Primary Clarifier Sizing in the Tropics: What Surface Overflow Rate Doesn't Tell You
Suggested readership: engineers laying out a new municipal plant in Indonesia, Vietnam or Thailand, and anyone retrofitting a primary tank that "worked on paper." Reading time: ~6 min. Companion tools: Primary clarifier · Rectangular primary · Lamella primary
The number everyone argues about is not the number that fails
Ask a room to size a primary clarifier and the conversation collapses to one figure: surface overflow rate. In temperate textbooks it is 1.5–2.5 m/h at average flow, 3–4 m/h at peak. Fine. But on a 10,000 m³/d plant with a 2.0 peak factor, a circular tank sized purely on SOR comes out at 278 m², 18.8 m diameter — and then quietly fails a different check entirely.
At ⌀18.8 m the peripheral weir is 59.1 m long. Peak flow of 20,000 m³/d over 59.1 m is 339 m³/m·d, well past the 250 m³/m·d that most guidelines treat as the ceiling for a single peripheral weir. The tank is big enough. The weir is not. The field symptom is familiar: a clarifier that looks calm at average flow and starts pulling fine solids over the effluent troughs every evening when the peak arrives — not because of hydraulics in the basin, but because of the drawdown cone at the weir.
That is why the geometry choice is not cosmetic.
| Geometry | Governing dimension | Weir behaviour | Best fit |
|---|---|---|---|
| Circular (radial) | Diameter set by SOR | One peripheral weir; length grows only as √area | Greenfield, 1–4 tanks, simple scraper |
| Rectangular | Length set by SOR and velocity | Long, straight weirs; easy to add length | Greenfield with land, shared-wall layouts |
| Lamella / inclined plate | Projected plate area set by loading | Short weirs over a tiny plan area | Retrofit, footprint-constrained sites |
The same plant in three geometries
Run the same 10,000 m³/d, peak factor 2.0–2.5, 250 mg/L TSS duty through all three:
Circular. 278 m², ⌀18.8 m, 3.0 m side-water depth, 833 m³, 2.0 h HRT at average flow. Peak SOR checks at 3.0 m/h. The weak point is the weir, as above — and the fix is straightforward: an inboard weir set at 0.7–0.8 of the radius roughly doubles effective weir length, or you simply specify ⌀22 m and accept 1.1 m/h average SOR as insurance.
Rectangular. Two tanks, 12 m wide, 3 m deep, sized at 2.0 m/h on peak flow (25,000 m³/d): 521 m² total, 21.7 m per tank, 1,563 m³, 3.8 h HRT. Weir loading lands at 288 m³/m·d — still above 250, but the number is easier to bend because weir length in a rectangular tank scales with width × number of bays, not with √area. Add a cross-baffle or a second effluent trough and you are compliant without touching the footprint.
Lamella. 260 m² of projected plate area in 10 modules, 17 m² of plan footprint — roughly one-fifth to one-tenth of the conventional alternative. That is the whole pitch. It is also the whole risk: a lamella primary has minutes of detention, not hours, and no meaningful storage volume. If your influent is a variable municipal sewage with a 2.5× peak, you need enough upstream equalisation that the plate pack never sees a slug, and you need to accept that you have traded sludge storage for area.
What primary treatment actually removes
Designers routinely over-promise here. The honest bands are 50–65 % TSS and 25–40 % BOD₅. On our example that is 150 mg/L TSS removed (1,500 kg/d) and 60–68 mg/L BOD removed (600–675 kg/d, about 27 %).
Note the ratio: you remove roughly two units of TSS for every one unit of BOD. Primary sludge is therefore bulky and only weakly organic. At 5 % DS, 1,500 kg/d of captured solids is 30 m³/d to pump, thicken and dispose of — a cost line that belongs in the comparison from day one, not after commissioning.
The tropical complication nobody budgets for
Primary tanks in a 30 °C climate go septic fast. Two hours of nominal HRT at 30 °C with a sludge layer that has not been drawn off is enough to generate sulphide, float black scum and lift a blanket of gas-entrained solids into the effluent. Three practical consequences:
- Desludge on a clock, not on a level. Automatic sludge draw-off every 2–4 hours beats a blanket-level trigger that only fires once the damage is visible.
- Watch HRT at peak, not at average. The rectangular option above gives 1.9 h at peak flow; the circular one around 1.0 h. If septicity is your historical problem, buy the volume.
- Ventilation and covers are structural, not architectural. Downwind odour complaints close plants faster than any effluent limit.
When the right answer is no primary tank at all
This sounds heretical and is frequently correct. Every mg/L of BOD you settle out is a mg/L of carbon your biological zone will not have for denitrification or bio-P release. For a plant aiming at TN < 10 mg/L or TP < 1 mg/L without bought-in carbon, a primary clarifier can be the single most expensive mistake in the flowsheet — you remove the readily biodegradable carbon and then dose methanol to replace it.
The decision rule is simple: if your influent BOD₅/TN is below about 4, think very hard before installing primary sedimentation. Above 5–6, primary is usually justified by sludge-handling savings alone. Between the two, run the numbers for your own site rather than inheriting a default.
Instruments that earn their keep here
A primary clarifier is usually the least instrumented unit in a plant and one of the few where a single sensor prevents a visible failure:
- Sludge blanket level — ultrasonic (S3000L class). This is the variable that decides whether you desludge in time and whether you are septic.
- Influent flow — electromagnetic flowmeter (SE10A class). Without it, "HRT" is a design number, not an operating fact, and peak behaviour is invisible.
- Primary sludge flow / density on the draw-off line. The simplest way to know whether you are pumping 3 % or 1 % solids.
Everything downstream inherits what this tank does. A primary clarifier that misbehaves does not produce a slightly worse effluent — it changes the entire load the biology was designed around.