rheo concepts

8/4/2019 Rheo Concepts

1/13

3.1 PreamBle

Current practice in the minerals and oil sands industries oten involves pumping the bottom stream rom a clarier at lowconcentration to a waste disposal site or tailings storage acility (TSF). For conventional low-density tailings this is invariablycontained by earthen embankments, although, in some parts o the world, this waste is still being pumped into rivers or

into the sea. There is no doubt that the industry is changing and moving towards thickening the claried waste to a varyingextent, which can lead to thicker tailings, paste tailings, or a direct mine stope ll. There are many reasons to move in thisdirection related to water recovery, the environment and, perhaps, even economic reasons (see Chapter 2). The industry ismoving towards a situation that will ensure a disposal site involves less risk, is stable and environmentally acceptable. In theuture we will even see rehabilitation taking place while deposition is still taking place and not only at the end o the mineslie, which is the current practice i the company is still in operation.

Once the claried waste stream in these industries is thickened above a certain consistency it will exhibit non-Newtonianbehaviour. Thereore, the determination o the disposal system design and operating conditions requires a thoroughunderstanding o the rheological characteristics o the material, both in shear and compression. The implementation andoptimisation o a thickened disposal strategy involves three concurrent and independent rheological studies to determine:

The concentration required to achieve the optimum characteristics or management and/or storage o the particulatewaste fuid.

The optimum conditions or pipeline transport to the storage area.

The easibility o dewatering the slurry to the required concentration prior to transportation.

Thickened disposal is dened here as any system designed to take advantage o a high-density thickened and/or pastetailings. This chapter provides an inormed background to ensure the appropriate rheological issues are addressed as part othe overall tailings design and management strategy.

The main criterion or the success o a tailings management system is to implement and operate a sae and environmentallyresponsible system at minimum cost. Exploitation and manipulation o the transport and deposition characteristics o thetailings are vital to ensure that the appropriate disposal system is operated successully.

A thickened disposal system is selected and designed to minimise the pumping energy required whilst delivering a materialat the velocity, yield stress and viscosity needed or the desired slope and spreading characteristics or surace deposition,and/or the desired characteristics or mine stope ll. This chapter elucidates the signicance o rheological characterisation

or the planning, design, operation and optimisation o thickened disposal systems that are considered here to include drystacking, central thickened tailings disposal, and paste backll.

3Chapter 3

Rheological Concepts

lead Author

David Boger The University of Melbourne, Australia

25


2/13

26

3.2 SUmmarY floWSHeet for tHe

Selection, deSign and teSting

for tHickened diSPoSal

imPlementation

Figure 3.1 summarises the rheology-based decisions,considerations and test work required or a proposedimplementation strategy in fow chart orm. Any greeneldsite should examine the entire process, ranging romthickening through to the ultimate disposal. Oten the issueis dealt with by particular consultants and/or contractorsor each part o the process; one or thickening, one orpumping, and another or the construction and managemento the disposal site. An integrated approach is required. Notethat in addition to the rheology and rheological properties,the chemistry can also be quite important; clay is a majorissue in many mineral waste tailings. When clays are present

in the orebody the waste product can be very dicult tohandle and will oten not consolidate to any signicantconcentration. Some mineral, sand mine, and oil sandprocessors do not know the specic types o clay they are orwill be encountering. Understanding the basic chemistry othe clays in the waste material is important as sometimes itcan be changed to have a considerable impact on the ultimate

disposal strategy (see Chapter 5 or more inormation onslurry chemistry). As an example o the importance o claychemistry, Figure 3.2 illustrates a clay at 25% w/w solids,where the fow characteristics vary dramatically with methodo dispersion. Dispersion with water only (let) yields a pottersclay, while controlled dispersion with 1 M CaCl

2(right) yields

a settling, well-behaved suspension. Uncontrolled dispersion(centre) yields a non-settling viscous suspension. Clays in theuncontrolled dispersed state oten end up in tailings pondsand never consolidate.

figUre 3.1 Panning, design and testing fow sheet

Paste and Thickened Taiings A Guide (Second Edition)


3/13

27

3.3 tHe imPortance of rHeologY

in a tHickened diSPoSal

imPlementation StrategY

By considering tailings rheology at all points during thedesign phase, implementation strategies can be developed

or new operations or existing systems optimised to reduceoperating costs and/or the impact on the surroundingenvironment, and or the recovery o the water.

The strategy proposed in this chapter looks at designingthe tailings to suit the surrounding environment, with

consideration being given to the choice and operation othe disposal method. Figure 3.3 briefy summarises the stepsinvolved in determining a disposal strategy. It is importantto note that the design sequence begins at the disposal pointand works upstream (back) to the thickening stage.

Point 1 in the design sequence represents the choice

o the disposal method and the rheological propertiesrequired to achieve the desired ootprint and shape o thenal deposit (depositional requirements). Point 2 reers tothe pumping and pipeline conditions needed or optimaltransport whilst ensuring the tailings reach the disposal site

figUre 3.3 Suggested approach or the determination o taiings disposa system and requirements

Rheoogica Concepts

figUre 3.2 Controed dispersion o a montmoriinite cay


4/13

2

with the rheological properties necessary or the chosendisposal method (pipeline requirements). Rheologicalproperties sometimes change in the pipeline itsel.

Once the depositional and pipeline issues have beenquantied and optimised on paper, it is possible todetermine the thickening requirements represented bypoint 3 in Figure 3.3. The thickener should be designed

and operated to produce the required tailings propertiesvia determination o thickener size and geometry, andcontrol o focculation and pre-treatment requirements.I the thickener requirements cannot be achieved then theprocess is repeated.

For clarity, the design sequence is depicted in Figure 3.3as a linear sequence. However, in reality the disposal systemand thickener should be designed using an iterative approachin order to optimise the entire disposal operation.

3.4 imPortant rHeological concePtS

Rheology is the study o the deormation and fow omatter. In terms o fuid fow, materials may be classied aseither Newtonian or non-Newtonian fuids. The viscosity(h) o a fuid is dened as the ratio o the shear stress (t)tothe shear rate , as shown in Equation 3.1. In many fows,the shear rate is equivalent to the gradient in velocity.

(3.1)

Inelastic Newtonian fuids exhibit a linear relationshipbetween the applied shear stress and the shear rate, as shownin Curve A in Figure 3.4. Flow is initiated as soon as a shearstress is applied. The linear relationship between the shearstress and the shear rate indicates a constant viscosity.

Concentrated mineral tailings oten display non-Newtonian fow behaviour in that they possess a yield

stress. The yield stress (ty) is the critical shear stress thatmust be exceeded beore irreversible deormation and fowcan occur. For applied stresses below the yield stress, theparticle network o the suspension deorms elastically, withcomplete strain recovery upon removal o the stress. Oncethe yield stress is exceeded, the suspension exhibits viscousliquid behaviour where the viscosity is usually a decreasingunction o the shear rate. Yield stress behaviour is shownby the constitutive relationships o Equations 3.2 and 3.3.

(3.2)

(3.3)

Curve B on Figure 3.4 shows a yield stress ollowed bya linear shear stress shear rate relationship, commonlyknown as Bingham behaviour. Although not a true viscosityaccording to Equation 3.1, the gradient o this line isreerred to as the Bingham plastic viscosity.

Table 3.1 lists typical yield stress values. The values listedor thickened tailings disposal ranging rom 30 to 100 Paare indicative o those used in the alumina industry today.There are other thickened tailings strategies where thematerial is discharged and fows like a river delta, where the

yield stress would be greater than 0 but below 30 Pa. Minestope ll is a true paste material where the yield stress, inour experience, can become as high as 800 Pa.

taBle 3.1 Typica yied stress vaues

Substance Yied stress (Pa)

Tomato sauce 15

Yoghurt 80

Toothpaste 110

Peanut butter 1900

Thickened tailings disposal 30100

Mine stope ll 250800

In addition to yield stress behaviour, the viscosity othe material will vary with shear rate. As the shear rate isincreased, pseudoplastic or shear thinning materials exhibita decrease in the viscosity (Curve C, Figure 3.4). Dilatantor shear thickening materials exhibit an increase in viscositywith increasing shear rate (Curve D, Figure 3.4). Dilatant

behaviour, although relatively rare, is sometimes observedin mineral suspensions. For the dierent fuid categories,various empirical fow models are used to describe the fowbehaviour. The most commonly used equations are theOstwaldDe Waele model or shear thickening or shear

figUre 3.4 Typica fow curves. A) Newtonian;B) Bingham (yied-constant viscosity);C) Yied-pseudopastic; D) Yied-diatant



5/13

2

thinning materials, the Bingham model or yield stressmaterials and the Herschel Bulkley model or yield stress,shear thinning or yield stress, shear thickening materials.

OstwaldDe Waele power law model:

(3.4)

Bingham model:

(3.5)

Herschel Bulkley model:

(3.6)

In these equations, K and n are experimentally determinedconstants.

The shear thinning nature o industrial tailings isattributed to the alignment o particles and/or focs in thefow eld. An increase in the shear rate rom rest, resultsin instantaneous alignment o particles in the direction o

shear, thus providing a lower resistance to fow. As such, thesuspension will show a decreasing viscosity with increasingshear rate.

Data on mineral suspensions are oten not available atlow shear rates. A common practice is to extrapolate thelinear portion, say, o Curve C in Figure 3.4 to the y-axisand reer to this as the yield stress. This is a Bingham yieldstress which bears no relationship whatsoever to the trueyield stress o the material. The linear portion o the curvebeing tted in this way sometimes is appropriate or pipelinedesign but a yield stress determined in this way is ar rom

adequate or the design o a thickener; more accurate yieldstress measurements can be and should be made.

The most complicated non-Newtonian behaviour, whichcan be observed in mineral and tar sand, and associatedindustries, is time-dependent behaviour. Here, the shearstress is a unction o both shear rate and time o shear.Basically, this is the case where the alignment o the particlesand/or focs occurs on a timescale which can be observed inthe rheometer. The classical example o this behaviour isor red mud in the alumina industry (Boger et al., 2002).

For such materials, the viscosity depends both on shearrate and time, whilst the yield stress also varies with time.When such a material is encountered it is necessary to seehow the yield stress varies with time and to measure theshear stress in the material in a rheometer as a unction otime at particular shear rates. Alternatively, with a constantstress rheometer, x the stress and watch how the shear

rate varies with time. Thixotropic behaviour is the mostcommon, where the viscosity decreases with shear rate andtime. Rheopectic behaviour, the time-dependent dilatantbehaviour analogue, is observed ar less oten, but when itdoes occur, it can present disastrous consequences. Here, theviscosity increases with shear rate and time. Such behaviouris quite perplexing or the engineering community, wherethe tendency is, when in doubt put in more energy, or hitit harder with a hammer. For example, in a mixing vessel,i the viscosity increases with shear rate and time then anyeort to get better mixing by increasing the rotationalrate o the impeller will, and can, result in catastrophic

consequences. We have seen this happen on a ew occasionsin the past with mineral suspensions. It is very importantto have a basic classication o the particulate fuid to beencountered in your industry.

3.5 rHeological ProPertY

meaSUrement

There are entire books written on rheological propertymeasurement (rheometry). For particulate fuid suspensionsand, in particular, more concentrated suspensions like thosethat will be encountered or thickened tailings disposal,

many o the rheological instruments available will not beapplicable. All o those relying on a small gap between theshearing suraces, like a cone and plate geometry or a parallelplate geometry, generally are not used and are not applicablein the industry. The most common instrument used is that which relies on Couette fow, involving connement othe sample between a rotating cup and a stationary bob,or (vice versa) a rotating bob and a stationary cup. Here,the torque is measured as a unction o rotational speed

Flow curves for paste sample Yield st ress = 250 Pa

Shear rate (1/s)

1 10 100 1000 10000

100

1000

Vane

yield

stress

10

Capillary results

Concentric cylinder results

Vane and cup results

Vane yield stress

Shearstress

(Pa)

figUre 3.5 Fow curves or a mine stope sampe, yied stress = 250 Pa

Rheoogica Concepts


6/13

30

and can be interpreted to determine the shear stress as aunction o shear rate. Extrapolation o data to zero shearrate rom such an instrument is generally not a good wayto determine the yield stress. Sedimentation o the samplein a Couette rheometer and slip at the wall are commonproblems. Another way to obtain shear stress shear ratemeasurements generally is with a capillary rheometer or

rom actual pipe loop-test data obtained in laminar fow.Here, the wall shear stress:

(3.7)

where p, the ully developed fow pressure drop in a pipeo length l is measured as a unction o the apparent wallshear rate 8V/D, where Dis the pipe or tube diameter and Vis the average velocity in the pipe. The wall shear rate:

(3.8)

can be determined rom the slope o a loglog plot otw

versus 8V/DThe shear rate dependent viscosity is thenh= t

w/

w. Data o this type are very useul or design

purposes (see Chapter 8).

Figure 3.5 shows fow curves or a mine stope ll materialwith a yield stress o 250 Pa. Note the discrepancy in the dataat lower shear rates. The capillary results and the concentriccylinder results agree or shear rates in excess o 300 s-1 butdeviate and drop away signicantly at the lower shear ratesrom those obtained with the vane and cup. I the inner

rotating cylinder in a Couette apparatus is replaced by avane, problems associated with slip are generally eliminated.The true yield stress o the material shown in Figure 3.5was about 250 Pa, whereas the data obtained both rom the

capillary and rom the concentric cylinder present valueso 64 Pa or 17 Pa, depending on the extrapolation used.The huge dierence would have a very signicant eect ondesign, particularly o a thickener. The rst lesson to learnis that data obtained with a concentric cylinder rheometerand sometimes with a capillary rheometer will oten bein signicant error at low shear rates. Great care must be

taken, particularly i used to determine a yield stress.In summary, the rheological measurements o most

interest or thickened tailings are generally obtained witha concentric cylinder apparatus and/or rom a capillaryrheometer. The vane and cup, which is a relatively newtechnique, represents what we believe is the best way togo in terms o determining both the yield stress and basicshear rate shear stress data or viscous shear thinningsuspensions.

3.6 rHeological reQUirementS for

tHe cHoSen diSPoSal metHod

We will now move step-by-step through the procedurerecommended in Figure 3.3.

3.6.1 dps qus P 1,

fu 3.3

Oten an essential depositional requirement is to delivertailings to the disposal point with a yield stress sucientto support the largest particles and ensure a homogeneoussuspension where segregation does not occur. Dierentialsegregation upon deposition will result in a non-uniorm

deposit slope and inecient use o the disposal area. Therelationship between the solids concentration and the yieldstress should be determined in the rst instance to indicate theminimum concentration necessary to obtain this yield stress.

figUre 3.6 The vane technique



7/13

31

A signicant amount o work on the measurement o theyield stress or mineral suspensions has been completed.From these works, novel and simplied measurementtechniques have resulted. The vane-shear instrument andtechnique allows direct and accurate determination o theyield stress rom a single point measurement (Nguyen, 1983;Nguyen and Boger, 1983, 1985). Many workers worldwidehave adopted the vane-shear method and conrmed itsapplicability or all types o yield stress materials (Yoshimuraet al., 1987; James et al., 1987; Avramidis and Turian, 1991;Liddell and Boger, 1996; Pashias, 1997).

Figure 3.6 shows an illustration o the our-bladed vaneand the technique or yield stress measurement. Basically,the vane, attached to a torsion measuring head, is careullyinserted into the test material and rotated at a low speedwhere the torque as a unction o time is measured. Themaximum in the torque curve is related to the yield stress

via Equation 3.9, where Tm

is the maximum torque; Dv isthe diameter o the vane; and H/Dvis the aspect ratio o thevane cylinder.

(3.9)

In an attempt to urther simpliy yield stress measurement,the slump test has been modied to accurately evaluatethe yield stress o mineral suspensions (Showalter andChristensen, 1998; Pashias et al., 1996; Boger et al.,2002). The technique has typically been used to determinethe empirical fow characteristics o resh concrete and isimportant in mine stope ll. The slump test is conductedusing an ASTM standard slump cone, or with a cylinderand ruler, eliminating the need or sophisticated equipmentand allowing easy, on-site yield stress measurement by

Variabe Coa taiings God taiings leadzinc taiings

Specic gravity (kg/m3) 1450 2800 4100

Solids concentration (%w/w) 36 75 75

Slurry density (kg/m3) 1120 1930 2310

Slump height (mm) 203 203 203

Cacuated yied stress (Pa) 160 275 330

Predicted pressure drop (kPa/m)* 5.07 .13 .60

taBle 3.2Sump height versus yied stress: density dependence

*Pressure drop prediction assumes:

Bingham material

Bingham viscosity 1 Pa.s

Horizontal pipeline

Pipeline internal diameter = 200 mm

Pipeline velocity 1 m/s

Rheoogica Concepts

figUre 3.7 Comparison o yied stress measuring techniques


8/13

32

plant operators. We recommend the use o a cylinder. Theequation or evaluating the yield stress rom a slump heightmeasurement is given in Equation 3.10:

(3.10)

wheretyis the yield stress dimensionalised with respect togh; and S'is the dimensional slump measurement, whichis dimensionalised with h, the height o the slump cylinder;

is the density o the suspension; andgis the accelerationdue to gravity.

A comparison o the yield stress derived rom the slumpmeasurement and the yield stress measurement with thevane is shown in Figure 3.7. Users o the slump methodpersist in measuring the slump in a linear dimension,centimetres o slump. This is an empirical measure and willvary rom material to material depending upon its density.Table 3.2 eloquently illustrates the point that the slumpis not a unique physical property measurement. Here, acoal tailings, a gold tailings, and a leadzinc tailings, are all

figUre 3.8 Yied stress concentration data or dierent industry waste streams

figUre 3.9 Yied stress concentration data or cay waste rom the phosphate industry



9/13

33

concentrated to a stage where they produce the same slump,203 mm. Hence, at rst glance they all appear to be thesame in terms o their fow characteristics. This, however,is not so. The three materials are vastly dierent in thattheir yield stress varies rom 160 Pa, to 275 Pa, to 330 Pa

respectively, or coal, gold and leadzinc tailings. I, indeed,we use these slump measurements as a Bingham yield stress,and i all the materials have the same Bingham viscosity o1 Pa, the pipeline pressure drops will be vastly dierent.However, i the slump is converted to a yield stress, as perEquation 3.10, then we certainly see the dierence in thematerials.

The fow properties o concentrated mineral suspensionsvary signicantly with solids concentration, particle sizeand size distribution, pH and ionic strength. However, anumber o common characteristics have been observed orconcentrated suspensions in general. The strong dependence

o rheology on concentration is illustrated in Figure 3.8,which shows the yield stress as a unction o concentrationor eleven dierent mineral tailings. The rst thing that isobvious rom the gure is that every material is dierent,

even within the same industry, as was illustrated or thewaste rom bauxite residue (red mud) in the rst editiono this Guide (Boger et al., 2002). All materials exhibit anexponential rise in the yield stress with concentration, withthe yield stress generally beginning to rise rapidly or yield

stresses beyond about 200 Pa. Remember now the range oyield stresses used or the various thickened exercises. Forthickened tailings the yield stress might range rom a verylow value to about 30 Pa. High-density thickened tailingsor distribution on the surace generally are handled atyield stresses o about 30-100 Pa and it is now believed thatsuspensions can be pumped up to a yield stress o about200 Pa with a centriugal pump. Paste backll materials areoten handled at yield stresses signicantly higher than 200Pa. The type o curve shown in Figure 3.8 is the rst stepin evaluating the potential or a particular waste-disposal

strategy, based on the material properties o the tailings.The concentration at which the yield stress begins to riserapidly is very signicant, particularly when optimisingpumping energy requirements.

Rheoogica Concepts

figUre 3.10 Viscosity shear rate data or a coa washery thickener underfow


10/13

34

As already stated, or thixotropic materials shear history

must be taken into account when determining the yield

stress concentration relationship. During a design phase

it is unlikely that the extent o structural breakdown

during pipeline transer will be known. As such, the yield

stress concentration relationship should be determinedor structural states ranging rom the initial (rest) to the

equilibrium state (ully sheared).

65 tonnes/day

0.1524 m pipe diameter

Pumpingenergy(W

/m)

100

200

300

400

0

5 10 15 20 25 30 35 40Volume fraction (%)

Turbulentflow

Laminar

flow

figUre 3.11 Optimum concentration or pumping thecoa washery underfow

figUre 3.12 The predicted pumping energy as a raction o soids concentration or a 50. cm (20 inch) diameter pipeineat various capacities

From data like that shown in Figure 3.8, the concentrationrange required to produce the desired yield stress canbe determined. The slope obtained or a given materialenables the disposal area to be sized. Increase in the yieldstress will generally produce a steeper slope and reduce theootprint; however, even today it is not possible to preciselypredict the slope o the angle o deposition rom yield stress

measurements only.Figure 3.9 shows the yield stress as a unction o solids

concentration or the ne particle clay waste stream in thephosphate industry. The ne particle waste produced in thisindustry causes the most diculty in terms o consolidation.Generally, these materials are pumped to disposal areas atconcentrations less than 10% by weight and consolidationin a dam to concentrations greater than 30% is rare. Insome cses it may be possible to shit the curve in Figure3.9 signicantly to the right by controlling clay chemistry,resulting in a higher solids content and thus higher in situdry density in the TSF.

3.6.2 Pp sp qus

P 2, fu 3.3

The data required as a rst step in the design o atransmission pipeline or a tailings are viscosity shearrate data as a unction o concentration, such as the resultsshown in Figure 3.10 or an Australian coal waste stream.The material exhibited no signicant thixotropy, washighly shear thinning, and demonstrated normal viscosityconcentration behaviour. Also, the yield stress or thismaterial was obtained as a unction o concentration. Using

the yield stress data and the viscosity shear rate data ordierent concentrations, the results shown in Figure 3.11were obtained. Here, the pumping energy requirements in



11/13

35

watts per metre length o pipe are expressed as a unction ovolume raction or the delivery rate expected in a 0.1524m diameter pipeline. Notice that there is a minimum in thecurve that coincides with the minimum energy requirementor pumping this material at this concentration. The basiso this calculation is the delivery o dry solids, hence theminimum in the curve. At the lower concentrationsturbulent fow is present and, as the concentration isincreased, eectively by removing water to maintain the

dry solids constant, there is a transition between laminarand turbulent fow which occurs basically at the minimum. Almost all pipelines in the world today are transportingmaterial in the turbulent fow regime using a critical velocityto keep the particles in suspension.

There is a great reluctance to move into the laminar fowregion, even though the results in Figure 3.11 show that onecould deliver the material at a much higher concentrationor the same energy expenditure. There is a movementtowards handling materials at higher concentrations orgood environmental reasons and water recovery reasons,and thus we will have to deal with laminar fow pumping.

The ear in dealing with a material in laminar fow isthe sedimentation and blockage o the pipeline. In someinstances, particularly or relatively uniorm ne-particlesuspensions, the type o prediction shown in Figure 3.11 isquite accurate. In others, where the particle size distributionis very broad with large particles, sedimentation and/or slipfows may occur in the pipeline, and hence the predictionmay not be valid (see Chapter 8).

Figure 3.12 shows similar predictions, based on thebasic rheological measurements or a 50.8 cm diameteriron ore pipeline at various throughputs. It is interesting

to note that the minimum in the curves at about 65%solids is approximately where the pipeline is operated. Thisoptimum has been determined by trial and error.

Finally, Figure 3.13 shows similar results or the washedwaste rom a bauxite mine in comparison to one data point

o the measured pressure drop in the pipeline. Notice theexcellent agreement. From the results shown in Figures3.11 to 3.13, one might be tempted to conclude that onthe basis o basic rheological measurements one can in actpredict and design a tailings pipeline rom rst principles.Unortunately this is not true. As already stated, sometimesthe prediction and actual perormance agree, normally with

ne particles o a relatively uniorm size, but invariablywith waste streams there is little agreement. Thereore, inaddition to basic rheological measurements, most designerswill resort to actual fow loop studies (see Chapter 8). We willnot deal with design considerations or thixotropic materialsas these were dealt with in the rst edition o this Guide.

3.6.3 th qus:

pss hy P 3,

fu 3.3

The aim o thickening is both to recover water intothe process stream and produce an output slurry o a

suitable rheology or pumping and, more particularly, ordeposition purposes. To achieve this goal, knowledge o thecompressional rheology o the slurry is desirable.

The compressional rheology o a slurry has threeparameters o relevance to the prediction o the propensityo this same slurry to increase in solids in a thickener: thecompressive yield stress, hindered settling unction, and thegel point. The derivation o these parameters is well capturedby the continuum theory o dewatering developed byBuscall and White (1987). In simple terms, the parametersdescribe an inter-phase drag or rate o fuid expression rom asuspension under compression (hindered settling unction)

and the strength o the particulate network in compression(compressive yield stress). This approach implicitly assumesthat the suspension o particles is both focculated and, inthe case o compression measurement, is at a concentrationsuch that interparticle interactions within the suspensioncause a continuous network to exist that must deormunder load. The concentration equivalent to the onset othis networked behaviour is known as the gel point.

The measurement o the hindered settling andcompressibility behaviour o a slurry is now well establishedor mineral suspensions, although complete characterisationis still conned to research laboratories. The compressive

yield stress is measured through a combination o ltration,centriugation and settled bed height experiments (deKretser et al., 2001; Green et al., 1996) and the hinderedsettling unction can be established rom a combination otransient settling tests and ltration (de Kretseret al., 2001;Lester et al., 2005).

In a conventional gravity thickener, it is usual to addfocculants to the incoming slurry (see Chapter 6). Theeect is both to produce aggregates with a aster settling rateand to improve the clarity o the overfow liquor throughenmeshment o small particles into the aggregates. The aster

settling rate o the focculated aggregates is an importantelement in the prediction o thickener perormance, sinceit has been established that the operation o most thickenersis permeability limited. In simple terms, although thecompressive yield stress o a slurry dictates the maximum

Rheoogica Concepts

figUre 3.13 Prediction versus observation in an

existing bauxite-washing operation o thewaste taiings


12/13

36

possible solids that could be achieved through thickeningat a given bed height o solids in a thickener, this solidsconcentration is rarely, i ever, challenged in a steady statethickener because the rate o escape o liquor rom the bedo particles limits the process.

One approach to achieving higher solids content in athickener is to increase the residence time in the device.

Thereore, the practice o running a large compressionalbed in a thickener is opportune in that it both allows alonger residence time or the escape o fuid and increasesthe potential upper solids limit in terms o the compressiveyield stress. Another option is to increase the aggregatesettling rate. This is usually achieved through an increase inthe focculant addition rate. The overall eect is to improvethe permeability o operation whilst limiting the maximumsolids slightly. Given that thickener operation is permeabilitylimited, the result is a higher output o solids. Figure 3.14presents a typical fux operating curve or a thickenershowing the regions o permeability and compressibility

limited operation or three focculation conditions. Thevertical dotted line indicates the gel point or the slurry.The data are or an operating bed height o 5 m. Increasingthe focculant dose shits the fux curve upwards and isbenecial or all but compressibility limited operations.The transition rom permeability to compressibility limitedbehaviour is usually limited to settlement in TSFs.

The above analysis or a thickener can be achievedthrough either taking a sample on site that has alreadybeen focculated, or focculating in the laboratory andthen, through settling, sedimentation and ltration tests,constructing a compressive yield stress and hindered settling

unction prole or the slurry at that particular focculationcondition. Changing the dose or manner o focculationwill change the prole and necessitate urther testing. Thisinormation is then ed into a phenomenological pseudo

figUre 3.14 Prediction o the soids fux through athickener as a unction o underfow soidsor sampes foccuated at three dierentfoccuant doses. An increase in foccuantdose shits the prediction to higher fux

two-dimensional model o steady-state thickening (Usherand Scales, 2005).

Intuition associated with the above discussion wouldsuggest that the way to get to very high solids in a thickeneris in act to overdose with focculant. Practitioners knowthis is not useul since the component o compressionalrheology that has not been discussed is the eect o shear/

raking processes in enhancing the rate o dewatering andimproving the gel point o slurries. Recent work has begunto quantiy these eects and has shown that the eect oraking processes in thickening is to dramatically lower theinter-phase drag in the system at a given solids contentand increase the gel point o a suspension (Gladman et al.,2005). This eect is diminished as focculant dose increases.Thereore, quantication o the compressional rheology oslurries shows that the appropriate interpretation o therole o focculants in thickening is that at low focculantdose, and a xed bed height in the thickener, the underfowsolids increase due to enhancement o the settling rate and

reduction o the drag, and an increase in the gel point dueto raking. There is a transition at higher focculant dosesto the enhancement being solely due to improvements insettling. This produces an optimum focculant dose wherethe permeability is a maximum. This maximum will befocculant specic and shear rate dependent.

3.7 conclUSion

Industries in the uture producing particulate suspension wastes and delivering these wastes to storage areas atlow concentration will move towards more responsiblebehaviour by thickening their waste, thereby recovering water, reducing the space required or waste managementand, more importantly, reducing the risk associated withconventional TSFs. In order to remove water eectivelyrom the waste stream, it is essential to have a basicknowledge o both shear and compression rheology, as thematerials are non-Newtonian. Basic knowledge o bothshear and compression rheology will allow a more optimaldesign o the disposal system. The disposal strategy mustinclude examination o the whole process, which includesthickening, pumping and deposition, concurrent with

optimum focculation practice o the waste stream leavingthe plant. For a waste disposal strategy to be eectivelymanaged, it is essential that the specication o the wastestream be regarded as just as important as the specicationo the product produced by the plant. Maybe one day wewill even start to think about cleaner production instead odealing with these wastes as an end o pipe problem.

The implementation and optimisation o dry disposalmethods involves three concurrent and interdependentrheological studies to determine i) the optimumconcentration required or the management o the tailingsstorage area, ii) the optimum conditions or pipeline

transport, and iii) the easibility o dewatering the slurry tothe required concentration. The technical methods outlinedin this chapter provide the rheological characterisationrequired to complete these studies.



13/13

As environmental actors translate into economic issues,the push or minimising waste production using dry disposalmethods is gaining popularity. The use o rheologicalinormation is o high importance in evaluating thisnew technology. The principles outlined in the examplesgiven in this chapter may be applied to many industriesencompassing a wide range o waste materials.

acknoWledgementS

The authors would like to acknowledge the ParticulateFluids Processing Centre, a Special Research Centre othe Australian Research Council at The University oMelbourne, or the encouragement received or this work.In addition, the lead author would like to express hisextreme thanks to the Alcoa Company o Western Australiaor involving him in their work on red mud disposal,starting in the 1970s. Particular thanks to Peter Colombera(deceased), Don Glenister (Alcoa, retired), and to David

Cooling, currently the International Disposal Manager orAlcoa.

Rheoogica Concepts

aUtHor detailS

lead Author

David Boger The University of Melbourne

David Boger is Laureate Proessor at The

University o Melbourne. His research is

in non-Newtonian fuid mechanics withinterests ranging rom basic polymer

and particulate fuid mechanics to applications in the

minerals and other industries. Proessor Boger has

received numerous awards including the Victoria Prize in

2002 and the Prime Ministers Prize or Science in 2005

in recognition o his research achievements.

c-auhs

Peter Scaes The University of Melbourne

Peter is Head o the Department o

Chemical and Biomolecular Engineering

and a team leader in the Particulate Fluids

Processing Centre, responsible or solid-

liquid interacial processes. Peter is also a Director o

Rheological Consulting Services. His research interests

include focculation and coagulation o suspensions,

dewatering processes including thickening and

ltration, and the rheology o concentrated focculated

suspensions.

Fiona Sora Rheological ConsultingServices

Fiona completed her PhD on the

rheological aspects o tailings disposal at

the Particulate Fluids Processing Centre,

The University o Melbourne, Australia in 2000 beore

orming Rheological Consulting Services Pty Ltd (RCS).

Fiona is currently Managing Director o RCS. Particular

areas o expertise are: rheological characterisation o

non-Newtonian thickened mineral tailings and pastes,

colloid and surace chemistry, and the evaluation o

material dewaterability.

rheo concepts

Documents