+ All Categories
Transcript
Page 1: Chapter 13. Solid - Liquid Separation

Chapter 13. Solid - Liquid Separation

13. INTRODUCTION

The process of grinding and classification involves the use of large quantities of water. In thegold industry for instance, the rule of thumb is a tonne of water for a tonne of ore. This bulkof water has to be separated or reduced for down stream treatment for recovery of the mineralin the ore. The separation of solids from liquids is usually achieved by gravity sedimentationin thickeners. For fine particles this is a slow process. In general 75-80 % of the water can beseparated and removed by thickeners. For further water removal, filters are used where inexcess of 90% of the water can be removed. The thickener operation can be a batch orcontinuous process, with either co-current or counter-current flow of underflow and overflowslurries. The filtering operation may also be batch or continuous.

For rapid solid-liquid separations, centrifugal forces are used and equipment similar tothose described under classification are employed, ha this chapter, we shall deal mostly withthickeners working under gravitational forces.

13.1. Design Features of ThickenersThickeners are essentially clarifiers producing a clearer over flow. The design

considerations are based on the settling rates of the slowest settling particles and conditionsfor minimum disturbance of the medium (water) through which the solid particles areallowed to settle. To achieve these objectives cylindrical tanks with conical or flat bottomsare used and the velocity of the feed slurry entering the settling tank is minimised to reduceturbulence in the settling tank. A schematic diagram of a typical thickener is shown in Fig.13.1. The feed in the form of slurry is generally guided by a launder, which is laid at a slopejust sufficient for the slurry to flow without depositing any solids. The feed launderterminates in a feed well located at the centre of the tank. The feed well is designed to breakthe fall of the slurry and dissipate the energy.

The feed well is concentric with the rake driving shaft. The rakes are bolted or welded onto this drive shaft and for long and large rakes they have additional support from cables.Usually four rakes are employed of which two may be short and two long. Attached to therakes and below them are spikes, particularly in situations where the sludge is thick. Thespikes help to break up the sludge and render it more suitable for pumping. The rakes aredriven by a motor which is mounted on a plate above the well. An alternative is to mount thedrive motor on a track running along the rim of the tank. A bridge usually runs from theperiphery to the centre of the tank. It is supported by the wall of the feed well and the rim ofthe tank. The bridge serves as a walkway and also carries an open launder (or pipe), whichcarries the slurry to the feed well. In some designs the bridge spans the entire length of thetank. As in clarifiers, the bottom of most tanks slope towards the centre where the thickenedunderflow sludge accumulates. When a flat bottomed tank is designed, the settled sludgebuilds up to form its own slope depending on the angle of repose of the material thus formingan artificial sloping tank bottom. The sludge collected at the bottom is discharged through anoutlet shaped like a cone with steep cone angle. Alternately, the thickened slurry is swept

Page 2: Chapter 13. Solid - Liquid Separation

402402

towards a trench at the bottom of the tank. Usually a scraper is installed for smooth deliveryof sludge from the discharging cone or trough. A slush or centrifugal pump subsequentlyremoves the sludge.

The thickener tanks are usually fabricated using steel sheets. But tanks with concrete sidesare quite common. Some small tanks (usually < 30 m in diameter) are made of plastics. Thewhole assembly is installed either above ground sitting on pillars or at ground level with thedischarge well below the ground level. In the latter case, an access tunnel is provided wherethe discharging pump is located. In some installations the discharge pump is located abovethe tank; in such cases, a suction pipe runs down the centre column to the bottom well.Alternatively, a submerged motor pumps the under flow slurry to the top of the tankdischarging its contents to a holding tank.

Several variations are known to exist. For instance the rakes are either supported by crossbeams or truss above the tank or supported by the central column and cables. The cables arealso connected to torque meters.

Fig. 13.1 is a sketch of a bridge thickener where the bridge runs across the thickener tank.The bridge support the rakes and the motor rotating the rakes sit on a platform in the centre ofthe tank. The rakes are bolted to the central column which is rotated by the motor. TheBridge thickeners have a maximum diameter of about 30 meters.

When the rakes are supported entirely by the central pillar, the access bridge usually runshalf way on the tank surface terminating on the central pier. The centre pier thickeners areconsiderably larger than the Bridge type. The diameter of the tank ranges from about 35-180meters.

A variation is the tray thickener where trays or compartments are placed one on top of theother. Each tray acts as a thickener and the assembly operates in parallel with a common pieror shaft where the rakes are fixed. Clarification takes place in series operation, that is, thethickener underflow from the top compartment serves as feed to the lower compartment.Ultimately the underflow from say, a six tray thickener, form the final thickened underflow.Similarly all the overflow from each tray combine forming the final overflow slurry. Fig. 13.2is a schematic diagram of a 3-compartment clarifier. Up to seven compartments are available.

drive motor

access bridge

feed

Fig. 13.1. Sketch of a thickener showing the access bridge, feed well, rakes supported by the centralcolumn and cables and the underflow discharge.

Page 3: Chapter 13. Solid - Liquid Separation

403403

overflow feed box

Sludge

Fig. 13.2 . Schematic diagram of a tray type clarifier, after Dahlstrom and Fitch [1].

The thickening process is accelerated by the addition of flocculants Hi-Capacity thickenersallow a mixing arrangement in the feed box where the flocculant is intimately mixed. Theother design features of the Hi-Capacity thickeners are similar to the Bridge thickeners.

While installing the feed pipe or launder to thickeners, the slope is held at 1 to 1.5. King[2] suggests that this slope provides minimum turbulence of the settling slurry in the tank.The feed is actually made to enter about a meter below the surface of the tank level thushelping to minimise turbulence.

The feed well diameters are between 1 and 1.2 m with lengths of 1.2 to 5 m. Tank sizesvary according to feed characteristics and the sedimentation time. Manufacturers such asDorr-Oliver-Eimco, [3] have suggested that the water depth should be between 3.0 and 3.6 mand the feed well size about 25% of the basin area.

The rake-drives in bridge clarifiers are either centre driven (as shown in the Fig. 13.1)where the motor is mounted on a support plate or are peripheral driven. When the sludge istoo thick and the rakes struggle to move or in extreme cases cease altogether, the rakes aredesigned to rise either mechanically of pneumatically. Usually the torque on the rakes ismonitored and the rakes rise automatically at a fixed torque level. This precautionaryprocedure is generally attached to thickeners of diameter greater than 10 meters. Theallowable torque is about 5-30 times greater than normal operating torque [1].

A recent innovation is the Dorr-Oliver Eimco E-Cat thickeners which has dispensed withthe rakes and introduced clarifying cylinders through which the suspension passes to producethe clear over flow (Fig. 13.3). These thickeners are designed for rapid sedimentation by theuse of flocculants. The clarified slurry then passes through filters producing a clear overflow.

Page 4: Chapter 13. Solid - Liquid Separation

404404

feed

feed mix

overflow

cones

clarifyingcylinders

dewatering cones

underflow

Fig. 13.3. Clarifier/Thickener, Dorr Oliver Eimco [4].

13.2. Thickener Design-Batch ProcessThickeners have been designed using the basic laws of sedimentation. Empirical methods

devised by manufacturers are also used for rapid work. For designing , the chief criterion is todetermine the relation between the settling velocity and the dimensions of the vessel to beused for each particular slurry. The settling velocity for a particular slurry can be easilydetermined in the laboratory by using small-scale tests. The tests consist of determining thedownward movement of the boundary of the clear liquid and the suspension. It has beenfound that this rate is initially constant but the rate decreased as the particles slowly settled tothe bottom and the interface met the sludge zone. This can easily be visualised from Fig. 13.4where the progressively increasing concentration with depth is shown. It is obvious that thedeeper the vessel and longer the time given for settling, the clearer will be the supernatantliquid and the thicker will be the sludge.

The decrease in the settling rate is due to hindrance by increased crowding of the particlesas they settle and collect at the bottom of the vessel. At the sludge-forming layer, the particlespack down by displacing the liquid in between. In so doing, the clear liquid level rises. Theseconsiderations apply both to batch and continuous processes, with the difference that in thecontinuous process a balance between the flow rate of the overflow stream and the removalrate of the sludge has to be maintained.

Page 5: Chapter 13. Solid - Liquid Separation

405

.

405

Clear fluid

time

Fig. 13.4. Sedimentation in a thickener.

These considerations originally used by Coe and Clevenger [5] are now in use extensively.The quantitative basis for designing the thickener area assumes that:

1. Settling rate was a function of concentration,2. The volume rate of discharge of the clear supernatant liquid was equal to the difference

of the rate of feed of the slurry minus the rate of removal of the thickened layer.

For determining the thickener area, Coe and Clevenger assumed that the liquid movingupwards is always greater than the movement downwards. The mass of liquid flowingupwards is given by:

t/h (13.1)

where FD

= the feed mass ratio (liquid/solids, also known as the feed dilution),= discharge mass ratio (liquid/solid) and= Feed capacity by mass, t/h

At equilibrium, the upward velocity of liquid equals the downward velocity of the solids.Thus if vs is the velocity of sedimentation, A the cross-sectional area of the tank, in m2, andPL the specific gravity of the liquid, then at equilibrium:

F-D(13.2)

F-Dhence, A = | I Q

VSPLM{F) (13.3)

In practice, to determine the design value of the thickener area, a number of laboratorysedimentation tests are run using 2 litre cylinders and determining the value of vs for a range

Page 6: Chapter 13. Solid - Liquid Separation

406406

of F values. The maximum value of A is taken as the design cross-sectional area of thethickener tank.

Dahlstrom and Fitch [2] has analysed each of the settling zones and arrived at a practicalexpression similar to the expression of Coe and Clevenger for sizing a thickener. Consideringthat the flow rate in the clear zone should be less than the settling rate of the smallest particlethat has to he removed by settling, they derived the velocity of sedimentation as:

vs = ^ Q (13.4)

This equation is similar to Eq. (13.2) by Coe and Clevenger. Dahlstrom and Fitch [2]suggested that the actual sedimentation rate must be multiplied by the area! efficiency factor,AEF to obtain a realistic value. The area! factor is a function of the tank dimensions (heightand diameter) and ranges between 0.20 and 0.25.

Eqs. (13.3) and (13.4) are extensively used to determine the cross-sectional areas of tanks.The laboratory estimations are performed at different concentrations of F and D and thelargest value of A is taken as the designed size of the tank as in the Coe and Clevengermethod. For practical purposes they suggest a scale-up factor of 1.25 - 1.5 for thickener unitsless than 15.2 m in diameter and 1.3 — 1.5 for units greater than 15.2 m in diameter.

13.3. Thickener Design-Continuous ThickenersFor designing continuous thickeners, the three most important parameters that need to beestablished are:

1. Cross sectional area of the tank2. Depth of thickened layer3. Depth of the clarifying zone

Other factors include discharge slurry properties, such as liquid/solid ratio, viscosity andthe characteristics of pumping.

13.3.1. Estimation of Cross-Sectional Area of TankCoe and Clevenger's equation fails to accurately estimate the cross-sectional area of the tankwhen the slurry is treated with a flocculating agent. In such cases the mathematical approachof Kynch [6] as applied by Talmage and Fitch [7] is more suitable. A particular advantage isthat while several determinations of settling velocities, vs, are required by Coe andClevenger's method, a single estimation is sufficient when analysis of the sedimentationcurve is made. To apply Kynch's method the following assumptions are made:

1. The concentration of particles in any horizontal plane is uniform,2. Differential settling due to differences in shape, size or composition of mineral

particles do not take place,3. The sedimentation velocity is a function of concentration and tends to zero at a

concentration equivalent to the sediment layer at the bottom of the container,4. The wall effect is negligible.

A single laboratory test therefore involves the suspension of a slurry in a 2 litre talltransparent cylinder and measuring the clear fluid interface with the slurry at different times

Page 7: Chapter 13. Solid - Liquid Separation

407407

till the level falls and all particles settle at the bottom as sludge. Where the sedimentation rateis very slow or the supernatant liquid remain turbid and unclear, flocculants are added.

If required, rakes are introduced to break up agglomerated particles. A typicalsedimentation curve indicating the height of the interface with time and the structure of theslurry in the cylinder is shown in Fig. 13.5, It can be seen that at the initial stages, the rate offall of the interface is nearly constant. When the settling rate of the bulk of the slurrydiminishes, (as seen in cylinder 4), the clear zone-sludge interface merges and the curve thenflattens out. At this stage, further lowering of the clear level interface can take place by theexpulsion of water between the particles in the sludge. Fig.13.5 shows that at time t = 0, theheight of the interface is Ho. As it is assumed that the concentration of slurry is uniformacross the cross-section of the tube, at any height, Hi, the concentration of the sludge will bethe same across the settling tube.

For a dispersed slurry, the solids start settling at a uniform velocity which is a function ofthe local solids concentration [5,6]. As the settled solids build up at the bottom of thecontainer, the boundary between the settling solids and slurry of the initial concentrationstarts to rise in the slurry as indicated in Fig. 13.5. Zones of intermediate concentrationbetween the initial and final concentrations will move upwards from the bottom at a raterelated to the concentration of solids in that zone. When the rising and settling zones meet,the settling slows and is controlled by the extraction of retained water from the solids as itgoes through compaction.

The rise velocity of the zone of concentration C, from the bottom of the cylinder to theinterface of the settling mudline, vR, given by;

dHdt

d\j/

dC(13.5)

and is represented by the line OY in Fig. 13.5. y is the settling flux, kg/m2/s.

Time (s)

Fig. 13.5. Settling curve -Kynch's interpretation.

Page 8: Chapter 13. Solid - Liquid Separation

408408

If line OY represents the initial uniform concentration Co then higher concentrationsresulting from settling solids at the bottom of the cylinder are represented by lines of lowerslope, OYj (for intermediate concentration) until the maximum concentration of the settledsolids is reached and represented by CMAX and line OYMAX- Any line parallel to these willrepresent the rise velocity of zones of the same concentration, C, so line H1Y1 will represent azone of concentration Co which originates from height Hi in the slurry propagating upwardsand reaching the mudline interface at Yi after time ti.

Since the sedimentation rate is dependant on concentration only, until the zone of initialconcentration from the cylinder bottom reaches the interface, the sedimentation rate of theinterface will be constant and hence the rate vs0 will be represented by a straight line, HoY.

According to Kynch, if a tangent is drawn to the settling curve at point Yt, the slope, a,corresponds to the settling velocity,vst, of the layer or zone of concentration Q just below thesettling interface. The intercept of the tangent on the Y-axis, Ht, corresponds to the height ofslurry of uniform concentration equal to Ct. Then by a mass balance:

HtCt = H0C0 (13.6)

for a cylinder of constant cross-sectional area. Consequently, a plot of settling rate versusconcentration can be constructed from a single settling curve.

Kynch's theory has been tested experimentally on many occasions and found to hold forthe batch settling of equi-sized rigid spheres in water but deviates for flocculated suspensionsthat form compressive sediments [8].

Yalcin [9] reported the sedimentation curves of a copper-nickel tailings for several initialpercent solids. By constructing tangents to the low density pulp curve at different higherpercent solids, using the Kynch construction, estimates of the settling rates can be comparedto the actual measured sedimentation rates of these slurries. Fig. 13.6 shows such aconstruction on the settling curve of an unflocculated slurry having an initial concentration of5% solids. The estimates of the settling rates of the higher % solids are obtained from thetangents to the 5% sedimentation curve, intersecting the Y-axis at the mudline heightscorresponding to 15, 25, 35 and 45% solids. Fig. 13.7 shows the measured sedimentationvelocities versus the Kynch estimates from the slope of the tangents. The plot for theestimates from the 5% solids curve shows considerable difference from actual measuredvalues being higher than the estimates according to the Kynch theory. If the estimates areconstructed from the 15% solids curve for slurries of higher densities, Fig. 13.7 shows acloser correlation between the estimates and real sedimentation velocities. The estimatesconstructed from the 25% solids curve are similar to that obtained from the 15% solids curve.

Figs. 13.8 and 13.9 show similar constructions for a flocculated gold tailings at 20, 30 and40% solids. In this case, the Kynch estimates of the settling velocities are in close agreementwith the actual measured velocities.

Although the Kynch theory is not considered suitable for all mineral slurries, especiallyflocculated slurries, nevertheless it can give satisfactory results as indicated in Fig. 13.9. It isstill used for thickener design calculations [8],

Talmage and Fitch [7] showed that the settling velocity was related to the concentration.For a point on the settling curve of time t and height Ht, the equation is:

Ct = MC ° H ° (13.7)

H t + v s t t t

Page 9: Chapter 13. Solid - Liquid Separation

409

0.0

0.4

0 10 20 30 40 50 60 70 80 90 100

Time (min)

Mu

dlin

e h

eig

ht (

m)

5% solids

15%

25%

35%

45%

0.0000

0.0001

0.0002

0.0003

0.0004

0.0000 0.0001 0.0002 0.0003 0.0004

measured, m/s

Kyn

ch e

stim

atio

n, m

/s

based on 5% solids curve

15%

25%

Ho

H15

H25

H45

409

0.4

Ho

si"35

a>c

•o

H15

H-25

H45

0.0

I ? -•—.D

"—

—A

D

- • - 5 % solids-D-15%-A-25%-0-35%-A-45%

A

0

10 20 30 40 50 60 70 80

Time (min)

90 100

Fig. 13.6. Cu-Ni tailing sedimentation data replotted from Yalcin [9] with Kynch construction on the5% solids curve.

0.0004

(A 0.0003

4-1

ns. i 0.0002(A

U>

0.0001

0.0000

—•— based on 5% solids curve

- a - 1 5 %

- • - 25%

y

, ' ' m

yyss

*

y

yy

ys

J~l

y•

yy

y

s'

0.0000 0.0001 0.0002

measured, m/s

0.0003 0.0004

Fig. 13.7. Kynch estimated sedimentation rates compared to measured rates for different % solidslurries, (data from [9]).

Page 10: Chapter 13. Solid - Liquid Separation

410

0.0

0.3

0 16

Time (min)

Mu

dlin

e H

eig

ht(

m)

20% solids

30%

40%

0.0000

0.0001

0.0002

0.0003

0.0004

0.0005

0.0000 0.0001 0.0002 0.0003 0.0004 0.0005

measured, m/s

Kyn

ch e

stim

atio

n, m

/s

H40

Ho

H30

H1

H

t

410

0 tTime (min)

Fig. 13.8. Sedimentation curves of a flocculated gold tailing with Kynch construction on the 20%solids curve.

0.0005

0.0004

I.2 0.0003is

0.0002

0.0001

0.00000.0000 0.0001 0.0002 0.0003 0.0004 0.0005

measured, m/s

Fig. 13.9. Kynch estimated sedimentation rates compared to measured rates for a flocculated goldtailing.

Page 11: Chapter 13. Solid - Liquid Separation

411

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0 2 4 6 8 10 12 14 16

Time (min)

Mu

dlin

e H

eig

ht (

m)

Critical pointH

HU(1)

tU(1)

HU(2)

tU(2)

411

In a batch settling test, the mass of solids in the test cylinder is given by CoHoA. If thetime taken for all the solids to settle past a layer of concentration C is tu then CoHoA/turepresents the quantity of solids that can be brought through the concentration layer per unittime. The area of thickener required to settle 1 tonne of solid per unit time is then given by:

A = m2/t/h (13.8)

The time tu is obtained by drawing a line from mudline height H, corresponding to theconcentration C, at a tangent to the settling curve. The intersection of this tangent with themudline corresponding to the underflow concentration is the value tu on the time axis. This isillustrated in Fig. 13.10.

The maximum thickener area requirement will occur when the tangent is drawn throughthe compression point on the sedimentation curve since this tangent will give the highestvalue of tu in the free settling range which, according to Talmage and Fitch [6] is the zonedetermining the unit area. When the line, corresponding to Hu, intersects the settling curveabove the compression point, the value of tu corresponding to the maximum thickener areawill be the point of intersection with the settling curve, shown as tu(i) in Fig. 13.10.

Fig. 13.5 shows that a near steady concentration is reached at about YMAX. Assuming thisto be an equilibrium state, a material balance of solid and liquid can be made. Svarovsky [10]expressed the area of the tank in terms of the overflow rate. From a material balance, the

0.30

0.25

-g- 0.20

I 0.15

I H3S 0.10

0.05

0.00

\

\ V Critical point

•—

— — - —j

•—

HU(2)

tu(l) 10

Time(min)14 tu ( 2 ) 16

Fig. 13.10. Talmage and Fitch construction for determination of tu; tu<n is the value where Hu liesabove the critical point and tup; is the value where Hu lies below the critical point.

Page 12: Chapter 13. Solid - Liquid Separation

412412

ratio of the overflow rate to feed rate can be determined in terms of feed and underflowconcentrations as:

(13.9)

where Qv(o); Qv(F) = volumetric overflow rate and feed rate, m3/s,Cu, CF = concentrations of the underflow and feed respectively, expressed

as the mass of solid/volume of slurry, kg/m3.

The overflow rate can be easily measured providing the velocity of the underflow is measuredand is constant. The overflow rate for the system is related to the liquid rise velocity by theequation:

Qv(o)=Av(o) (13.10)

where V(o) = liquid rise velocity or overflow velocity.

Substituting this value in Eq. (13.9) and simplifying, the area of the tank may be expressedas:

A =QV(F) C —CJX5 (13.11)

Eq. (13.11) gives the area of the cross-section of the tank at a known feed rate, knownconcentrations of feed and underflow and liquid rise velocity.

13,3.2. Determination of Critical PointThe Talmage and Fitch and other methods of thickener design require the determination ofthe critical point on the sedimentation curve. As the solids settle they pass from free settlingto hindered settling to compression conditions. At each of these transitions there is adiscontinuity in the sedimentation curve, hi the free settling region, the settling rate isconstant and representative of the initial solids concentration. When the solids concentrationincreases to the point where the near neighbours start to influence the settling rate of theparticles, the settling rate slows and is affected by the concentration of nearest neighbours.The slurry is in a hindered settling condition and the decrease in settling rate is referred to asthe first falling rate. The settling behaviour becomes non-linear, inversely proportional to thesolids concentration. When the solid concentration increases to the extent where the solidstouch, settling ceases and further consolidation of the solids occurs by compression. Thisfurther drop in sedimentation rate is referred to as the second falling rate and a seconddiscontinuity in the settling curve will occur (Fig. 13.11). The end of hindered settling andthe start of the compression zone is referred to as the compression or critical point. Thisdiscontinuity in the settling curve is not always readily discernable and some procedures havebeen suggested to try and locate the compression point on the settling curve.

These procedures try to replot the data to accentuate the discontinuity in the settlingbehaviour, making some assumption as to the shape of the curved sections:

Page 13: Chapter 13. Solid - Liquid Separation

413413

1. replot on log-log axes. The upper and lower sections of the curve generallyapproximate straight lines which intersect at the critical point.

2. draw tangents to the sedimentation curve at both ends and bisect the angle formed. Theline bisecting the angle often intersects the sedimentation curve close to the criticalpoint. This however will change as the scale on the axes changes [11]. See Fig. 13.12.

3. The mudline height at the critical point, He, can be obtained from a plot of dH/dt versustime as indicated by Mondal and Majumdar [12], Fig. 13.13. Again, the change inslope at the critical point should be evident. Barnea [13] also plots the differential:

H ° - ' ~ H » 1 versus H ° ~ H "HH

where n is one data reading on the sedimentation curve. Hn is defined as the mean ofHn., andHn+i. See Fig. 13.14.

4. Plot the distance log (H - Ho,) versus time where Hm is the final (equilibrium)sedimentation height (at infinite time). If the curved sections of the sedimentationcurve are represented by an inverse exponential function, then plotting the log of theheight vs. linear time will give a straight line and a change in slope will occur at thesedimentation discontinuity [14]. See Fig. 13.15.

5. Dahlstrom and Fitch [1] assumes the start of the compression of the sediment takesplace at a mudline height halfway between the initial height and the final height of thesediment; ie. Hc = Ffo-FLa> For the sedimentation curve in Fig. 13.11, this gives anestimate of the critical point at 56 s. This point appears to be the start of the hinderedsettling zone rather than the compression zone. This method and the bisected anglemethod are not recommended.

From the sedimentation curve given in Fig. 13.11, the critical point estimation by thevarious methods is given in Table 13.1

Table 13.1Critical point estimates, tc, from the sedimentation data in Fig. 13.11 by various methods.

MethodLog-log plotBisected angleRobertsMondal and MajumdarBarneaDahlstrom and Fitch

Hindered settling, s-58

-608040-

Critical point, s-41014559058046056

The Roberts, Barnea and Mondal methods appear to give similar estimates; The log-log plotgives little deviation from a straight line and the critical point is less easily identified for thisdata. On the Barnea plot, for this data, it is also difficult to identify the critical point.

Page 14: Chapter 13. Solid - Liquid Separation

414

0

50

100

150

200

250

300

0 200 400 600 800 1000 1200

Time (s)

Mu

dli

ne

He

igh

t (m

m)

0

50

100

150

200

250

300

0 200 400 600 800 1000 1200

Time (s)

Mu

dlin

e H

eig

ht

(mm

)

First falling rate Second falling rate section

First discontinuity

Compression or critical point

H

414

300

250 -

50 -

0

Firstdiscontinuity

Compression or criticalpoint

Second falling rate section

200 400 1000 1200600 800

Time (s)

Fig. 13.11 Batch settling tests showing discontinuities at the transition from free settling to hinderedsettling and to compression.

300

200 400 600 800 1000 1200

Time (s)

Fig. 13.12. Rough location of the critical point by bisecting the angle formed by two tangents to theextremities of the sedimentation curve [11]. Critical point at 145 s.

Page 15: Chapter 13. Solid - Liquid Separation

415

0.0

0.2

0.4

0.6

0.8

0 100 200 300 400 500 600 700 800 900 1000 1100 1200

Time (s)

dH

/dt

0.01

0.10

1.00

10.00

0.1 1.0

dHn

(dH

/dt) n Compression

point

Constant rate

Hindered settling

Compression

460 s

40 s

Critical point

.

415

0 100 200 300 400 500 600 700 800 900 1000 1100 1200

Time(s)

Fig. 13..13. Plot of change in slope versus time [12]. Critical point at 580 s.

10.00

1.00 -

5X•a

0.10 -

0.01

Compressionpoint

460 s

Hindered settling

Compression

dHn

40 s

Constantrate

1.0

H —H H —HFig. 13.14. Barnea (1977) plot where (dH/dtV = — ^ ^ anddHn=—= *

tn+l~tn-l

Page 16: Chapter 13. Solid - Liquid Separation

416

10

100

1000

0 200 400 600 800 1000 1200

Time, s

H -

Hin

f Critical point

416

1000

100 -

200 400 600 800

Time, s1000 1200

Fig. 13.15. Plot of log (H - H») versus time to accentuate the change in slope at the twodiscontinuities in the sedimentation curve [9,14]. Critical point at 590 s.

13.3.3. Determination of Settling FluxInstead of using the concentrations of streams it is more convenient to express Eq. (13.10) interms of the mass of sedimentation per unit area, known as the settling flux (y) and given by:

V = C t vs

Substituting the value of Ct from Eq. (13.6):

COHO• v e

(13.12)

(13.13)

Thus on the sedimentation curve, if tangents are drawn at several points, then the slopesand intercept with the H axis gives the corresponding flux-concentration curve as shown inFig. 13.16.

The settling flux curve can only be reconstructed from the sedimentation curve for theconditions where the gradient to the flux curve is decreasing with increasing concentration.That is, conditions which are found in the normal sedimentation test and are represented byconcentrations higher than the point of inflection on the flux curve.

A non-graphical approach was proposed by Yalcin [9] and is dependant on a power lawrelationship between the slurry % solids and sedimentation time being established. For the20% solids slurry in Fig. 13.8 the underflow % solids corresponding to each mudline height isplotted against time in Fig. 13.17. From Fig. 13.17:

%S = k f = 31.06101961 (13.14)

Page 17: Chapter 13. Solid - Liquid Separation

417

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 100 200 300 400 500 600 700 800

Concentration, kg/m3

Set

tlin

g F

lux

, kg

/m2 s

graphically from sedimentation curve

Yalcin calculation

y = 31.06x0.1951

R2 = 0.9643

10

100

0 1 10 100

Time, min

% s

olid

s o

f U

/F

417

1

0.9

0.8

J" 0.7

I 1 0.6

3 0.5U.g1 0.4

m 0.3

0.2

0.1

0

//

I

II

j

/ x

' \

\ 9\ \\ \

—•—graphically from sedimentation cune

—o— Yalcin calculation

>O <Xi

100 200 300 400 500 600 700 800

Concentration, kg/m

Fig. 13.16. Settling flux calculated from tangents to the 20% solids pulp curve in Fig. 13.8.

100

at•aoin

10

/ = 31.06x°

R2 = 0.96

X

1951

43

*f

»

1 10

Time, min

100

Fig. 13.17. Relationship between underflow % solids and sedimentation time for a gold tailing at 20%solids initial concentration.

Page 18: Chapter 13. Solid - Liquid Separation

418418

At any mudline height, H and time tt on the sedimentation curve (Fig. 13.5), the % solids(%S) corresponding to the slurry near the interface having a settling rate of vs is given by theintercept of the tangent to the curve, Ht. The % solids at Ht is given by [9]:

1 _ 1 (H 0 -H t )p w (13 15)%St %SO 100CoHo

where %St = % solids at time t and%S 0 = initial % solids (at t = 0).

Substituting for %S from Eq. (13.14) into Eq. (13.15) and rearranging gives:

- + HJ1-—-9- (13.16)pwk

Then, differentiating Eq. (13.16) with respect to time:

dH 100nHoC (n+1)

at P ^

Thus for any sedimentation time, tt, the value of H is obtained from Eq. (13.16), vs fromEq. (13.17) and Ht from Fig. 13.5:

H, = H + vstt (13.18)

The % solids below Ht (%St) is then obtained from Eq. (13.15). Thus a series of vs valuescan be obtained for different underflow % solids, which are related to the solids concentrationby the equation:

C = f^ ^ kg/m3 (13.19)%S flOO%SN

Ps ^ Pw

Points on the settling flux curve can thus be evaluated as shown in Fig. 13.16 for a goldtailing sample. Good agreement is found between the calculated flux and the graphicallyestimated flux values.

A second graphical method was advocated by Jernqvist [15,16] and described by Kelly andSpottiswood [17] and is based on the laboratory determination of batch settling in a tallcylinder. The time axis of the height-time curve is reversed and the following steps taken:

Step 1: The Y-axis is extended to form the Y-axis of the solid flux-concentrate curve.

Step 2: Draw a horizontal line to form the x-axis of the flux curve.

Step 3: Let the initial concentration of the slurry be Co located on the C axis. Through Codraw a vertical line.

Page 19: Chapter 13. Solid - Liquid Separation

419

Concentration, kg/m3

m/gk , x

ulF

gniltte

S2 s

Sedimentation curve

Yalcin calculation

graphically from sedimentation curve

Sedimentation time, min HO

O

F

α1

α1

1

CO

H1

O’

C1

2

Mu

dili

ne

hei

gh

t, m

419

Step 4: Through Ho draw a horizontal line, indicating concentration (could be the C axis).

Step 5: Draw a tangent to the sedimentation curve, say a\. Through the origin of the fluxcurve draw line OF parallel to the tangent a i .

Step 6: From the intersection of ai with the y-axis, draw a horizontal line to cut the verticalline through Co at Hi. Join origin O' to H] and proceed to cut the x-axis at Cj. Drawa vertical line through Ci to cut Hie OF line at 1. This point of intersection is a pointon the flux-concentration curve.

Step 7: Repeating steps 5 and 6, several points can be obtained which on connecting, providethe \|/ - concentrate curve.

The construction of the (\|/-C) is illustrated in Fig. 13.18 and compares well with othertechniques. The scale used for the flux axis must be consistent with the dimensions of theother 3 axes. This procedure also assumes that the sediment is not compressible.

E

'a

-Sedimentation curve

-Yalcin calculation

-graphically from sedimentation curve

Sedimentation time, min 'Concentration, kg/m

Fig. 13.18. Jernqvist method for construction of the flux curve from a sedimentation curve [15,16].

Page 20: Chapter 13. Solid - Liquid Separation

420420

The thickener area is then obtained from the settling flux value according to the equation:

A = -^Ha (13.20)V

For smooth operation of a thickener and to achieve the required properties of the productstreams it is imperative to know:

1. the maximum allowable concentration of the underflow,2. the optimum conditions of the overflow and the concentration of the feed slurry.

This information is derived from the flux-concentration and flux-time curves,hi continuous sedimentation process two forces are simultaneously in operation:

1. Sedimentation flux (\|/s),2. Withdrawal flux (\|/w ).

Thus the total flux is given by:

WT= VS + Vw (13.21)

The two flux-concentration curves and the total flux curve resulting from the combinationof each set of data are shown in Fig. 13.19. The combined flux curve shows a minimumvalue at some critical concentration, CCRIT- The corresponding minimum flux, \\icmi, is themaximum that the thickener can handle. At a concentration less than CCRJT> solids enter thesludge layer faster than can leave via the underflow and hence the concentration, C, increasesup to CCRIT- At concentrations greater than CCRIT> solids leave the underflow faster than isentering sludge layer and hence C drops to CCRIT-

Coe and Clevenger [5] suggested that at this critical concentration, the flux of solids to theunderflow of a continuous thickener would be a minimum and the critical flux, V|/CRIT, is themaximum flux that can flow through the thickener into the underflow at steady state. Thiscritical flux is rate determining and will determine the thickener area for a given feed rate andunderflow density, according to Eq. (13.20).

Yoshioka et al. [18] obtained the critical flux from the settling flux-concentration curve byconstructing a tangent to the curve passing through the underflow concentration Cu on the x-axis (Fig. 13.19). The tangent is called the Operating Line and the intercept on the flux axisis the critical flux, and the thickener area is:

A = ^ L = Qv(F) C ° (13.22)

It should be noted that if \|/o > YCRIT then the thickener is overloaded and corrective actionis necessary.

Oltmann [19] suggested a simple empirical approach to the critical flux determination andhence thickener area, hi cases where the settling rate at the beginning of the sedimentationtest is non-linear due to turbulence resulting from mixing, an extrapolation of the linearsettling rate section to the horizontal extension of Ho will give the start time ta (Fig. 13.20).

Page 21: Chapter 13. Solid - Liquid Separation

421

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 100 200 300 400 500 600 700 800

Concentration, kg/m3

Set

tlin

g fl

ux,

kg

/m2 .s

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0 2 4 6 8 10 12 14 16

Time (min)

Mu

dlin

e H

eig

ht (

m)

CCRIT

ΨCRIT

CU

HU

tU ta

HO

Critical point

421

0 100 200 300 400 500 C u 600

Concentration, kg/m3

Fig. 13.19. Flux curves for a continuous thickener [17].

700 800

0.30

0.25

•=• 0.20

si'55

0.15

C

•c3

0.10

0.05

0.00

• A — Ho

\

VN \ Critic

s

——-».

v - -

al point

m- m-

Hu• -

0 ta6 tu 8 10 12 14 16

Time(min)

Fig. 13.20. Oltmann construction to determine the critical settling flux.

Page 22: Chapter 13. Solid - Liquid Separation

422422

A line drawn from the point (ta,Ho) through the critical point on the settling curve tointersect the underflow mudline Hu will determine the time tu. The detention time is thengiven by tn = (tu - ta) and the critical flux is given by;

Vour = ^ f (13-23)t JZ

The factor 1.2 is used to give a 20% safety factor. The thickener area is then given by Eq.(13.22).

13.3.4. Long Tube Method for Estimating Thickener DimensionsSome pulps containing fines or materials of colloidal dimension do not settle with a clearinterface. Some do not tend to settle at all. For such suspensions the long tube method ofdetermining the dimension of thickeners or clarifiers are useful. The test consists ofdetermining the rate of rise of a fluid with different detention times. The rate of rise is relatedto the concentration of solids in the overflow.

The test is carried out in a long plastic or glass tube about 3-4 meters in length (height) andabout 0.075 m in diameter. Along its length are a number of sampling points. The firstsampling point is about 100 mm from the top of the tube with the remaining sampling pointsevery 200-300 mm. The slurry mixed with an appropriate flocculant is charged and the topmost level is established at the top most outlet by opening the value to allow any excess slurryto drain off. The pulp is allowed to settle for a time till a visibly clear supernatant liquid levelis seen. Pulp samples are then rapidly taken from all the sampling points, starting from thetop and the solid concentrations in each sample is determined. The procedure is repeated 4-5times for different settling times depending on the expected operating conditions of thethickener/clarifier. The results are tabulated for cumulative depth (H) and solid concentration.

The test results indicate that the clarification zone clarity is a function of overflow rate (vo)and the detention time. The overflow rate is given by:

vo = f = ^ (13.24)

where H = the cumulative height (or depth),t = time,Qv(O)= volumetric flowrate in the overflow andA = area.

The detention time is related to the feed rate, depth and area as:

AHQv(F)

where Qv(F) = the volumetric feed rate.

(13.25)

Page 23: Chapter 13. Solid - Liquid Separation

423423

The overflow rate determined by the long tube test is the ideal overflow rate, vo(i), that isrequired for a certain feed concentration and therefore Eq. (13.24) can be rewritten as

Vow = Y (13-26)

hi the operation of a thickener or clarifier, generally a volume overflow rate, Qv(O)» isspecified (or chosen) for a particular feed solid concentration, CF. At this overflow amaximum solid concentration Co is tolerated. From the data it can be seen that:

™ , 4 overflow rate required Qvroi ,,-,-,-^The pool area, A = = w (13.27)

ideal overflow v0(i)

The pool volume V = Overflow rate required x time =Qv(O)t, and (13.28)

^ i J *. TT P°°l volume V „ . __.The pool depth H = = — (13.29)

pool area A

The units for Qv(O) is m3/h, Vo(i) is m/h, t is h and the overflow solids concentrationtolerable is in ppm.

hi practice the overflow concentration is less than that obtained by the long tubeexperiments. To account for this discrepancy, Perry and Chilton [20] plot suspended solidsconcentration C against the rise rate and graphically integrate between C=0 and C=Cc. Theresulting value is then subtracted from the observed suspended solids concentration at thechosen rise rate.

Osbome [21] however suggests the use of a suspensoid factor, /(s), to correct the error.The corrected pool area would be given by:

The pool area A = - ^ i . - i - (13.30)

where /(s) = 0.7.

13.3.5. Estimating Height (depth) of the Compression LayerThe approximate depth of the thickened sludge layer can be readily determined by the methodoutlined by Osborne [21]. The height, H, of the layer would depend on the total volume ofthe solids and liquid in the compression zone and inversely as the area of the vessel. That is:

H = ^ (13.31)

where Vc = the total volume of the liquid and the solids in the compression layer.

Vc can be determined if the average concentration of solids in the layer is known. Hence,if the average concentration of solids in the compression zone is Cc (mass solid/volume of the

Page 24: Chapter 13. Solid - Liquid Separation

424

0

50

100

150

200

250

300

0 100 200 300 400 500 600 700 800 900 1000

Time (s)

Mu

dlin

e H

eig

ht (

mm

)

equal areas

B'

B

Hu

H

Ho

t1 t2

2)H(HO –

424

compression zone), then the mass of liquid in the zone per unit volume will be (pp - Cc) andthe liquid/solid ratio in the compression zone will be (pP - Cc)/Cc. The depth of thecompression zone will depend on the amount of sludge deposited and that would depend onthe retention time. The retention time is a function of the rate of discharge of the underflowand the underflow concentration. Thus if the feed rate is expressed as Qv(F), the feedconcentration as CF and the retention time in the compression zone as to, then the volume ofsolids plus the volume of liquid in the compression zone would be:

Vc =V(F)

Ps

Qv(F) ^F *D [ Pp ~

PL

(13.32)

Thus if A is the cross-sectional area of the thickener, then the height of the compressionzone, He, will be:

Qv(F) C F *D L + Ps

A p s I(13.33)

The units are: Qv(F) = m /S, to = s, A = m , C = kg/m and p = kg/m .

Dahlstrom MethodDahlstrom and Fitch [1] obtained the compression zone volume from the settling curve asshown in Fig. 13.21.

100 200 300 400 500 600 700 t2 800 900 1000

Fig. 13.21. Settling curve showing Dahlstrom's construction for compression zone height [1].

Page 25: Chapter 13. Solid - Liquid Separation

425425

The settling test is carried out in a 2 L cylinder with a picket rake and run for 24 hours toobtain the final sediment height, Hoc. It is assumed that the start of the compression of thesediment occurs at the point B, which is located at the height (Ho- Hx)/2 and time ti. Theheight corresponding to the underflow solids concentration is given by Hu which intersectsthe sedimentation curve at time t2. The detention time, to, (residence time) in thecompression zone required to achieve the desired underflow density is then given by t2 - ti.

The compression zone volume is calculated from the average solids concentration in thecompression zone obtained by integrating the area under the curve from ti to t2.Alternatively, a line bisecting the area under the curve from ti to t2 will intersect thesedimentation curve at B \ The sediment height at this point can be converted to obtain theaverage concentration, Cc in kg of solid/m3 of pulp. The compression zone volume is thenobtained from:

where SF = a scale factor, normally taken as 1.75 andQM(F) - the mass flowrate in the thickener feed.

The compression zone height is then obtained by dividing the compression volume by thethickener area. Empirically it is found that if the calculated compression height is greaterthan 1 meter then the thickener area should be increased or the thickener underflow densityreduced to maintain a maximum compression height of 1 meter [1,22].

13.3.6. Estimating the Depth of the Clarifying ZoneFor estimating the depth of the clarifying zone, a tall cylinder is again taken and filled withslurry. Sample points are inserted every 200 mm and the clear level recording startedimmediately and continued at regular intervals. If a clear level is not obtained, a flocculant isadded and a height-concentration curve drawn. The overflow rate, Vo, is determined as [11]:

_ H 0 F _ Qv(Q)V

That is: H0F = **m± (13.35)A

where vo = overflow rate or overflow volume flux expressed as m/s andHQF = height of the clarification zone.

13.3.7. Estimating the Retention TimeDahlstrom and Fitch [1] described a procedure for determining the retention time ofcontinuously operating thickeners. The overflow rate was related to the retention time by therelation :

tD = ^f~ (13.36)

Page 26: Chapter 13. Solid - Liquid Separation

426426

where Qv(o) = overflow rate, m3/h,A = cross sectional area of the tank, m 2 , andAEF = areal efficiency factor.

The overflow flow rate can be measuring by determining the overflow velocity, vo, and usingthe relationship:

AEF (13.37)

Dahlstrom and Fitch reported areal efficiency values between 0.1 and 0.6, depending on theheight to diameter of the thickener and feed well, with typical values of 0.20 - 0.25. Theoverflow velocity must be less than the settling velocity of the smallest particle to beremoved. This maximum velocity for thickeners is generally 0.00034 - 0.0020 m/s.

Examples 13.1-13.2 illustrate the use of the above concepts.

Example 13.1The volume rate of flow of slurry from a dust catcher was 3.7 m3/min. The concentration ofslurry (by mass) was 10% and the specific gravity of the solid is 2.75. The slurry is to bethickened to produce a sludge containing 47% minimum solids by mass in a continuousthickener. Settling tests on the sample gave the following data:

Rate of settling, m/min 0.72 0.36 0.24 0.051 0.01Concentration, % solids by mass 10 15 25 35 45

Estimate the cross-sectional area that will separate 1000 tonnes of solids per hour.

SolutionTake as the basis, 1 kg solid and time in seconds.

Step 1Solid to be separated/min = 1000 x 1000/60 = 16,667 kg = 277.8 kg/sThe underflow contains 47% solids and 53 % water, hence the underflow water/solids ratio =53/47=1.12.

Step 2Estimate the water to underflow from the given data as shown in the table below.The table shows that the maximum flowrate in the overflow stream, per unit sedimentationrate = 858.2 s/m.According to the given conditions, the sludge contains 277.8 kg/s (QM(F))-Assuming a density of water = 1000 kg/m3, the thickener area, from Eq. (13.3) = 858.2 x277.8/1000 = 238.4 m2.

Page 27: Chapter 13. Solid - Liquid Separation

427427

.0.5.Hence the diameter of the thickener = (238.4/3.14)"3 x 2 = 17.4 m.

solids

(1)

Feed rate0 / T7

water (Mw/Ms)

(2) (3)

Water distributionD = U/F(Mw/Ms)

(4)

0/F = F-D(Mw/Ms)

(5)=(3)-(4)

Sedimentation rate, vsm/min

(6)

m/s

(7)

(F-D)/vs

s/m

(8)=(5)/(7)1015253545

9085756555

9.005.673.001.861.22

1.1281.1281.1281.1281.128

7.874.541.870.730.09

0.720.360.240.0510.01

0.01200.00600.00400.00090.0002

656.0756.5468.1858.2567.4

Example 13.2A slurry containing 300 kg solid per cubic meter of slurry is to be dewatered in a thickenersuch that the underflow will contain 750 kg/m3. The feed rate to the thickener was expectedto be 0.5 m3/min. A batch settling test of the slurry gave the following results:

Solid concentration,C, kg/m3

Settling velocity, Vs,mm/min

300.0

26.667

362.3

15.588

497.4

7.148

774.2

1.610

960.0

0.455

1010.5

0.271

1078.7

0.111

1128.1

0.068

Estimate the maximum area and diameter of the thickener.

SolutionSteplDetermine i|/ from vs and C values using the expression:

\\i = vs . C (kg/m2.s)

This can be determined for different velocities as illustrated in the table below:Settling test results from laboratory test

Settlingmm/min

26.66715.5887.1481.6100.4550.2710.1110.068

velocity, vsm/s x 10"°

444.0260.0119.026.8

7.584.521.851.13

Solid concentration, Ckg/m3

300362.3497.4774.2960.0

1010.51078.71128.1

Settling Flux, \\i = vs . Ckg/m2.s

0.1330.0940.0590.0210.0070.00460.00200.0013

Page 28: Chapter 13. Solid - Liquid Separation

428

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0 200 400 600 800 1000 1200

Solids concentration, kg/m3

velo

city

, kg

/m2 .s

CU

ψCRIT

428

Step 2Plot settling velocity against solid concentration as in Fig. 13.22. Since the underflow has tobe 750 kg/m3, draw a line, tangent to the curve and passing through 750 kg/m3 on the x-axis.This line cuts the y-axis at \|/CRIT = 0.17 kg/m2/s.

Step 3As a first approximation, using Eq. (13.22):

area A =(Qv(F) CO)A|>CRIT = 0.5 x300/(0.17x60)m2

= 14.7 m2

and diameter = 4.3 m

Considering a safety factor of 1.5, the practical diameter = 6.5 m

0.18

0.16

CM

°>

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0.00

^ —

\

\

VCRIT

\ \

\

V /Cn

/

200 400 600 800 1000 1200

Solids concentration, kg/m

Fig. 13.22. Settling flux vs solids concentration.

In the discussions and computations explored in examples 13.1 and 13.2, the effect ofdifferent particles sizes (and possibly density) has not been considered. The velocity ofdescent of different sized particles will obviously be different. In such a case thesedimentation profile will consist of more than three zones (Fig. 13.23) due to the upward

Page 29: Chapter 13. Solid - Liquid Separation

429429

Fig. 13.23. Sedimentation layers resulting from particles of different size and density

flow of the displaced liquid by the movement of the different size of particles. The lines ofdemarcation between these zones are not well defined and flux determinations are difficult.

Due to such difficulties, adjustments to experimentally computed design parameters havebeen published from time to time to yield realistic approximations of the different parameters[1,20]. These modifications are summarised in Table 13.2.

Table 13.2Multiplying factors for different thickener parameters [20].

Parameter Multiplying factorTank sizeSedimentation timeCross-section area of tank

Transition zone depthCompression zone depth

0.5 - 0.7 to rise rateratio of static detention time/volume efficiency1.2 for diameter > 30 m1.5 for diameter < 4.6 m1.1-1.25 for safetyAdd about 2 m1.75

The estimated area of the tank is increased by multiplying by a factor of 1.2 for tankdiameters greater than 30 meters and a factor of 1.25 for tanks with estimated diameters lessthan 5 meters [20].

Often the thickener area and depth are calculated by manufacturers from standard tablesestablished from a large number of field operations. However, no two circumstances are thesame and the following method adopted by Eimco [3] is of interest for rapid estimates andmay be accepted with reservations. The effective clarification area is obtained from:

AE =Average daily flow rate

Specified overflow rate(13.38)

Page 30: Chapter 13. Solid - Liquid Separation

430

0

2

4

6

8

10

12

14

16

18

0 50 100 150 200

Effective clarification area, m2

Tan

k d

iam

eter

,m2

430

18

16

14

2

I 10ECOT3 8

50 100 150 200

Effective clarification area, m

Fig. 13.24. Tank diameter selection [3].

where AE = the effective clarification area = tank area - feedwell area.

The average daily flow rate is in gallons per day or m3/h and the specified overflow rate isin gallons per day per ft2 or m3/h.m2. The relation between the effective clarification area andthe diameter of tank is given in Fig. 13.24.To calculate the required area for a thickener, the recommended expression is:

A =Daily solid load in kilograms

Floor Loading Rate(13.39)

where the solid load is in kg/day and the Floor Loading Rate is in kg/m2/day, obtained fromTable 13.3.

Table 13.3Thickener floor loading [3].

% Sludge in feed Floor Loading kg/m /day Typical % Solids in underflow106532

025

35-5075100

107.5173.3048.8729.3219.55

Page 31: Chapter 13. Solid - Liquid Separation

431431

13.4. Operation of thickenersThe operation of thickeners involves a delicate balance of the feed rate, the overflow rate andthe underflow withdrawal rate and is dependant on the concentration of the feed, overflowand underflow streams.

The feed stream generally enters the feed well at a speed of about 15 m/min but this woulddepend on its characteristics, such as concentration (liquid/solid ratio), particle size, particleshape and viscosity. The characteristics of the overflow and underflow streams depend on thesedimentation time and particle properties like, specific gravity, shape , size and wettability.If the particles are very small, the associated surface charge or zeta-potential is of importance.Flocculants play an important role in affecting the surface charge on particles and help toaccelerate or reduce the rate of sedimentation by dispersion or agglomeration.

Rakes help to increase the sedimentation rate and also break up large agglomerates. Therakes are operated between 8-18 m/min. To prevent damage to the rakes and torque metersthe recommended operation is to discharge the sludge at regular intervals at predetermined setconditions. It is necessary for the operator to detect the build up on the rakes and operate toavoid the jamming and seizure of the rakes. Usually the built-up mud tends to form islandswhich grows and develops a moment that could easily damage the rake mechanism. Duringnormal operation the rise rate varies from about 0.01—0.03 m /min/m of cross-sectional areaand the detention time is between 2-5 hours.

Some common operating parameters and cross-section of tank sizes for selectedmetallurgical operations are given in Table 13.4.

Table 13.4.Thickener and clarifier operating conditions [1,23].

Material Feed Underflow Area Overflow rate% solids % solids m2/tonne/day m3/h/m2

Copper concentrateIron ore (concentrate, coarse)Iron ore (concentrate, fine)Lead concentrateNickel carbonate ore

(acid leach residue)Uranium (acid leach residue)Iron making blast furnace flue 0.2-2.0 40-60 - 1.5-3.7

dustSteel making BOF flue dust 0.2-2.0 30-70 - 1.0-3.7

13.5. Thickeners in CircuitsThickeners used to produce low solid overflows (eg. about 1% solids), may be referred to asclarifiers. Both thickeners and clarifiers are extensively used in metallurgical operations fordewatering purposes. In processing gold, nickel, iron, copper ores etc. thickeners are used toproduce overflows suitable for use as process water in circuits such as flotation. The clearoverflow water is used for re-pulping the flue dusts or fine dust from precipitators. Thereforethe feed to thickeners vary considerably. A common arrangement is illustrated in Fig. 13.25

14-5025-4015-3020-2515-25

10-300.2-2.0

40-7560-7560-7060-8045-60

25-6540-60

0.2-2.00.02-0.10.15-0.40.5-1.00.3-0.5

0.02-1.0_

Page 32: Chapter 13. Solid - Liquid Separation

432432

Feed

Hydrocyclone

Overflow

Sludge

Fig, 13.25. Thickener arrangement.

Thickeners serve as classifiers when a near clear overflow is required. For exampleclarifiers used in iron blast furnace dust cleaning plant or electrostatic precipitator circuits arerequired to produce clean overflows as the water is for reuse and the sludge is for secondaryuse. In such cases the sludge is washed continuously by counter current decantation., wherethe underflow from a thickener/clarifier is pumped to the next thickener/clarifier (connectedin series) forming the feed to the second tank. A typical set up is illustrated in Fig. 13.13consisting of three units of thickeners/clarifiers.

Such setups are structured so that the overflow from one clarifier/ thickener flows bygravity to the adjacent clarifier. The sludge is usually pumped to the next clarifier. Make upwater is added at the third thickener.

make up water

Feed

Clarifiedoverflow Underflow sludge

U/F

Fig. 13.26. Thickeners in a counter-current decantation (CCD) arrangement.

Page 33: Chapter 13. Solid - Liquid Separation

433433

13.6. Problems

13.1.Settling tests in a cylindrical tube was performed on a slurry containing 300 ppm solids.After a detention time of 80 minutes, the overflow fluid was found to contain 10 ppm solidsThe overflow from the test data was found to be 8.2 m/h. The classifier was required to acievean overflow rate of 120 m3/h. Estimate:

1. The pool volume,2. Pool area,3. Pool depth,4. Pool diameter.

13.2.Laboratory tests on a sample of slurry showed the heights of the clear interface with time as:

Height,Time, t,

H, mmsec

6000

516100

434200

285400

176650

147800

1281000

The slurry containing 15% solids (by volume) was required to feed a continuous thickener toproduce an overflow containing no more than 1% solids (by volume). Specific gravity ofsolids was 2.65 and water 1.0. If the feed rate is 75 t/h and the desired underflow density is75% solids by mass, determine:

1. The settling velocity at each time interval,2. The concentration of solids corresponding to each settling velocity,3. The flux-concentration curve and4. The area of the thickener.

13.3.Using the data of problem 13.6.2, determine:

1. The volume of sludge in the underflow and hence the compression zone height2. The height of clarification zone.

13.4.A batch settling test on a flotation tailing gave the following results.

time(min)

013567

mud height(mm)340290236189175150

time(min)

8910111520

mud height(mm)1401251201078168

time(min)

2530405060

mud height(mm)

6360585555

Page 34: Chapter 13. Solid - Liquid Separation

434

0

50

100

150

200

250

300

350

400

0 10 20 30 40 50 60 70

Time, min

Inte

rfac

e H

eig

ht,

mm

434

Calculate the thickener area required, in m2/t/day for the following conditions.

C o = 50 kg/m3 pulp

C u = 340 kg/m3

13.5.A settling curve of a 15% solids (by mass) copper concentrate pulp is shown in the graphbelow. Estimate the thickener diameter required to dewater this material to an underflow of55 % solids (by mass) at a rate of 1000 t/h. Density of the solid is 4100 kg/m3 and water is1000 kg/m3.

400 n

350 '

E 300

E* r 250 -U)

I 200a>jo 150

a£ 100

500 J

C

H

)

JV

%-A

10 2

• - <

0 3

1

0

firn( m

4

ir0

I

50 60

t

70

13.6.A flocculated pulp settles according to the settling curve in the graph below. From theseresults, and a desired 1000 kL of clarified process water per hour from a feed slurry of 300tph at 20% solids;

1. locate the critical point on the plot,2. What is the initial concentration of solids (Co) and the underflow concentration (Cu) in

kg solid/m3 of pulp,3. estimate the size of thickener required using the method of Talmage and Fitch.

Data: Solids density = 2800 kg/m3 water density = 1000 kg/m3

Page 35: Chapter 13. Solid - Liquid Separation

435

0

50

100

150

200

250

300

0 100 200 300 400 500 600

Tim e (s)

Mu

dlin

e h

eig

ht

(mm

)

435

250

?>§, 200

jj 1500c=O 1003

s :50

C

r

"*• m- • .—a •

) 100 200 300 400 500 600

Time (s)

13.7A slurry of 20% solids (by mass) is to be dewatered to produce a product of 8% moisture at75 tph. A settling test is carried out on the slurry. The critical point of the settling curveoccurs at a mudline height of 80 mm and 250 seconds. The initial mudline height in the testcylinder is 300 mm. Solid density is 2500 kg/m3 and the water density is 1000 kg/m3.

1. If the mudline height corresponding to the thickener discharge is 70 mm, what would bethe thickener discharge % solids?

2. What method could be used to calculate the thickener area requirement for this slurry?Calculate the thickener diameter using this method.

13.8

A slurry of 20% solids (by mass) is to be dewatered to produce a product of 50% solids (bymass) at 75 tph (solid). A settling test is carried out on the slurry at 20%, 30% and 40%solids. The initial settling rates of the slurries are recorded below. Calculate the thickenerdiameter requirement for this slurry.

slurry R (mm/s)20%30%40%

0.77960.07800.0242

Page 36: Chapter 13. Solid - Liquid Separation

436436

13.9According to the Kyneh theory, the settling velocity of a slurry of a concentration given bythe settling interface, is given by the slope of a tangent to the settling curve. If the slope tothe settling curve of a flocculated copper flotation tail is 0.4 mm/s at the critical point whichoccurs at a point (67s, 115mm) on the settling curve, calculate the thickener area requirementto treat this slurry if the desired underflow is 65% solids (mass), the feed density is 22%solids (mass) and the initial mudline height in the settling test is 250 mm.

Throughput = 150tph

Density of solid = 2750 kg/m3

Density of water = 1000 kg/m3

REFERENCES[I] D.A. Dahlstrom and E.B. Fitch, in Mineral Processing Handbook, N.L. Wiess (ed),

SME/AIME, Chapter 9,1985, pp. 2-14.[2] D.L. King, in Mineral Processing Plant Design, A.L. Mular and R.B Bhappu (eds),

AIME, 1980, pp 541-577.[3] Eimco 2005, Retrieved September 1,2005 from http://www.glv.com:

http://www.glv.com/docs/product_docs/435/CompClarf50ft%20(pg)LR.pdf[4] Eimco 2006, Retrieved January 24,2006 from

http://www.glv.com/ProductList.aspx?secID=2&catID=131[5] H.S. Coe and G.H. Clevenger, Transactions AIMME, 55 (1916) 356.[6] GJ. Kynch, Trans. Faraday Society, 48 (1952) 166.[7] W.P. Talmage and E.B. Fitch, Industrial and Engineering Chemistry, 47 No.l (1955)

38.[8] F. Concha and R. Burger, KONA, No. 20 (2002), 38.[9] T. Yalcin, Bulletin of the Canadian Institute of Metallurgy, 81 No. 910 (1988) 69.[10] L. Svarovsky, Solid-Liquid Separation, Butterworths, 1977.[II] J.J. McKetta, Unit Operations Handbook, vol.2 Mechanical Separations and Materials

Handling, Marcel Dekker Inc., 1993.[12] P. Mondal and C.B. Majumdar, Journal of the Institution of Engineers (I) - Chemical

Engineering, The Institution of Engineers (India), 85 (2004) 17.[13] E. Barnea, Chemical Engineer, (1977) 75.[14] E.J. Roberts, Transactions, AIME, 184 (1949) 61.[15] A. Jernqvist, Svensk Papperstidn., 68 (1965) 506, 545,578.[16] A. Jernqvist, Svensk Papperstidn., 69 (1966) 395.[17] E.G. Kelly and D.J. Spottiswood, Introduction to Mineral Processing, Mineral

Engineering Services, Denver, 1989.[18] N. Yoshioka, Y. Hotta, S. Tanaka, S. Naito and S. Tsugami, Chemical Engineering

Japan, 21 (1957) 66.[19] H.H. Oltmann, Filtration and Separation, 12 No. 6 (1975) 636.[20] R.H. Perry and C.H. Chilton, Chemical Engineering Handbook, R.H. Perry and C.H.

Chilton (eds), 5th Edition, McGraw-Hill Book Co., 1973.[21] D.G. Osborne, Solid-Liquid Separation, Chapter 5, L. Svarovsky (ed), Butterworth,

London, Boston, 1977, pp. 75-99.

Page 37: Chapter 13. Solid - Liquid Separation

437437

[22] B. Fitch, Industrial and Engineering Chemistry, 58 (1966) 18.[23] R.H. Perry, R.H., Perry's Chemical Engineering Handbook, R.H. Perry, D.W. Green

and J.O. Maloney (eds), 6th edition, Chapter 19, McGraw-Hill, 1984, p. 64.


Top Related