Thickener Feedwell Design: Dilution and Flocculation

Last updated: 2026-11-03

Thickener feedwell design determines more settling performance per square meter than any other part of the machine: a well-diluted, gently flocculated feed typically lifts high-rate thickener underflow density by 3-6 percentage points (indicative) and cuts flocculant consumption by a quarter, while a poor feedwell caps whatever the tank below could achieve. This guide covers sizing, dilution, injection layout, and troubleshooting.

Key Takeaways

• Thickener feedwell design, not tank geometry, sets floc size; inlet energy is dissipated or wasted in seconds.
• Dilute to 7-12% w/w solids before flocculant contact; viscosity and reagent demand both punish dense feeds.
• Target 60-120 s feedwell residence time and 20-80 mm/s downward velocity (typical ranges).
• Roping, short-circuiting and turbidity are usually feedwell geometry problems, not flocculant problems.
• Flocculant at 10-40 g/t is the largest operating lever; retrofits pay back mainly through reagent and throughput.

What Is a Thickener Feedwell and What Does It Do?

A thickener feedwell is a submerged, open-bottomed chamber that receives incoming slurry, dissipates its inlet energy, blends dilution liquor and flocculant into the pulp, and hands a gentle, uniform stream to the settling zone below. It sits at the center of every gravity solid-liquid separation circuit, from concentrate duty in a zinc plant to tailings duty in a high-rate or paste machine, and it is the only place where the three variables that decide settling performance (energy, dilution, and polymer contact) can still be manipulated at once.

Consider what arrives at the feedwell. Slurry typically reaches the tank through a launder or pipeline at 1.5-3 m/s (typical), carrying kinetic energy proportional to the square of that velocity. If that jet were released straight into the clarification zone it would shear polymer chains and growing aggregates, entrain air, and suspend fines all the way to the overflow weir. The feedwell exists to absorb this energy in a controlled way: as a near plug-flow downward displacement, or as a tangential swirl that spreads flow evenly around the chamber before it exits into the bed. Energy dissipation is the feedwell’s first job, and it happens within a second or two of entry.

Feed dilution is a controlled addition of clarified liquor that lowers the solids concentration of incoming pulp before flocculant contact. Most ground concentrates flocculate best between 7-12% w/w solids (typical, to be confirmed by settling test work), while mill discharge often arrives at 25-45% w/w. At high solids, viscosity climbs and polymer chains cannot bridge particles efficiently; the same dose that builds coarse aggregates at 10% solids may build half the size at 35%. Dilution also lowers reagent demand per tonne and reduces the liquor volume dragged down into the bed.

Finally, why does the feedwell, not the tank, set floc size? Polymer adsorption and aggregate growth take seconds, not minutes. By the time slurry has traveled one feedwell diameter below the exit, the aggregate size distribution is essentially fixed. The tank below can preserve that distribution or destroy it, but it cannot rebuild it. Tank diameter, sidewall depth, and rake design govern consolidation and clarity at the margin; floc size is decided above, which is why thickener feedwell design deserves the same rigor as thickener sizing itself.

Thickener Feedwell Design Parameters and Typical Ranges

Feedwell sizing is a chain of coupled choices: the more dilution you need, the more volume the chamber must hold, and the larger the chamber, the slower the vertical velocity and the gentler the exit conditions. The table below collects the parameters that matter, with the ranges we use as starting points before settling tests refine them. Every value is typical or indicative and must be confirmed by test work on your own pulp.

Three of these rows deserve comment. The diameter fraction is not arbitrary: a feedwell below about 0.25 of tank diameter cannot house the dilution liquor volume a dense feed needs, while one much above 0.45 eats into the clarification area you paid for. Downward velocity is the single best early warning number; when a client reports fresh feed breaking through into the underflow, the exit velocity is the first thing we recompute. And the flocculant rows interact: a weak solution at 0.025-0.05% w/w needs more pumping volume but distributes far more evenly through the pulp, which matters more than it sounds because local overdose produces small, dense, shear-brittle flocs even at the correct total dose. Dose and strength should come from bench work with a laboratory flocculant jar test apparatus run on your own liquor, then verified in a settling column before they are frozen into design. Where the machine sits between conventional and high-rate duty, the choice of feedwell depth and dilution arrangement should follow the high-rate thickener selection logic rather than being inherited from a previous project.

How Do You Size a Feedwell for Dilution?

The sequence below reflects how thickener feedwell design is executed in practice, with a worked example carried through every step so the arithmetic can be checked and reused. The example: a concentrate thickener treating 500 t/d of dry solids, feed at 25% w/w solids, dilution target 10% w/w, solids specific gravity 4.2 (typical for a base-metal concentrate; confirm by measurement), 90 s residence time.

1. Fix the design feed and slurry properties. Dry solids: 500 t/d ÷ 24 h = 20.8 t/h. Feed slurry mass at 25% w/w: 20.8 ÷ 0.25 = 83.3 t/h, of which water is 62.5 t/h (≈ 62.5 m³/h). Solids volume: 20.8 ÷ 4.2 = 5.0 m³/h. Total feed slurry volume ≈ 67.5 m³/h.
2. Choose the target solids concentration for flocculation. Settle a series of dilutions in a laboratory settling test column and pick the concentration where initial settling rate stops improving; for ground concentrates this is typically 7-12% w/w. We take 10% w/w.
3. Compute the dilution liquor flow. Required slurry mass at 10% w/w: 20.8 ÷ 0.10 = 208 t/h. Required water mass: 208 − 20.8 = 187.5 t/h ≈ 187.5 m³/h. Dilution liquor flow = 187.5 − 62.5 = 125 m³/h. The dilution ratio is 125 ÷ 67.5 ≈ 1.9 : 1 by volume, comfortably inside the typical band.
4. Compute the total volumetric feed to the tank. Water 187.5 m³/h plus solids 5.0 m³/h ≈ 192.5 m³/h. This, not the 67.5 m³/h arriving from the mill, is the flow the feedwell internals must handle, and it is the number most often forgotten in retrofits.
5. Size the feedwell volume from residence time. At 90 s: V = 192.5 × 90 ÷ 3600 ≈ 4.8 m³ of internal, working volume below the dilution ports and above the exit.
6. Convert volume into diameter and depth. Pick a downward velocity of about 35 mm/s: A = 0.0535 m³/s ÷ 0.035 m/s ≈ 1.5 m², giving a diameter of ≈ 1.4 m. Depth = 4.8 ÷ 1.5 ≈ 3.1 m. Rounding to a 1.5 m diameter × 3.0 m depth gives ≈ 5.3 m³, residence ≈ 99 s, velocity ≈ 28 mm/s. All checks pass.
7. Check the flocculant duty and injection layout. At 20 g/t: 20.8 t/h × 20 g/t = 416 g/h of active polymer. At 0.04% w/w solution strength that is 416 ÷ 0.0004 ≈ 1.04 m³/h of prepared solution, split across 6 injection points at roughly 0.17 m³/h each, injected below the dilution ports where the pulp is already thin.

Step 4 is the one that causes retrofit grief: designs based on the incoming slurry flow instead of the diluted flow end up with residence times half the design intent and velocities twice it. Where the wider machine dimensions are still open, run the same checks against the thickener sizing basis so the feedwell and the tank are sized on one consistent flow sheet.

Auto-Dilution vs External Dilution: Which Fits Your Circuit?

In thickener feedwell design terms there are only two ways to source the 125 m³/h of the worked example, and each buys something different. Auto-dilution is an internal recirculation arrangement that draws clarified liquor through ports in the feedwell wall, driven by the density difference between incoming feed and the clear zone; external dilution pumps liquor from the overflow launder, a process-water header, or both, through metered lines. Auto-dilution is self-regulating and needs no pumps or controls, but its flow cannot be set independently: it falls exactly when a dense feed slug arrives and dilution is most needed. External dilution costs pump capital and power and needs flow metering, but it is controllable and measurable, which matters when flocculant cost dominates.

The verdicts are deliberately concrete. For a new machine on a dense concentrate, auto-dilution wins on simplicity and reagent performance, and it is the default in our own proposals. For retrofits, external dilution usually wins because the mechanical scope stays outside the tank; a hybrid, in which internal ports are supplemented by a metered external line, is common when some independent control is wanted. For paste thickener design the answer shifts again: deep feedwells with staged dilution are standard because yield stress, not just viscosity, is the constraint, and that is a design conversation of its own.

What Goes Wrong in Thickener Feedwells? 7 Symptoms, Causes and Fixes

Most thickener feedwell design defects announce themselves through a handful of recurring signatures. The list below is ordered by how often plant surveys encounter them, with symptom, root cause, and corrective action for each.

1. Roping in the thickener. Symptom: a spiral-shaped bed surface, erratic rake torque, and underflow density that cycles instead of holding. Cause: excess downward momentum at the feedwell exit pushes into the bed and drags a column of settled solids around with the rake, usually combined with too little dilution and a high yield-stress bed. Action: recompute exit velocity against the 20-80 mm/s band, add dilution, and increase feedwell bottom clearance; verify bed rheology using slurry rheology test data before changing rake hardware.
2. Short-circuiting down the feedwell wall. Symptom: coarse or poorly settled solids appear in the underflow almost immediately after a feed change, and the overflow stays turbid at the same time. Cause: slurry hugging the inside wall and plunging to the exit without mixing, typically from a single-side inlet aimed tangentially at the wall or a missing diffuser. Action: convert to balanced twin or swirl entry, add an outlet diffuser ring, and recheck the diameter fraction against the dilution duty.
3. Flocculant overdose. Symptom: reagent consumption drifts upward over months with no improvement in clarity, and the bed turns viscous and slow to raking. Cause: dose crept up to compensate for a dilution or mixing defect the feedwell actually owns; above roughly 40 g/t (typical), extra polymer produces small dense flocs that blind the bed. Action: return the dose to jar-test values, fix the real defect, and audit the addition point before increasing pump size again.
4. Feedwell sanding and blockage. Symptom: intermittent slumps of coarse solids, rising rake torque spikes, and lost dilution flow. Cause: feed launder velocity below roughly 1.2 m/s lets particles settle in the launder and slough into the feedwell in slugs; long horizontal runs and size creep from upstream mills make it worse. Action: re-slope or shorten the launder, restore design velocity, and add flushing connections.
5. Vortexing from unbalanced feed. Symptom: air entrained into the underflow, floating floc rafts at the overflow, and uneven bed draw-off. Cause: a single feed entry sets the whole chamber swirling asymmetrically, so one sector of the bed is starved while another is overloaded. Action: split the entry into two balanced inlets or adopt a true tangential swirl design, and verify the feedwell is level.
6. Persistent overflow turbidity. Symptom: overflow carries fines continuously, above roughly 100-200 NTU (nephelometric turbidity units, plant-specific basis), after dose and bed level are confirmed correct. Cause: dilution ratio too low, so the pulp flocculates dense and small, or residence time shortened by a dilution flow that was never included in the design. Action: recompute the diluted flow, enlarge the chamber or raise the dilution flow, and re-run settling tests at the corrected concentration.
7. Flocculant injection at high shear. Symptom: clarity is poor despite dose, and dilution liquor is milky with unsettleable micro-flocs. Cause: polymer is injected into the inlet jet or the launder, where local shear exceeds what polymer chains tolerate; the chains break before they ever reach the pulp. Action: move injection below the dilution ports into the thinned pulp, split the flow across a multi-point ring, and keep local injection velocities low.

Notice the pattern: five of the seven causes are geometric or hydraulic, and only one is genuinely a reagent problem. That distribution matches what we see in the field, and it is why the troubleshooting order should always run geometry first, dilution second, chemistry last.

What Does Thickener Feedwell Design Cost, and Where Is the Payback?

On capital, the feedwell, its dilution internals, and the associated feed launder and injection system typically represent 5-10% of a complete thickener package price (indicative; varies with diameter, materials, and duty). That is a modest share of capex, and the temptation to value-engineer it down is exactly backward: it is the cheapest part of the machine to improve and the most expensive to get wrong, because a feedwell defect is partially corrected only by spending more flocculant forever.

On operating cost, flocculant consumption of 10-40 g/t of dry solids is the dominant lever, and it dwarfs the dilution pump power it replaces. The arithmetic is simple enough to plan with. A plant treating 5,000 t/d of solids, 350 d/y, cutting dose from 25 to 17 g/t saves 5,000 × 350 × 8 g = 14 t of polymer per year; at a delivered price of roughly 3-5 $/kg (typical, indicative), that is 40-70 thousand dollars annually. A feedwell retrofit costing several hundred thousand dollars then pays back in roughly three to seven years on reagent alone, faster where underflow density gains also relieve a downstream filtration bottleneck or where overflow clarity credits solution losses back to the circuit. Smaller plants may not justify retrofit capex on reagent savings alone; there the case usually rests on availability, torque stability, and turbidity rather than payback arithmetic.

Two honest caveats. First, all figures here are typical and indicative ranges compiled from published practice and our own project records; they must be confirmed by test work and a site-specific survey before being used in any budget, and the dose side of the analysis is developed further in our flocculant dosage guide. Second, the payback logic only holds if the diagnosis is right: a retrofit that fixes a misdiagnosed defect, for example adding dilution where the real fault is a sanding launder, buys nothing. Scope a feedwell retrofit the way you would scope any process change: bench confirmation first, then a quoted mechanical scope. You can find the supporting test equipment and further reading in our resources library.

Field Data: What Changes When a Feedwell Is Retrofitted?

Anonymized records from retrofits we have supervised show what thickener feedwell design changes actually deliver, with each figure carrying its basis. The numbers are indicative, plant-specific, and rounded; treat them as direction, not warranty.

The lead-zinc numbers settled within a week of restart, with underflow density measured by vacuum-dried samples averaged over two weeks. The gold tailings retrofit is the instructive one: rake torque peak events fell from about four per day to fewer than one per week while the dose never moved, because the gain came entirely from energy dissipation rather than chemistry. The copper cleaner case shows the step 4 arithmetic trap: the original feedwell had been sized on the mill discharge flow, residence time was well under 40 s, and resizing on the diluted flow restored the designed 85-90 s and lifted underflow density by six points (indicative).

Every one of these retrofits was first reproduced at bench scale in settling columns on the plant’s own pulp, and the predicted direction held in every case even where the magnitudes shifted. That two-step discipline, bench prediction followed by mechanical correction, is the cheapest way to de-risk the spend.

Frequently Asked Questions

What solids concentration should flocculation happen at in a thickener feedwell?

For most ground concentrates and tailings, flocculation performs best between roughly 7-12% w/w solids (typical, to be confirmed by settling tests on your own pulp). Above that band, rising viscosity restricts polymer chain extension and the same dose builds smaller, denser aggregates. Below it, you are paying to pump and dilute water without gaining settling rate. The optimum is found by running settling rate curves at several dilutions and picking the plateau.

How deep should a thickener feedwell be?

Typical internal depths run 2-6 m depending on machine size and duty (indicative). The bottom must sit safely above the bed interface at maximum inventory, while the chamber must be deep enough to hold the dilution liquor volume at the design residence time. Deep, staged-dilution feedwells are standard on paste and high-compression duty, where the bed is high and yield stress governs entry into it.

Can I add flocculant in the feed launder instead of the feedwell?

Occasionally, but it is rarely the best choice. Launder injection exposes polymer and growing aggregates to transit velocities of 1.5-3 m/s and to bends and drops, all of which shear chains and break flocs before they reach the pulp. If upstream mixing time is genuinely needed, a low-shear pipe reactor section can work; otherwise inject below the dilution ports, in the thinned pulp, through a multi-point ring.

What is auto-dilution and when does it fail?

Auto-dilution is an internal recirculation arrangement that draws clarified liquor into the feedwell through wall ports, driven by the density difference between feed and clarified zone. It fails when ports block or scale over, when the density difference collapses because the overflow itself is turbid, and during feed slugs, when dilution is needed most precisely when driving head is weakest. Pairing it with a small metered external line covers those cases.

Why is my thickener roping after a feed grade change?

Roping is a bed-flow failure in which settled solids spiral with the rake, usually triggered when a finer, denser, or faster-settling feed raised exit momentum or bed yield stress beyond what the feedwell was sized for. The fix is rarely the rake: recompute feedwell exit velocity against the 20-80 mm/s band, restore the dilution ratio, and check bed rheology before changing any drive hardware.

How much dilution water does a thickener feedwell need?

Compute it from a mass balance, not a rule of thumb. For each tonne of dry solids, the dilution liquor flow is the difference between the water content at the target flocculation concentration and the water arriving in the feed. In our worked example, 500 t/d of solids diluted from 25% to 10% w/w required about 125 m³/h, a ratio of roughly 1.9 : 1 against the incoming slurry volume (typical band 0.5-3 : 1).

How do I test feedwell changes without shutting the thickener down?

Bench scale first, always. Settling tests in a column reproduce dilution effects and flocculation windows on your actual pulp, and jar tests fix dose and solution strength. Predict the expected change in settling rate, turbidity, and underflow density, then implement the mechanical change at the next available shutdown and verify against the prediction. Discrepancies between bench and plant usually point to launder sanding or inlet imbalance.

Conclusion

Thickener feedwell design is where settling performance is actually won: the chamber that dissipates inlet energy, dilutes dense feed into the 7-12% w/w flocculation window, and places polymer where chains survive decides the floc size the tank below must merely protect. Size it on the diluted flow, not the incoming flow; check exit velocity and residence time against the ranges in the parameter table; and put geometry, dilution, and chemistry in that diagnostic order when clarity or underflow density drifts. The capex share is small, the flocculant lever is large, and retrofit records consistently show double-digit percentage gains in underflow density and reagent consumption when the diagnosis is right. If you want a specific feedwell sized, reviewed, or retrofitted against your own settling test data, request a quote with your pulp properties and duty, and our process engineers will return a design basis you can check line by line.

References

• Wills, B.A. and Finch, J.A., Wills’ Mineral Processing Technology, 8th ed., Butterworth-Heinemann.
• SME Mineral Processing and Extractive Metallurgy Handbook, Society for Mining, Metallurgy and Exploration, Englewood, CO.
• Svarovsky, L., Solid-Liquid Separation, 4th ed., Butterworth-Heinemann.
• Perry’s Chemical Engineers’ Handbook, section on gravity sedimentation and thickening, McGraw-Hill.
• Kynch, G.J., “A theory of sedimentation,” Transactions of the Faraday Society, 1952.
• U.S. Environmental Protection Agency, process design manual series on sludge thickening and handling.
• Australasian Institute of Mining and Metallurgy, conference proceedings on thickening and settling practice.

Changelog

• 2026-11-03: First publication.