effectively separated, compared to the gravity field?

Eq. (4), answers this question, that is, as the centrifugal

acceleration increases, the size of the critical particle (the

finest particle that can be recovered) decreases [9].

√

(4)

Where, K, m, and c are coefficients, H is the thickness of

fluid film, and R is the average radius of the rotation drum.

From this equation, it is also easy to understand why the

Falcon Concentrator (or Super Bowl) can effectively recover

even finer particles than the Knelson Concentrator does, since

their centrifugal acceleration can reach 200 gs, compared to 60

gs of a regular Knelson Concentrator [7], [10].

The recovery of gold was calculated using the following

equation:

( )

( ) (5)

II. METHODOLOGY

The Knelson Concentrator XD-20 was used for the

optimization test works at KMP. The operating variables such

as G-Force, Fluidization water flow and cycle time were

selected using Knelson proprietary ―Independent Control

System‖ (ICS). The pilot test work was carried out over two

campaigns. A sample of 500g was collected on each campaign

per day from the sampling points as follows: Knelson

Concentrator 1 feed and Knelson Concentrator 1 tails; Knelson Concentrator 2 feed and Knelson Concentrator 2 tails

as shown in Figure 1, (a) and (b).

Samples from the two streams were drawn and filtered

using a pressure filter and then dried in the oven at 60 °C for 2

hours. The dried samples were then pulverized. Individual

samples of 50 g each were sent for chemical analysis. Gold

analysis was performed by KMP in Solwezi, Zambia. Fire

analysis of gold solid samples was also carried out. The tests

were repeated in triplicates.

For the particle size analysis, 2 kg sample was taken from

three sampling points: Autoclave discharge point (before the

thickener), Knelson Concentrator 1 feed, Knelson Concentrator 2 tails. The pulverized sample was then split into

1 kg batch each using the riffle splitter. Each portion of each

sample was then put on the 150 µm sieve with the 106, 75, 53

and 25 µm sieves arranged in descending order then placed on

a sieve shaker. The retained sample on each screen size was

then weighed and sent for gold analysis.

(a)

(b)

Fig. 1. Schematic flow sheet of the Knelson Concentrators XD-20: (a) Series configuration, (b) Parallel configuration.

TABLE I

OPERATING CONDITIONS OF THE KNELSON CONCENTRATORS

G-Force Cycle time (min) Fluidization water flowrate (m3/h)

100 30 -60 9 and 13

150 30-60 9 and 13

200 30-60 9 and 13

III. RESULTS AND DISCUSSION

A. Effect of G-Force

According to Cherest [11] the recovery is directly

proportion to the G-force. Increased G-force was found to be beneficial for the recovery. Figure 2 shows the effect of G-

force on Knelson Concentrator XD-20 unit recovery.

Intl' Conf. on Chemical, Integrated Waste Management & Environmental Engineering (ICCIWEE'2014) April 15-16, 2014 Johannesburg

166

Fig. 2. Effect of G-Force on Gold recovery.

B. Effect of cycle time

In general, for Knelson Concentrator application in a

secondary circuit or in other circuits with high levels of pyrite

in the ore, a phenomenon known as ―concentrate bed erosion‖

can occur, whereby gold already recovered on the surface of

the concentrating ring is removed due to the effect of high

specific gravity particles such as pyrite. As such, short

concentrating cycle times are generally of benefit. The test

work carried out proved this to be the case in this application.

Figure 3 shows more stability with the minimum cycle time

giving an experimentally accurate peak for all cycle time (30,

45 and 60 min) investigated. Initially, 60 minute cycles were used, but the results suggested shorter cycles would be of

benefit without causing significant dilution of the final copper

concentrate. Thus 30 minute cycles were adopted. Figure 3

shows the differences in unit GRG recovery between the 30

and 60 minute concentrating times.

Fig. 3. Effect of recovery related to cycle time.

C. Effect of fluidization water flowrate

The effect of using different fluidization water flowrate values on the gold recovery was investigated. Laplante et al.

[12] reported that the effect of fluidization water pressure was

minimal in their findings. Optimum fluidization occurs when

the force of the slurry against the cone is comparable to that of

the fluidization water flow rate.

Figure 4 shows gold recovery for two tests, one at low

fluidization water flow rate and another at high water flow rate

to understand the effect of fluidization water flow rate on gold

recovery. The tests conducted at low fluidization flow rate

(9m3/h) produced low recoveries in all sizes with an overall

recovery of 64% whereas the test performed at a high fluidization flow rate (13m3/h) produced high size by size

recoveries and an overall recovery of 71%. The low recoveries

are attributed to the combined effect of coarse size, and low

fluidization flow rate, 9m3/h, which act synergistically to

erode the concentrate collected in the riffles. Typically for an

operating Knelson, low recoveries have been reported due to

the suboptimal fluidization water flow rates [13].

However, the recovery trend is consistent for both high and

low fluidization water flow rates. However, as fluidization

water flows are increased beyond this range, a loss of recovery

of fine GRG is incurred.

Fig. 4. Effect of fluidization water flow rate on gold recovery.

D. Particle Size Analysis Table II (see appendix) shows the amount of gold contained

on each screen size. It is expected that a higher grade of gold

would be on the smaller size fraction as more gold would have

been liberated and there would be less gangue material.

Fig. 5. Cumulative % passing and retained as related to screen size

for the series test work.

0

20

40

60

80

1 2 3 4 5 6

Recovery %

Test runs

Recovery for 100G

Recovery for 150G

Recovery for 200G

0

10

20

30

40

50

60

70

80

1 2 3 4 5 6

Recovery

%

Test runs

recovery30

recovery45

recovery60

55

60

65

70

75

1 2 3 4 5 6

Recovery

%

Run number

recovery for 9recovery for 13

Intl' Conf. on Chemical, Integrated Waste Management & Environmental Engineering (ICCIWEE'2014) April 15-16, 2014 Johannesburg

167

Furthermore, the results showed that gold grade was higher

on the 53µ screen and highest on the 25µ screen. The least

gold grade was found to be on the 106µ screen followed by

the +150µ screen. In Figure 5, the cumulative passing and

retained indicated that bulk of the feed material was a bigger

size, falling in the range of 150µm and above.

According to Grewal and Fullam [14], it is expected that that

the finer size particles would have the most gold losses.

However, optimization test work showed that the most gold

losses were in the courser size fractions (+53 µm). This could

have been due to gangue material being mostly in the courser

size fraction and less in the finer sizes.

IV. CONCLUSION

The objective of the work described in this paper was to

determine the optimum parameters and investigate the effects

of G-force, fluidization water flowrate, cycle time and particle

size distribution on the recovery of very fine gold in leach

tailings.

The optimum recovery of gold was achieved at different

conditions. The optimization results indicated an improvement

of 12-25% of gold recovery and this was achieved at 150G-

force, 13m3/h fluidization water flowrate and a 30min cycle

time. Replicate tests suggest that the G-force giving greater

than 150G-force do not offer additional gold recovery for

these feed materials.

The particle size distribution was found to be a contributing

factor to the overall yield. This was mainly due to either

interstitial trickling or composition of gangue material in the

overall courser materials. The recovery was best in the -75µm

size fraction. Extending the size analysis to -25µm

demonstrated that the Knelson Concentrator is concentrating

gold bearing silicate particles down to 25 µm. However,

particles finer than 25 µm are beginning to follow the water

flow into the tailings.

APPENDIX

TABLE II AMOUNT OF GOLD CONTAINED ON EACH SCREEN SIZE.

Screen size (µm) 150 106 75 53 25

Sample collection point

Weight %retained 49 24.5 23 2.5 1

Autoclave Discharge Cumulative % retained 49 73.5 96.5 99 100

Gold % contained 28 12.6 14.6 22.7 21.6

Weight % retained 76 17 2.5 3.2 1

Knelson Feed Cumulative % retained 76 93 95.5 99 100

Gold % contained 40.77 7.7 4.34 26.844 20.26

Weight % retained 52.5 25.5 13 6.5 2.5

Knelson Tails Cumulative % retained 52.5 78 91 97.5 100

Gold % contained 49.95 27.46 11.29 6.67 4.74

ACKNOWLEDGMENT

The authors would like to thank the Kansanshi Mining Plc, Solwezi, Zambia and the Copperbelt University, Kitwe,

Zambia.

REFERENCES

[1] Press release, ―Kansanshi fact, last update May,‖ Solwezi, 2007.

[2] B .V. Knelson, ―Centrifugal concentration and separation of precious

metals,‖ 2nd International Conference on Gold Mining, Vancouver, pp.

303-317, 1988.

[3] B. V. Knelson, ―Development and economic application of Knelson

concentrator in low grade alluvial gold deposits,‖ The AusIMIM Annual

Conference, Rotorua, New Zealand, March 18-21, pp. 123-128, 1990.

[4] L. Huang, ―Upgrading of Gold Gravity Concentrates. A Study of the

Knelson Concentrator. Ph.D. Thesis, pp. 294, 1996.

[5] B. V. Knelson, ―Batch Knelson Concentrator,‖ (Booklet), pp. 6, 2004.

[6] J. M. Coulson, J. F. Richardson, J. R. Buckhurst, and J. H. Harker,

―Chemical Engineering, vol. 2, 4th Ed., Particle Technology and

Separation Processes,‖ Pergamon Press, Oxford, New York, pp. 968,

1991.

[7] J. Sutter, ―Super Bowl nominal specifications‖, 1997 [Online]

http://www.rmsross.com/sb_spec.htm (Accessed on 13/10/2013).

[8] G. G. Stokes, ―Mathematical and Physical Papers‖. Trans. Cambridge

Phil. Soc., 9, Part II, 51ff, 1901.

[9] Y. Sun, ―Gravity Separation,‖ Metallurgy Press, Beijing, pp. 310, 1982.

[10] S. McAlister, ―Case study in the use of the Falcon gravity concentrator,‖

Proceedings of 24th Annual Meeting of the Canadian Mineral

Processors, Ottawa January, pp.20, 1992.

[11] A. Cherest, ―Performance of Knelson XD-20 concentrator at golden

giant mine under different operating conditions‖, Research PROJECT

306-410B, McGill University, 2000.

[12] A. R. Laplante, Y. Shu, and Marois, ―Experimental characterization of a

laboratory centrifugal separator,‖ Canadian Metallurgical Quarterly 35

(1), pp. 23–29, 1996.

[13] F. A. Vincent, ―A comparison of Knelson Concentrator and jig

performance for gravity recovery,‖ M. Eng. Thesis, McGill University,

pp. 131, 1997.

[14] I. Grewal, and M. Fullam, ―Gravity circuit optimization via

mathematical modeling by particle size classes,‖ [Online]. Available from www.knelsongravitysolutions.com [accessed January 2013].

Intl' Conf. on Chemical, Integrated Waste Management & Environmental Engineering (ICCIWEE'2014) April 15-16, 2014 Johannesburg

168