Chemical dosing systems for DAF
More DAF units underperform because of the dosing skid than because of the flotation cell. It is the cheapest part of the plant and the one most often value-engineered into uselessness.
A complete DAF dosing system has four parts: pH correction ahead of coagulation, coagulant dosing into a high-energy mixing point, polymer make-up and dosing into a low-energy flocculation stage, and the flocculation chambers themselves. Coagulant needs rapid mixing at a velocity gradient of 300–1,000 s⁻¹ for 10–60 seconds; polymer needs gentle mixing at 30–70 s⁻¹ for 1–3 minutes. Reversing those two conditions — which happens whenever both chemicals are dosed into the same inline mixer — shears the floc and the unit will never reach its design performance.
The correct sequence
| Stage | Chemical | Mixing energy G | Time | Purpose |
|---|---|---|---|---|
| 1. pH correction | NaOH, lime, H₂SO₄ | 200 – 500 s⁻¹ | 1 – 3 min | Bring into coagulant working window |
| 2. Coagulation | PAC, alum, FeCl₃ | 300 – 1,000 s⁻¹ | 10 – 60 s | Charge neutralisation, destabilisation |
| 3. Flocculation stage 1 | Polymer | 50 – 70 s⁻¹ | 1 – 2 min | Bridging, initial floc growth |
| 4. Flocculation stage 2 | — | 25 – 40 s⁻¹ | 1 – 2 min | Floc maturation without shear |
| 5. Flotation | — | Minimal | 3 – 20 min | Bubble attachment and separation |
Dosing skid specification
- Coagulant pump: diaphragm metering pump, turndown at least 10:1, with a calibration column so dose can be verified rather than assumed. Duty and standby on anything above 30 m³/h.
- Polymer dosing: progressive cavity or low-shear diaphragm pump. High-shear pumps degrade long-chain polymer before it reaches the water.
- Polymer make-up: for powder polymer, an automatic wetting, mixing and maturation unit with at least 30 minutes of ageing. Powder polymer dumped into a stirred drum forms fisheyes and delivers a fraction of its rated performance. Emulsion polymer needs an inline dilution and activation loop.
- Dilution water: clean water, never raw effluent. Polymer made up in effluent is partly consumed before it does any work.
- pH control: inline probe with automatic dosing, dead-band control and a second probe at the DAF inlet for verification. One probe with no cross-check gives you a number, not control.
- Storage: bunded to 110 % of the largest vessel, shaded, with eyewash and shower. Sodium hypochlorite, ferric chloride and caustic all degrade or crystallise in African ambient temperatures — shade is a process requirement, not a comfort item.
- Materials: PVDF, PTFE or PE for ferric and acid lines. PVC embrittles quickly in direct sun and is a frequent source of chemical spills.
Chemical consumption and storage sizing
Chemical supply reliability varies widely across the continent. Size storage on the realistic resupply interval, not on the theoretical one.
| Chemical | Typical dose | Consumption at 30 m³/h | Recommended storage |
|---|---|---|---|
| PAC 10 % Al₂O₃ | 100 – 300 mg/L | 72 – 216 kg/day | 21 – 30 days |
| Ferric chloride 40 % | 150 – 500 mg/L | 108 – 360 kg/day | 21 – 30 days |
| Anionic polymer (powder) | 1 – 5 mg/L | 0.7 – 3.6 kg/day | 60 – 90 days |
| Caustic soda 32 % | Variable | Site specific | 30 days |
| Sulphuric acid 96 % | Variable | Site specific | 30 days |
Automatic dose control
Three levels, in increasing order of cost and capability:
- Flow-proportional. Dose pump speed tracks inlet flow. The minimum acceptable standard and adequate where influent concentration is stable. Cheap and reliable.
- Flow plus feedback. Adds a turbidity or streaming-current monitor on the outlet, trimming dose against measured performance. Typically cuts chemical consumption by 10–25 % and pays back within two years at industrial scale.
- Feedforward on influent. Inlet turbidity or streaming current sets the dose before the water reaches the mixer. Best performance on variable streams, highest instrumentation burden, and only worth it where someone will maintain the probes.
Be honest about the last point. An unmaintained streaming-current detector is worse than no detector, because the control loop will act on a bad signal. Where instrument maintenance is not realistic, flow-proportional dosing with a disciplined daily jar test is the more robust plant.
Frequently asked questions
Can coagulant and polymer be dosed at the same point?
No. They need opposite mixing conditions. Coagulant requires rapid, high-shear mixing at G of 300–1,000 s⁻¹ to disperse before it hydrolyses. Polymer requires gentle mixing at 30–70 s⁻¹ so the long chains can bridge particles without being torn apart. Dosing both into one inline mixer gives you neither: the coagulant is under-dispersed and the polymer is sheared. Separate them by at least 30 seconds of travel and change the mixing energy between them.
How much polymer does a DAF need?
0.5 to 6 mg/L as active polymer, depending on the stream. Municipal and paper duty sits at 0.5–3 mg/L, food and dairy 1–3 mg/L, oily and tannery streams 2–6 mg/L. Overdosing is as damaging as underdosing: excess polymer re-stabilises particles and produces a slimy float that blinds the skimmer. Confirm by jar test and re-check whenever the process changes.
Should I use powder or emulsion polymer?
Powder is cheaper per kilogram of active polymer, stores for years and is not affected by heat, but it requires a proper make-up unit and 30–60 minutes of maturation. Emulsion is easier to handle and activates in minutes, but costs more per active kilogram and has a shorter shelf life. On remote African sites with unreliable resupply, powder plus a good make-up unit is generally the better choice.
What is the biggest dosing mistake you see?
Polymer made up in raw effluent or in untreated borehole water. Hard water and effluent both partially consume the polymer before it reaches the flocculation stage, so the plant runs at two to three times the necessary dose and still underperforms. Clean, low-hardness dilution water at the correct concentration — typically 0.1–0.5 % for powder — solves it, and it costs almost nothing.