Chapter 23 · Water, Slurry Transport, and Dewatering

Sedimentation, thickener duty and inventory

What the two product streams describe

A gravity thickener removes clarified liquor in overflow and withdraws a concentrated slurry in underflow. The main dry-solids, water and hydraulic objectives differ. High underflow solids concentration improves downstream water recovery, but sustained operation requires enough solids transport through the settling and compression regions. An overflow suitable for reuse may require a low residual suspended-solids load irrespective of underflow concentration. In a steady mass-balance model, a specified split can reproduce a measured operating point. A capacity model asks whether the material can settle at that load. An inventory model asks how long feed and draw imbalances can persist.

Use S and W for dry solids and water feed in t/h. If a fraction r of solids reports to underflow, then Sᵤ = rS and Sₒ = (1−r)S. With positive underflow mass fraction wᵤ, the water required there is Sᵤ(1−wᵤ)/wᵤ. An imposed overflow solids fraction wₒ with positive solids flow there demands Sₒ(1−wₒ)/wₒ. The two targets and r each constrain the same water balance.

Wᵤ + Wₒ = W.
If both percent-solids targets are exact, W = Sᵤ(1−wᵤ)/wᵤ + Sₒ(1−wₒ)/wₒ.

The Basic Thickener places all feed solids in underflow and assigns water toward its target underflow solids fraction. Water to underflow is capped at feed availability and the remainder overflows. The Steady-State Thickener accepts solids recovery and water-target inputs. It first reserves available water toward the underflow target. Overflow gets the residual when no positive overflow solids target is applied. When there are overflow solids and a positive target, it uses up to the required overflow water and sends any further water to underflow. Therefore a low overflow solids target may be water-limited. Conversely, an exact overflow target with surplus water drives the actual underflow solids below its first target.

For example, at S = 60 t/h, W = 95 t/h, r = 0.95, and an initial underflow solids target of 0.60, underflow reserves 38 t/h water. This leaves 57 t/h water and 3 t/h solids in overflow, or 5% solids. Setting overflow solids to 4% asks for 72 t/h water and is water-limited at 57 t/h. Setting overflow solids to 6% asks for 47 t/h. Its actual concentration is 6% and underflow receives 48 t/h, leaving 57/(57+48) = 54.286% solids. The configured nominal residence time in this steady model is a reported input under positive flow and area conditions. It does not increase capacity or constrain these allocations.

Settling tests and solids flux

In the Stokes regime for an isolated sphere, particle settling velocity is v∞ = gΔρd²/(18μ). Its use requires a low particle Reynolds number, measured particle density and a suitable Newtonian liquid viscosity. In a concentrated suspension, particles impede one another. Richardson and Zaki’s fluidization and settling work showed how voidage can represent a decrease in settling velocity with increasing solids volume fraction. Denoting that fraction by φ, a fitted hindered settling law is

v(φ) = v∞ (1−φ)ⁿ.

At positive n, v decreases with φ. For settling solids mass concentration C on a suspension-volume basis, gravity mass flux is Jg = C v. Its units are t/(m²·h) for C in t/m³ and v in m/h. A solids flux based on particle volume is φv with units m³ solids/(m²·h). Multiply the latter by solids density for the mass flux when material density is uniform.

Batch-Settling Thickener Sizing

A batch settling cylinder gives suspension interface heights hᵢ at recorded times tᵢ. The active Batch-Settling Thickener Sizing calculation estimates a positive settling speed from consecutive interval slopes, expressed in m/h, and an estimated concentration using the mean height h̄. With initial concentration C₀ and height h₀, its interval result is as follows.

vᵢ = (hᵢ₋₁ − hᵢ)/(tᵢ − tᵢ₋₁).
Cᵢ = C₀ h₀/h̄ᵢ.
Jᵢ = Cᵢ vᵢ.

Time is converted to hours. The calculation ignores non-decreasing height intervals and selects the smallest positive interval flux. For correction factor F and dry solids throughput S, estimated required area is F S/Jmin. The depth estimate uses the larger of the configured minimum and the observed height range. The smallest positive flux can be set by a slow late interval or by measurement noise. Check whether the limiting interval belongs to settling or compression and assess its repeatability. Short time increments or single rounded readings can create a false limit.

Take h = 0.80, 0.68, 0.55 and 0.49 m at times 0, 4, 11 and 21 min, with initial concentration 0.20 t/m³. The first interval has settling speed 1.80 m/h and concentration 0.2162 t/m³, giving flux 0.3892 t/(m²·h). The second gives 1.1143 m/h and 0.2602 t/m³, or 0.2899 t/(m²·h). The third gives 0.36 m/h and 0.3077 t/m³, or 0.11077 t/(m²·h). With F = 1.20 and S = 90 t/h, the required area estimate is 975 m². The large area derives from the late 10-minute interval, which can be investigated in replicate settling tests.

Hindered-Settling Flux Thickener

The active curve model fits v₀ and n to a mass concentration scale Cmax, and samples interior concentrations between zero and Cmax. It calculates each gravity flux and divides feed throughput by the highest sampled positive flux for an area estimate.

v(C) = v₀ [1 − C/Cmax]ⁿ.
Jg(C) = C v(C).

For n > 0, the continuous maximum occurs at Cmax/(n+1). The grid resolution influences the reported maximum. The model’s “limiting” flux in this context is the maximum pure gravity flux on that sampled curve. Withdrawal, compression and the full concentration profile can introduce stricter conditions.

Ideal Kynch thickening and operating feasibility

For a settling layer with bulk underflow downward velocity u, the ideal total solids volumetric flux adds gravitational settling and advective withdrawal. In a steady vessel with area A, dry solids volume flow Qₛ, target underflow volume fraction φᵤ and feed volume fraction φf, physical feasibility requires transport across all intermediate concentrations.

Qᵤ = Qₛ/φᵤ.
u = Qᵤ/A.
Jtotal(φ) = φ[v₀(1−φ)ⁿ + u].
Qₛ/A ≤ min over φf ≤ φ ≤ φᵤ of Jtotal(φ).

The active Ideal Kynch Thickener checks both endpoints and any interior minimum. Its derivative is u + v₀(1−φ)ⁿ⁻¹[1−(n+1)φ]. It checks the increasing portion of this derivative for a zero within the interval. Water availability separately requires Qᵤ − Qₛ ≤ Q𝓌. The model sends all dry solids to underflow when the duty is feasible. It rejects a duty beyond the flux limit or requiring unavailable underflow water. Physical overflow clarity, rake torque, compressive yield stress and flocculant distribution require evidence beyond the ideal flux calculation.

Worked area and underflow case

Use 90 t/h dry solids with particle density 3.0 t/m³ and 120 t/h water at 1.0 t/m³. Set target φᵤ = 0.55, v₀ = 8 m/h and n = 4. Solids volume rate is 30 m³/h, feed φf = 30/(30+120) = 0.20 and target underflow slurry volume is 54.545 m³/h. At area 55 m², u = 0.99174 m/h and the applied solids volume flux is 0.54545 m³/(m²·h). The endpoint fluxes are 0.85371 at φ = 0.20 and 0.72588 at φ = 0.55. The derivative on the upper branch remains negative at φᵤ and gives no interior minimum. The minimum of 0.72588 exceeds the applied load.

Underflow carries 24.545 t/h water and 90 t/h solids, or 78.571% solids by mass. Overflow carries 95.455 t/h water. If area is reduced to 20 m², applied flux is 1.5 and u = 2.72727 m/h. The feed-concentration flux is 1.20081. Since this candidate is already smaller than the applied flux, an admissible minimum cannot accommodate the duty. The model rejects the lower-area case. Report both the volume fraction requested and the mass concentration delivered. An unrealistically high underflow target can be mathematically feasible in the ideal model if material limits at high concentration are omitted.

Dynamic sedimentation inventory

The active Dynamic Sedimentation Thickener updates a lumped solids and water inventory over a time step Δt. It first adds feed masses. It then removes the requested underflow solids up to available inventory and assigns water toward the underflow mass fraction, capped at availability. When remaining wet inventory exceeds capacity, the model sends excess wet mass to overflow, with solids fraction equal to the inventory solids fraction times a specified carryover factor. The remainder is overflow water. For sufficiently wet inventories the update follows the two phase balances exactly.

Mₛ,next − Mₛ,now = Δt(Sfeed − Sᵤ − Sₒ).
M𝓌,next − M𝓌,now = Δt(Wfeed − Wᵤ − Wₒ).

Initial inventory, underflow draw and capacity must all be realistic. The level diagnostic uses wet inventory relative to capacity. It does not track a concentration profile or surveyed bed interface. For extreme low-water states check the water balance independently, since the update bounds final inventories at zero.

Consider initial solids 55 t and water 45 t. In a 0.25 h step, feed is 30 t/h solids and 65 t/h water, underflow draw 20 t/h solids and 50% solids by mass. After underflow, inventories are 57.5 t solids and 56.25 t water. At capacity 105 t, excess is 8.75 t. A carryover factor 0.20 gives 8.75 × 0.20 × 57.5/113.75 = 0.88462 t solids and 7.86538 t water in overflow. Final inventories are 56.61538 t solids and 48.38462 t water. Dividing overflow masses by the step gives mean outlet rates of 3.53846 and 31.46154 t/h. A measured instantaneous overflow during the step can differ from these interval rates.

Calibration and exercises

Measure solids and water rates to close the external balance. Establish solid density from representative mineralogy. For settling behavior, use several batch tests covering feed concentrations, flocculant dosage and dilution conditions. Fit kinetic parameters to interface velocities and check model predictions against independent throughput and underflow measurements. A successful mass split at one regime leaves the predicted capacity under a new solids distribution uncertain. Compare the applied flux to the model limit and confirm that the assumed thickening and compression regimes are covered by the tests.

  1. At 72 t/h dry solids and 110 t/h feed water, find Basic Thickener underflow and overflow water for a 65% underflow solids target.
  2. For the previous case, could the same feed produce 35% solids underflow by the Basic Thickener mechanism alone?
  3. Show that the maximum of C v₀ (1−C/Cmax)ⁿ occurs at C = Cmax/(n+1) for n > 0.
  4. What settling flux and estimated area result if the final height in the batch example is 0.52 m with the other inputs fixed?

Answers

  1. Underflow water is 72 × 0.35/0.65 = 38.769 t/h. Overflow water is 71.231 t/h.
  2. The target requires 133.714 t/h water, exceeding supply. All water and solids report to underflow, which remains at 72/(72+110) = 39.56% solids.
  3. Differentiate with respect to C to obtain v₀(1−C/Cmax)ⁿ⁻¹[1−(n+1)C/Cmax]. The sign changes from positive to negative at Cmax/(n+1).
  4. Last-interval velocity is 0.18 m/h and concentration is 0.16/0.535 = 0.29907 t/m³, giving limiting flux 0.05383 t/(m²·h) and area 2006.25 m². A shallow last-interval height change strongly affects this sizing estimate.

Kynch (1952), A theory of sedimentation, developed concentration-dependent settling theory. Richardson and Zaki (1954), Sedimentation and fluidisation, Part I80006-8), described voidage-dependent velocity. The stated active behavior follows DPSIM’s ThickenerSedimentationCalculations and IdealKynchThickener source.