Ballasted Flocculation
Microsand injected into the flocculation stage gives every floc a heavy, inert core — settling 30 to 60 times faster than conventional coagulation, in a tenth of the space and a fraction of the time.
Density from an added ballast
Ballasted flocculation solves the same separation problem from the opposite direction: instead of growing a denser floc, it adds density. Fine silica microsand — typically 100–150 µm, specific gravity 2.65 — is injected into the flocculation stage together with an organic polymer.
The polymer acts as a bridging agent, binding the coagulated particles onto the sand grains. Each floc is therefore built around a heavy, inert core rather than around water. Settling velocity rises by a factor of 30 to 60, so lamella loading of 40–120 m/h on projected area is routine and the whole train — from coagulant injection to clarified water — occupies 10 to 15 minutes of hydraulic residence time.
That short residence time has a second consequence: the process reaches full treated-water quality within minutes of start-up. It can be stopped and started at will, which makes it the natural choice for intermittent duties such as wet-weather flow, seasonal load and standby capacity.
The sand is not consumed. Sludge and sand are drawn from the clarifier hopper and pumped to a hydrocyclone; the dense sand reports to the underflow and returns to the injection tank, while the sludge leaves in the overflow. In steady operation more than 99 % of the circulating sand is recovered, and make-up is limited to attrition and carry-over.
The highest hydraulic rates, storm and peak duty, seasonal operation, and large drinking water plants where land is the binding constraint.
Process sequence
- CoagulationMetal salt coagulant is flash-mixed into the raw water, destabilising colloids and forming hydroxide floc. 1–2 minutes.
- Microsand and polymer injectionRecovered microsand from the hydrocyclone, plus make-up sand, is introduced together with an anionic polymer. Mixing energy is set to disperse the sand without shearing the developing floc.
- FlocculationA draft-tube reactor circulates the mixture on a controlled loop, giving uniform contact between floc and sand at low energy input. Polymer bridges attach the floc firmly to the grains. 4–8 minutes.
- Lamella settlingBallasted floc drops rapidly out of the rising water; inclined modules intercept the remainder. Loading is set on the projected plate area, 40–120 m/h depending on duty.
- Hydrocyclone recoverySludge and sand are pumped from the hopper to a cyclone. Centrifugal force sends the dense sand to the underflow, which gravity-returns to the injection tank; the sludge leaves in the overflow to dewatering.
Grain size is a design variable, not a detail. Finer sand offers more surface area per unit mass and therefore more attachment sites, but is harder to separate in the cyclone and easier to carry over. The 100–150 µm band is the working compromise between attachment efficiency, cyclone cut point and abrasion.
| Design parameter | Typical range |
|---|---|
| Microsand grade | 100–150 µm silica, SG 2.65 |
| Sand concentration in the reaction loop | 2–6 g/L |
| Sand make-up consumption | ~1–3 g per m³ treated |
| Coagulation time | 1–2 min |
| Injection / maturation time | 1–2 min |
| Flocculation time | 4–8 min |
| Total hydraulic residence time | 10–15 min |
| Rise rate — drinking & process water | 40–80 m/h |
| Rise rate — municipal wastewater, stormwater | 60–120 m/h |
| Rise rate — industrial / short-duration peak | up to 150–200 m/h |
| Sand + sludge recirculation rate | 3–6 % of feed flow |
| Start-up to full effluent quality | 5–15 min |
| Footprint vs conventional clarifier | 1/5 to 1/20 |
| Suspended solids removal | 90–99 % |
| Effluent turbidity | < 1–2 NTU |
| Effluent total phosphorus | < 0.1–0.3 mg/L |
| Sludge concentration | 0.3–1.5 % DS (to ~2 % with the thickening configuration) |
Ranges are typical engineering values for this process type and are given for orientation only. The design basis for a specific plant is fixed from raw water characterisation, jar or pilot testing and the agreed effluent guarantee.
Configurations
The same reaction and separation train, adapted to the treatment objective. Configurations can be combined.
KWS-BF · S
Coagulation, injection, flocculation and lamella settling in four stages. The reference configuration for drinking water, municipal and industrial duty across the full flow range.
KWS-BF · C
Injection and flocculation combined in a single high-efficiency draft-tube reactor. Roughly half the reactor footprint, lower mixing energy and start-up inside five minutes. Suited to retrofits with severe space limits.
KWS-BF · PAC
Powdered activated carbon dosed ahead of coagulation for taste and odour compounds, pesticides, dissolved organic carbon and trace micropollutants. Spent carbon is captured with the floc and leaves with the sludge.
KWS-BF · SF
Lime and, where required, soda ash dosing for carbonate hardness and silica reduction at high rate. The ballast keeps the dense carbonate precipitate under control at loadings a conventional softener could not sustain.
KWS-BF · T
An external thickener on the cyclone underflow raises sludge dry solids towards 2 % and cuts process water loss to roughly 0.3 % of throughput. Chosen where sludge transport or disposal cost dominates.
KWS-BF · WW
Configured for intermittent storm and combined sewer overflow duty: fast automated start-up, drain-down between events, and materials selected for long idle periods.
For small and mobile duties the whole train — reaction tanks, lamella settler, cyclone, dosing and controls — is supplied pre-assembled on a skid or in a container, factory tested and ready for connection. Typical packaged range is up to roughly 2,500 m³/h; above that, site-built concrete structures are normally more economical.
Common questions
Does the microsand damage pumps and valves?
The recirculation loop is designed as an abrasive-slurry duty from the outset: hardened or rubber-lined pumps, generous pipe radii, velocities held inside a defined band to keep the sand in suspension without scouring, and ceramic-lined cyclone components. Wear parts are identified as consumables with a defined replacement interval, and are listed in the spares schedule rather than discovered later.
How much sand do we actually have to buy?
After the initial charge, consumption is typically 1–3 g per cubic metre treated — a few tonnes a year on a mid-size municipal plant. Sand is a low-cost commodity, so the operating cost contribution is small; the recovery equipment exists mainly to keep the inventory stable, not to save on the material.
Can the plant be stopped and started?
Yes, and this is one of the strongest reasons to select the process. Full effluent quality is reached within 5–15 minutes of start-up, so a unit can be brought online only when it is needed — a few hundred hours a year for storm duty, or seasonally for peak demand — without a long stabilisation period or off-spec discharge.
What sludge does it produce?
Chemical sludge comparable to any coagulation process, at 0.3–1.5 % dry solids, with the microsand removed. Where sludge volume matters, the thickening configuration raises this towards 2 % and reduces water loss. If very concentrated sludge is a primary objective, high-rate clarification with integrated thickening is generally the better platform.
Will microsand end up in the sludge or the treated water?
A small quantity does, and that is the make-up consumption. Silica sand is inert and non-hazardous, so its presence in the sludge has no disposal consequence. Carry-over into the treated water is negligible when the settler and cyclone are correctly sized and the loop concentration is held in range.
Sizing a ballasted clarifier?
Give us the design and peak flow, the raw water solids and turbidity range, and the effluent target. We will come back with an indicative configuration, footprint and chemical demand.