Tiegel home Talk to a metallurgist

Thickener sizing from settling tests: the Coe–Clevenger and Talmage–Fitch unit area methods

Sizing a thickener from cylinder settling tests with the Coe–Clevenger and Talmage–Fitch unit area methods, with a metric worked example.

· 7 minute read · By

A thickener is a large circular tank that separates solids from water by settling. Slurry enters at the centre, the solids settle to form a dense layer on the floor, a slowly rotating rake moves the thickened slurry to a central outlet, and clarified water overflows a launder at the rim.

The sizing problem is to determine the floor area at which the water rises more slowly than the solids settle, and at which the solids have sufficient time to compact to the required underflow density. If the area is too small, solids are carried upward with the water and the overflow becomes dirty. A settling test in a glass cylinder addresses this problem, because the rate at which the interface between clear liquid and slurry falls in the cylinder represents the settling rate of the solids in the thickener.

Settling tests are inexpensive and quick, and remain the starting point for thickener sizing. This article describes the test procedure, interpretation of the settling curve, and conversion of the results to a thickener area with the two classical unit area methods.

Zones in a settling cylinderClear liquid at the top, a zone settling at constant rate, a transition zone and a compression zone at the bottom.Clear liquorFree settling at feed concentrationTransitionCompression (bed)Graduated cylinderCoe–Clevenger sizes the areafrom settling rates ofpulps at several dilutionsTalmage–Fitch uses one testand the time to reachunderflow concentration
Batch settling test. The height of the interface between clear liquid and pulp is recorded against time.

Settling test procedure

  1. Prepare the slurry at the expected thickener feed density from a representative sample at the plant grind. Use process water if it differs from fresh water, because salinity and dissolved salts affect settling.
  2. If the plant will use a flocculant, add it as a dilute solution at the planned plant dose in grams per tonne, and mix gently by inverting the cylinder or by plunging.
  3. Fill a graduated cylinder of 1 or 2 litres to a known height, start a timer, and record the height of the interface between clear liquid and slurry every 30 seconds initially and less frequently later. Continue until the height barely changes.
  4. Note the clarity of the supernatant and measure the final solids content of the settled bed.

The settling curve, a plot of interface height against time, has three parts: an initial straight line representing free settling at a constant rate, a curved transition, and a slowly falling tail as the bed compresses.

The Coe–Clevenger method

Coe and Clevenger reasoned that a thickener must pass, at every concentration between feed and underflow, the water that the solids release as they settle. At each concentration the upward velocity of the water must be lower than the settling velocity of the solids. Settling tests are run at several dilutions spanning the range from feed to underflow, the initial constant settling rate of each is measured, and a unit area is calculated for each:

UA=F−D24×R×ρl

where

UA
unit area (m² of thickener per tonne of dry solids per day)
F
liquid-to-solids mass ratio of the test pulp (dilution)
D
liquid-to-solids mass ratio wanted in the underflow
R
measured settling rate of that pulp (m/h)
ρl
density of the liquid (t/m³, about 1.0 for water)

Worked example: the underflow target is 50% solids, so D = 50 ÷ 50 = 1.0.

Test pulp (% solids)Dilution FSettling rate R (m/h)Unit area (m² per t/d)
20.04.03.100.040
25.03.01.900.044
33.32.00.950.044
40.01.50.480.043
45.51.20.150.056

For the 40% solids test, (1.5 − 1.0) ÷ (24 × 0.48 × 1.0) = 0.043 m² per tonne per day. The largest unit area, 0.056, governs. In the older imperial form of the equation the constant 1.333 appears because the result is in square feet per short ton per day.

The Coe–Clevenger method sizes the area for free settling. It does not account for the time the bed requires to compress to the underflow density, which is addressed by the Talmage–Fitch method.

The Talmage–Fitch method

Talmage and Fitch used a single test at the feed concentration and a graphical construction on the settling curve. The thickener area must give the solids sufficient time to reach the underflow concentration.

Hu=C0H0CuUA=tuC0H0

where

C0, H0
initial solids concentration (t/m³) and initial pulp height (m) in the cylinder
Cu
target underflow solids concentration (t/m³)
Hu
height the pulp would have if all of it were at underflow concentration (m)
tu
time where the tangent at the compression point meets the line at height Hᵤ (days)
  1. Calculate Hu. In the example, C0 = 0.15 t/m³ (150 g/L), H0 = 0.35 m, and a 50% solids underflow with a solids SG of 2.7 contains Cu = 0.73 t/m³ of solids, so Hu = 0.15 × 0.35 ÷ 0.73 = 0.072 m. Draw a horizontal line at that height.
  2. Locate the compression point, where the straight free-settling section of the curve turns into the slow compression tail. In the example it is at about 4.2 minutes.
  3. Draw the tangent to the curve at the compression point and extend it to the Hu line. The intersection gives tu ≈ 10.9 minutes, or 0.00756 days.
  4. Unit area = 0.00756 ÷ (0.15 × 0.35) = 0.144 m² per tonne per day.
Batch settling curve with Talmage and Fitch constructionThe interface falls fast at first, then slows as the bed compresses. The tangent at the compression point crosses the underflow line at about 10.9 minutes.01020300102030405060Time (min)Interface height (cm)tu ≈ 10.9 minHu = 7.2 cm (underflow line)Interface heightTangent at compression pointSettling curveTangentcompression point
Flocculated settling test. The underflow line is at H₀C₀/Cᵤ = 7.2 cm for a 50% solids underflow; the time at which the tangent reaches it sets the unit area. Illustrative figures.

Locating the compression point by eye is the weakest step, and two analysts may draw different tangents. Plotting the curve on log-log axes, where the compression point often appears as a distinct break, reduces this uncertainty. Kynch's theory, on which the method is based, assumes that settling rate depends only on local concentration, an assumption that flocculated and compressing pulps do not fully satisfy.

Thickener area and diameter

A=UA×Qs×marginD=4Aπ

where

A, D
thickener area (m²) and diameter (m)
Qs
dry solids to the thickener (t/d)
margin
design allowance chosen by the designer for variability in feed and test scale

The two methods gave 0.056 and 0.144 m² per t/d. The larger value governs, because the thickener must satisfy both free settling and compression. For a plant producing 2,400 t/d of dry tailings, with a design margin of 25% chosen for this example, A = 0.144 × 2,400 × 1.25 = 432 m², so D = √(4 × 432 ÷ π) = 23.5 m.

Depth is checked separately. The compression zone must be deep enough to hold the solids for the time they need to thicken, and the clear liquid zone deep enough to produce a clean overflow.

Limitations of batch tests

Modern high-rate thickeners use a feedwell that dilutes and flocculates the feed. Their unit areas are often a fraction of those of conventional thickeners, and they are sized from dynamic (continuous) bench or pilot tests together with supplier experience. Batch cylinder tests remain useful for ranking flocculants and doses.

Wall effects in small cylinders slow settling. The largest practical cylinder should be used, with a slowly rotating picket rake to represent the thickener rake in compression tests.

Grind, mineralogy (clays in particular), water chemistry and temperature all affect settling, and the tests should cover the range of conditions the plant will experience.

Glossary

Thickener
A large settling tank that produces a thick underflow and clear overflow water.
Unit area
The thickener floor area needed per tonne of dry solids per day (m² per t/d).
Dilution
Mass of liquid per mass of solids in a pulp; a 25% solids pulp has a dilution of 3.
Free settling
The early stage of a settling test, when the interface falls at a constant rate.
Compression point
Where the settling curve turns from free settling into slow compaction of the bed.
Flocculant
A long-chain polymer that binds fine particles into larger flocs, which settle much faster.
Feedwell
The central chamber where feed enters a thickener, is diluted and mixed with flocculant.
Underflow
The thickened slurry drawn from the bottom of the thickener.

Frequently asked questions

Which is better, Coe–Clevenger or Talmage–Fitch?

They address different requirements. Coe–Clevenger covers free settling over a range of concentrations; Talmage–Fitch includes the time to compress to the underflow density. Both should be calculated and the larger area used, then confirmed with supplier or dynamic test data.

What size cylinder should be used for settling tests?

The largest practical. One- and two-litre cylinders are common for screening flocculants; larger cylinders reduce wall effects in sizing work.

Why is the thickener overflow dirty when the area matches the design?

Common causes are a solids rate above design, poor flocculant mixing in the feedwell, a change in ore (particularly more clay), or a bed level high enough for solids to reach the overflow.

How much design margin should be added?

The margin is a design decision that depends on feed variability and confidence in the tests. The worked example uses 25% to illustrate the method; suppliers and design standards provide guidance for each application.

Discuss your plant with a metallurgist

Tiegel is a Perth metallurgy company. We help with testwork, plant optimisation, metal accounting and investigations, and we build the Tiegel Suite software.

Talk to a metallurgist

References and further reading

  1. Coe, H.S. and Clevenger, G.H. (1916). Methods for determining the capacities of slime-settling tanks. Transactions of the American Institute of Mining Engineers, 55.
  2. Talmage, W.P. and Fitch, E.B. (1955). Determining thickener unit areas. Industrial and Engineering Chemistry, 47(1), 38–41.
  3. Kynch, G.J. (1952). A theory of sedimentation. Transactions of the Faraday Society, 48, 166–176.
  4. Wills, B.A. and Finch, J.A. (2016). Wills' Mineral Processing Technology, 8th edition. Butterworth-Heinemann. Chapter 15, Dewatering.

Worked examples and charts marked illustrative use invented figures, not data from any operation.