International Journal of Mineral Processing 129 (2014) 22–26

Contents lists available at ScienceDirect

International Journal of Mineral Processing

j ourna l homepage: www.e lsev ie r .com/ locate / i jm inpro

Effect of wash water on the mineral size recovery curvesin a spiral concentrator used for iron ore processing

Maryam Sadeghi a, Claude Bazin a,⁎, Marilène Renaud b

a Engineering Department of Mining, Metallurgy and Materials, Laval University, Québec, QC, Canadab COREM, Québec, QC, Canada

⁎ Corresponding author at: Engineering Department of MLavaUniversity, Québec, QCG1V0A6, Canada. Tel.:+1418 6

E-mail address: Claude.bazin@gmn.ulaval.ca (C. Bazin

http://dx.doi.org/10.1016/j.minpro.2014.04.0060301-7516/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 February 2014Received in revised form 19 March 2014Accepted 24 April 2014Available online 4 May 2014

Keywords:Spiral concentratorsPartition curvesSize recovery curvesWash waterGravity concentrators

Spiral concentrators are widely used in the iron ore industries to concentrate heavy iron oxides from light silicagangue minerals. The operation of a WW-6 spiral for the concentration of an iron oxide ore is analyzed throughthe size recovery or partition curves of the minerals. Several tests conducted following a factorial design usingwash water addition, feed rate and slurry solid concentration as studied factors show that for the tested conditionsthewashwater addition has themost important effect on the spiral performancewith its effect beingmainly locatedon coarse particles.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Spiral concentrators are gravity based separation devices used for thepreparation of coal, iron and heavy mineral ores (Wills, 1992). A spiralcan be described as a centrifugal sluice in which the centrifugal forcecombines with gravity to operate the separation between light andheavy particles. The separation of particles in a spiral is also impactedby the drag force and the lesser known Bagnold force which specificallyacts preferentially upon particle size rather than weight (Atasoy andSpottiswood, 1995; Burt, 1984; Bouchard, 2001). For a given ore feedsize distribution the main operating variables for a spiral concentratorare the feed rate, the slurry concentration, the wash water addition andthe position of the cutters or splitters used to separate the tailings,middlings and concentrate streams.

Dallaire et al. (1978) investigated the effect of the feed rate andslurry solid concentration on a Humphrey spiral operation for iron ore.These authors reported that increasing the feed rate is favorable toconcentrate grade but hampers valuable species recovery and thatincreasing the feed slurry solid concentration should increase recoveryat the expense of concentrate grade. Few authors discussed quantita-tively the effect of wash water addition on the operation of spirals.The discussion indicates that wash water is an important control vari-able used to wash away entrapped light minerals from the concentratestream (Burt, 1984; Bouchard, 2001). Richards and Palmer (1997)

ining, Metallurgy andMaterials,56 5914; fax:+1418 656 5343.).

indicate that significant benefits were observed after the removal ofwashwater from spirals treating light minerals. These studies aremain-ly based on global performance indices such as grade and recovery butdid not investigate the effect of spiral operating variables on the recov-ery of particles as a function of their size and specific gravity.

This paper uses themineral size recovery curves to investigate, usinga 23 factorial design, the effect of the slurry solid concentration, slurryfeed rate and wash water addition on a WW-6 spiral processing aniron ore. The first section describes the set-up used for the tests andthe test results are analyzed in the second section.

2. Test set up and experimental conditions

The tests are conducted using a spiral operated in a close circuit atthe COREM pilot plant in Quebec, QC, Canada. The set up used for thetests is described first followed by the test conditions.

2.1. Test set-up

The spiral used for the tests is a 7 turn WW-6E from Roche Mining.The set-up used for the experiments is shown in Fig. 1. The feedmaterialis a ground ore from the Mount Wright Arcelor-Mittal mine in Quebec(Hyma and Meech, 1989). The ore is mixed with water and the slurryis loaded into pump box A (see Fig. 1). The slurry is pumped to a distrib-utor at the top of the spiral. The distributor is connected to 12 rubberpipes that can be easily moved in and out of the spiral feed wells. Thisset up allows a fast and accurate adjustment of the spiral feed rate.The slurry that is not used as the spiral feed is returned to pump box

Wash watertank

Dewateringhydrocyclone

Pump box A

Sample

12 pipesplitter

Pumpbox B

Pumpbox C

7 turn WW-6 spiral

Fig. 1. Experimental set-up.

23M. Sadeghi et al. / International Journal of Mineral Processing 129 (2014) 22–26

A. One pipe is used to sample the spiral feed (see Fig. 1). The spiral con-centrate and reject streams discharge into pump box A via soft rubberpipes that can be easily manipulated to collect samples and measureflow rates. Pump box A overflows to pump box B that feeds adewatering hydrocyclone. The hydrocyclone underflow is returnedinto pump box A while the overflow is pumped back, via pump C, to awater tank that provides the wash water addition to the spiral. Amanual valve (Fig. 1) is used to adjust the flow of wash water sent tothe spiral that is fairly constant providing that the level in the watertank is kept constant during the test work.

2.2. Experimental design

The tests conducted follow a 23 factorial design with the feed rate,feed slurry concentration and flow rate of wash water. The openingsof the concentrate cutters (splitters) are set at the beginning of thetest work and kept constant for all the tests. The 23 factorial designused for the tests is shown in Table 1 with the considered variation forthe factors. Tests 1 and 10 are carried out at the reference conditions.

Table 1The 23 factorial design used for the tests on the spiral.

Test Feed rate % Solids Wash water

F (l/min) S (% w/w) W (l/min)

1 60 (0) 37 (0) 16.7 (0)2 48 (−1) 32 (−1) 8.3 (−1)3 72 (+1) 32 (−1) 8.3 (−1)4 48 (−1) 42 (+1) 8.3 (−1)5 72 (+1) 42 (+1) 8.3 (−1)6 48 (−1) 32 (−1) 25.0 (+1)7 72 (+1) 32 (−1) 25.0 (+1)8 48 (−1) 42 (+1) 25.0 (+1)9 72 (+1) 42 (+1) 25.0 (+1)10 60 (0) 37 (0) 16.7 (0)

The amplitudes of the changes in the feed solid concentration and volu-metric feed rate are chosen so that the test conditions fall within themanufacturer's recommended values reproduced in Fig. 2. The spiralmanufacturer (Roche Mining, 2011) recommends a wash water addi-tion within 0.5 and 1.5 m3/h which corresponds to the levels testedfor that factor.

For each test the feed rate, slurry solid concentration andwashwaterare adjusted to the planned values. The spiral is then allowed to stabilizefor 10 min with regular measurements of the feed rate and slurry solidconcentration to ensure a stable operation. A timed sample of the feedstream is taken first to avoid disturbing the spiral operation. Thentimed samples of the spiral concentrate and reject streams are takensimultaneously. The samples are used to obtain:

the mass flow rate;the head assays (Fe, SiO2, Al2O3, CaO, MgO, Na2O, K2O, MnO);the size distribution (from 1.6 to 0.038mm) of the particles: a size of0.025mm is used as themean size of the particles passing through a0.038 mm sieve;the chemical composition (Fe, SiO2, Al2O3, CaO, MgO, Na2O, K2O,MnO) of all the size fractions;the solid concentration of the slurry;the specific gravity of the ore.

2.3. Data processing and estimation of the size recovery curves

All measurements are reconciled using the BILMAT algorithm(Hodouin and Everell, 1980). The estimation of the mineral content ofthe streams from the chemical assays is not yet incorporated into thedata reconciliation. However since iron oxides, mainly hematite, andquartz account for more than 95% of the considered ore (Hyma andMeech, 1989) the tracking of the behaviors of iron and silica (SiO2) isdeemed acceptable in a first approximation. A detailed analysis of thespiral operation would however require the estimation of the mineralcontents of each stream and size fraction as the spiral classifies theparticles as a function of their size and specific gravity the latter beingdependent of the mineral composition.

The recovery in the concentrate stream of a species m within sizeinterval i is noted as Ri;m and is calculated using:

Ri;m ¼ WCgC;ixC;i;mW FgF;ixF;i;m

: ð1Þ

The variable W stands for a solid flow rate, while g is used for theweight retained within the ith size interval and x is the species m con-centration of the ore within size interval i. The subscript m is used toindicate a species such as iron or silica. The indices F and C indicaterespectively the feed and concentrate streams. The use ofmass balanceddata in Eq. (1) filters out the measurement errors and provides morereliable performance indices than the direct use of measurements into

Testzone

Fig. 2. Recommended operating range for the WW6E spiral (Roche Mining, 2011).

Table 2Reconciled results for the size distribution and iron and silica contents for test 1.

Size interval (mm) Feed Concentrate Reject

% ret. % Fe % SiO2 % ret. % Fe % SiO2 % ret. % Fe % SiO2

Head assays 33.75 50.55 60.39 12.81 11.67 81.82+1.180 2.97 45.98 31.93 3.61 68.34 2.20 2.44 18.58 68.361.180/0.850 9.42 48.16 30.00 11.66 69.14 1.20 7.57 21.41 66.720.850/0.600 9.11 44.31 35.61 9.68 69.16 1.10 8.63 21.24 67.660.600/0.425 11.73 38.42 44.84 11.25 69.07 1.20 12.12 14.85 78.400.425/0.300 15.96 31.86 54.68 13.81 66.83 4.13 17.74 9.30 87.280.300/0.212 15.49 26.96 61.18 13.71 57.86 16.61 16.96 6.27 91.040.212/0.150 12.80 25.14 63.28 12.84 49.72 28.33 12.78 4.69 92.380.150/0.106 8.86 26.00 60.71 9.85 47.13 30.89 8.03 4.51 91.020.106/0.075 6.28 29.98 53.63 7.37 50.52 24.50 5.38 6.65 86.730.075/0.053 3.36 36.00 44.80 3.84 56.17 16.36 2.96 14.31 75.390.053/0.038 1.93 37.05 42.28 1.71 57.50 14.44 2.12 23.43 60.82−0.038 2.08 33.92 43.16 0.67 51.27 21.22 3.26 30.96 46.90

Table 3Main reconciled results obtained during the tests (F: feed rate (l/min); S: feed slurry solid concentration (% w/w); W: wash water flow rate (l/min)).

Test Factors Spiral feed Spiral reject Spiral concentrate Distribution

F S W t/h % Fe % SiO2 t/h % Fe % SiO2 t/h % Fe % SiO2 Solids Fe SiO2

1 60 (0) 37 (0) 16.7 (0) 1.63 33.75 50.55 0.90 11.68 81.82 0.73 60.4 12.81 44.8% 80.2% 11.4%2 48 (−1) 32 (−1) 8.3 (−1) 1.07 33.18 51.01 0.40 4.69 90.34 0.68 49.8 28.00 63.1% 94.8% 34.6%3 72 (+1) 32 (−1) 8.3 (−1) 1.35 34.55 49.15 0.51 6.15 89.39 0.84 51.8 24.76 62.3% 93.3% 31.4%4 48 (−1) 42 (+1) 8.3 (−1) 1.69 33.83 50.59 0.74 5.59 91.67 0.96 55.7 18.83 56.4% 92.8% 21.0%5 72 (+1) 42 (+1) 8.3 (−1) 2.01 33.38 51.16 0.77 5.39 90.77 1.24 50.8 26.49 61.6% 93.8% 31.9%6 48 (−1) 32 (−1) 25.0 (+1) 0.94 32.48 51.92 0.58 12.84 79.70 0.36 64.4 6.80 38.1% 75.5% 5.0%7 72 (+1) 32 (−1) 25.0 (+1) 1.46 31.65 53.28 0.89 12.26 81.11 0.57 61.7 10.11 39.2% 76.4% 7.4%8 48 (−1) 42 (+1) 25.0 (+1) 1.67 32.69 51.67 1.00 12.51 80.16 0.67 62.7 9.31 40.2% 77.1% 7.2%9 72 (+1) 42 (+1) 25.0 (+1) 2.54 32.36 50.73 1.46 12.01 78.55 1.09 59.7 13.42 42.7% 78.7% 11.3%10 60 (0) 37 (0) 16.7 (0) 1.75 32.83 51.60 0.99 11.67 81.64 0.75 60.7 12.00 43.1% 79.8% 10.0%

24 M. Sadeghi et al. / International Journal of Mineral Processing 129 (2014) 22–26

Eq. (1) (Hodouin et al., 1984). The calculated recoveries using Eq. (1)(Ri;m) are then plotted as a function of the mean geometric size of thesize intervals to obtain the size recovery curve for species m.

3. Analysis of the spiral performances

The effects of the studied experimental factors (wash water addi-tion, spiral feed rate and solid concentration) are analyzed first usingglobal performance indices such as concentrate grade and recoveryand then by using the size recovery curves for iron and silica carriers.

3.1. Effect of the factors on the global performance indices

An example of a set of reconciled data is presented in Table 2. Suchdata set is used to calculate the ore and species distribution and sizerecovery. A summary of the global performance indices calculatedfrom the reconciled data is given in Table 3. The results indicate thatthe feed composition remained fairly constant during the test work.The similar compositions of the concentrate and reject streams for thetests repeated at the reference conditions (tests 1 and 10) confirmthat the reproducibility of the tests is acceptable. It is also obvious

Table 4Independent and dependent variables considered in the analysis.

Independent variables Dependent variables (performance indices)

Feed volumetric flow rate Iron content of the concentrateFeed % (w/w) solid concentration SiO2 content of the concentrateWash water addition Solid distribution in the concentrateIron content of the feed Iron distribution in the concentrateSiO2 content of the feed SiO2 distribution in the concentrateDry solid feed rate

from the results of Table 3 that the variations imposed to the washwater strongly influence solid and element distribution in the concen-trate stream of the spiral. On the other hand the effects of the feedrate and solid concentration are difficult to assess from the results ofTable 3.

A stepwise regression (Himmelblau, 1970) analysis is used to iden-tify statistically significant effects for the independent and dependentvariables listed in Table 4. The usefulness of an independent variablein themodel is measured through the standard deviation of its estimat-ed parameter. A parameter value that is less than twice its standarddeviation is considered as non-significant. The application of the proce-dure shows that the washwater addition explainsmost of the variationfor all the considered performance indices or dependent variables ofTable 4. The introduction of any additional independent variables tothe wash water flow rate fails to produce a significant decrease in thecriterion. Some regression results for the iron content in the concentrateare given in Table 5. Similar results are observed for the other depen-dent variables of Table 4, i.e. that thewashwater addition is the variablethat has themost significant effect on the considered performance indi-ces. Any increase of the wash water increases the iron content of the

Table 5Regression models for the iron content of the concentrate (the number within brackets isthe standard deviation of the estimated parameter).

Model Criterion

xFe;C = 47.7 (2.0) + 0.61 (0.11) FWW 51.6xFe;C = 53.1 (4.8) + 0.61 (0.10) FWW − 0.09 (0.07) QF 42.4xFe;C = 53.1 (4.8) + 0.61 (0.10) FWW − 0.09 (0.07) %SF 51.5xFe;C = 53.1 (4.8) + 0.761 (0.19) FWW + 1.79 (1.76) XFe;F 45.0

xFe;C: iron content of the concentrate (%); XFe;F: iron content of the spiral feed (%); FWW: washwater flow rate (l/min); %SF: feed slurry % solids (% w/w); QF: volumetric feed rate (l/min).

a) Iron carriers b) Silica carriers

0

20

40

60

80

100

0,01 0,1 1 10

Rec

ove

ry t

o c

on

cen

trat

e (%

)

Particle size (mm)

Test 1

Test 10

Iron carriers

0

10

20

30

40

50

0,01 0,1 1 10

Rec

ove

ry t

o c

on

cen

trat

e (%

)

Particle size (mm)

Test 1

Test 10

Silica carriers

Fig. 3. Partition curves for the iron and silica carriers for the reference conditions (tests 1 and 10 of Table 1).

25M. Sadeghi et al. / International Journal of Mineral Processing 129 (2014) 22–26

concentrate and leads to a decrease in the recovery of iron, silica andsolids to the concentrate.

This result is primary due to thewashing of the silica carrier particlesaway from the inner part of the spiral trough from where the concen-trate is recovered. When light silica carrier particles are flushed awayfrom the inner trough it is possible that the density of the particle bedrestricts the return of these particles to the inner trough while denserhematite particles are able to travel back toward the concentratetraps. The size recovery curve for the silica carriers shown later confirmsthat the effect of the wash water is dominant for coarse particles. Thisresult is an indication that the Bagnold force is probably the dominantaction in cleaning the slurry within the inner trough of the spiral(Holtham, 1992). Several other variables such as the slurry viscosityand the particle bed thickness are difficult to measure and these vari-ables play a role in the rejection of the light minerals from the innertrough of the spiral. The way these variables interact to clean the con-centrate from light minerals is seldom discussed in a comprehensivemanner in the literature.

The insensitivity of the spiral performances to the changes of spiralfeed rate and slurry solid concentration is not consistentwith the obser-vations reported byDallaire et al. (1978). However the trend is probablymasked by the results of tests 2 and 4 which show amarked increase inthe iron content of the concentrate (from 49.8% to 55.7%) with anincrease in the slurry solid concentration (from 32% to 42%). On theother hand a comparison of tests 3 and 5, 6 and 8 and 7 and 9 shows a

a) Iron carriers b

0

20

40

60

80

100

0,01 0,1 1 10

Rec

ove

ry t

o c

on

cen

trat

e (%

)

Particle size (mm)

High wash water

Reference

Iron carriers

Low washwater

Fig. 4. Effect of the wash water addition on the partition curves for the iron and silica carr

systematic decrease of the concentrate grade with an increase in thefeed slurry solid concentration, a result that is consistent with the find-ings of Dallaire et al. (1978). The insensitivity of the spiral to changes inthe feed rate and slurry solid concentration could also be related to thefact that experiments are conducted within the manufacturer's recom-mended operating range (see Fig. 2) within which the spiral could berelatively insensitive.

3.2. Partition curves for the iron and silica carriers

Theprevious analysis shows that increasing thewashwater additionleads to a higher concentrate grade at the expense of recovery. The anal-ysis of the partition curves of the iron and silica carriers provides infor-mation about the particle size range that is mostly influenced by thewash water addition. The partition curves observed for the iron andsilica carriers for the reference operating conditions (tests 1 and 10 ofTable 1) of the spiral are shown in Fig. 3. The decrease in the recoveryof coarse particles is reported by different authors (Burt, 1984; Bazinet al., 2014; Atasoy and Spottiswood, 1995) and it is attributed tothe Bagnold force that preferentially acts on the coarse particles(Bouchard, 2001). The increase in the recovery of iron carrier particlesof 0.85 mm is not yet explained but it is also observed in data providedby Miller (1991).

The average partition curves for iron and silica observed with thespiral operating with high wash water addition (average of tests 6, 7,

) Silica carriers

0

20

40

60

80

0,01 0,1 1 10

Rec

ove

ry t

o c

on

cen

trat

e (%

)

Particle size (mm)

High washwater

Reference

Silica carriers

Low washwater

iers (high flow average of tests 6, 7, 8, and 9; low flow average of tests 2, 3, 4, and 5).

26 M. Sadeghi et al. / International Journal of Mineral Processing 129 (2014) 22–26

8 and 9) and lowwashwater addition (average of tests 2, 3, 4 and 5) arecompared to the average partition curves obtained at the reference con-ditions (tests 1 and 10) in Fig. 4. The effect ofwashwater is dominant onthe coarse iron carrier particles while for silica carriers it is distributedthrough all the particle size ranges. Results of these tests clearly showthat a WW-6 spiral that is operated within the manufacturer's recom-mended range of feed rate and slurry solid concentration is sensitiveto change in the wash water addition, making this variable a factorthat should be regulated accurately to a target value to avoid the pro-duction of erratic results from the operation of a spiral plant.

4. Conclusion

A series of tests were conducted using an iron oxide ore to analyzethe effect of the wash water addition, spiral slurry feed rate and solidconcentration on a WW-6 spiral performance. The test results clearlyshow that the spiral operation is fairly insensitive to change of feedrate and slurry solid concentrationwhen these variables aremaintainedwithin the operating range recommended by the manufacturer. On theother hand, the spiral operation is found to be sensitive to the washwater flow addition. Increasing the wash water flow rate increases theconcentrate grade at the expense of the iron recovery. The wash waterflow rate significantly impacts the recovery of coarse particles withhigher recoveries observed at low wash water addition rate.

References

Atasoy, Y., Spottiswood, D.J., 1995. A study of particle separation in a spiral concentrator.Miner. Eng. 8 (10), 1197–1208.

Bazin, C., Sadeghi, M., Roy, P., Bourassa, M., Cataford, D., Lavoie, F., Rochefort, C., Gosselin,C., Renaud, M., Mahieu, G., 2014. Simulation of an iron ore concentration circuit usingmineral size recovery curves of industrial spirals. Proceedings of the 46th CanadianMineral Processors, Ottawa, January 20–22, pp. 387–402.

Bouchard, S., 2001. Traitement du Minerai, Montréal, Les Éditions. Le Griffon d'argile.Burt, R.O., 1984. Gravity Concentration Technology. Elsevier, Amsterdam.Dallaire, R., Laplante, A., Elbrond, J., 1978. Humphrey's spiral tolerance to feed variations.

CIM Bull., August, No. 796 pp. 128–134.Himmelblau, D.M., 1970. Process Analysis by StatisticalMethods. Streling Swift publishing

company, Manchaca, Texas.Hodouin, D., Everell, M.D., 1980. A hierarchical procedure for adjustment and material

balancing of mineral processes data. Int. J. Miner. Process. 7, 91–116.Hodouin, D., Bazin, C., Trusiak, A.R., 1984. Reliability of calculation of mineral process ef-

ficiencies and rate parameters from balanced data. In: Herbst, J.A. (Ed.), Control 84.SME/AIME, p. 133.

Holtham, P.N., 1992. Particle transport in gravity concentrators and the Bagnold effect.Miner. Eng. 5 (2), 205–221.

Hyma, D.B., Meech, J.A., 1989. Preliminary tests to improve the iron recovery from the−212 micron fraction of new spiral feed at Quebec Cartier Mining company.Miner. Eng. 2, 481–488.

Miller, D.J., 1991. Design and operating experience with the Goldsworthy mining limitedBatac jig and spiral separator iron ore beneficiation plant. Miner. Eng. 4 (3/4),411–435.

Richards, R.G., Palmer, M.K., 1997. High capacity gravity separators: a review of currentstatus. Miner. Eng. 10 (9), 975–982.

Roche Mining, M.T., 2011. WW6E spiral separator. Technical Brochure. 11 Elysium Road,Carrara, Queensland, 4211, Australia (www.mineraltechnologies.com).

Wills, B.A., 1992. Mineral Processing Technology 5th Edition. Oxford Pergamon Press.