DAF tank design splits the vessel into two hydraulically distinct zones: a contact (reaction) zone where microbubbles collide with and attach to flocs, and a separation (clarification) zone where the buoyant bubble-floc aggregates rise to the float layer. Good design sizes each zone independently — contact time and upflow velocity for the first, surface loading (rise) rate and cross-flow velocity for the second — then controls density currents so real flow matches the plug-flow ideal.
What are the two functional zones of a DAF tank?
Hydraulically, a dissolved air flotation tank is not one vessel but two coupled reactors in series, separated by a baffle:
- Contact (reaction) zone. The pressurised, air-saturated recycle is released through nozzles at the base and mixes with the coagulated feed. Here the physics is collision and attachment: microbubbles (10–100 µm) and flocs must meet, and the bubbles must stick, forming aggregates of reduced bulk density. The zone is narrow and vertical, with a deliberately high upflow velocity to promote contact and carry aggregates over the baffle.
- Separation (clarification) zone. The aggregate-laden water flows over the baffle into a wide, quiescent basin. Here the physics is buoyant rise and separation: aggregates rise to form the float; clarified water is drawn off low down through perforated laterals. This zone sets the tank footprint.
Because the governing mechanism differs — kinetics and mass transfer in the contact zone, buoyancy and hydraulics in the separation zone — the two zones are sized by different equations. Treating the tank as a single well-mixed box is the most common design error. For the upstream sizing basis (flow, recycle and air), see our guide to sizing a DAF system.
How is the contact (reaction) zone sized?
The contact zone is designed for two coupled targets: a contact time long enough for bubble-floc attachment kinetics to complete, and an upflow velocity high enough to keep bubbles and flocs colliding while carrying aggregates upward.
where tc = contact time (s), Vc = contact-zone volume (m³), Ac = contact-zone plan area (m²), Q = feed flow and Qr = recycle flow (m³/s). Typical design: tc ≈ 60–120 s and vup ≈ 20–80 m/h (upflow), with modern high-rate units toward the upper end.
The attachment process is a first-order collision problem. The rate at which free bubbles are removed by attachment to flocs follows white-water (bubble-cloud) theory, in which the single-collector collision efficiency and the bubble number concentration set a first-order rate constant. In practice designers do not solve the population balance on site; they translate it into the two robust surrogates above — a minimum residence time so the reaction can proceed, and a minimum upflow velocity so the collision frequency stays high. A contact zone that is too short or too slow starves the separation zone of well-formed, low-density aggregates. The underlying collision-attachment physics is covered in our note on DAF bubble dynamics.
What sets the separation-zone surface loading (rise) rate?
The separation zone is sized by the surface (hydraulic) loading rate, also called the rise rate or overflow rate — the total flow divided by the horizontal projected area of the separation zone. It must not exceed the rise velocity of the slowest bubble-floc aggregate you intend to capture, or that aggregate is washed out with the clarified water.
where SLR = surface loading rate (m/h, equivalent to m³/m²·h), As = separation-zone plan area (m²) and vb = design rise velocity of the bubble-floc aggregate (m/h). The rise velocity of a single aggregate follows a buoyant Stokes law: vb = g·da2(ρw − ρa) / 18µ, with da the aggregate diameter, ρa its bulk (reduced) density and µ the water viscosity.
Note the recycle flow is included in the loading: the recycle water leaves with the clarified effluent, so it counts toward the rise rate. This is why raising the recycle ratio to supply more air simultaneously raises the hydraulic load on the separation zone — a real design tension. The choice of design SLR then defines the split between conventional low-rate and high-rate DAF, discussed next.
Low-rate vs high-rate DAF: how do the numbers differ?
The distinction is almost entirely one of separation-zone hydraulic loading — and the aggregate rise behaviour that permits it. Low-rate DAF relies on the individual buoyant rise of discrete aggregates (Stokes regime). High-rate DAF exploits hindered and bulk stratified rise of a dense bubble cloud, where a rising suspension behaves collectively and tolerates far higher throughput, often with lamella/tube inserts to shorten the rise path.
| Parameter | Conventional (low-rate) DAF | High-rate DAF |
|---|---|---|
| Separation SLR (rise rate) | 5–12 m/h | 15–40 m/h |
| Contact-zone upflow velocity | 20–40 m/h | 40–80 m/h |
| Contact time | ≈ 90–120 s | ≈ 60–90 s |
| Rise mechanism | Discrete Stokes rise | Hindered / stratified bulk rise |
| Tank footprint | Large | Compact (2–4× smaller) |
| Sensitivity to hydraulics | Moderate | High — needs uniform flow |
The trade-off: high-rate designs deliver a small footprint but leave almost no hydraulic margin, so density currents and short-circuiting that a low-rate tank shrugs off can ruin a high-rate one. That is where flow uniformity and CFD earn their keep.
Worked example 1: sizing the separation zone
Design a separation zone for a feed of Q = 120 m³/h with 50% recycle (Qr = 60 m³/h), targeting conventional low-rate operation at a design rise rate of SLR = 8 m/h.
- Total flow: Q + Qr = 120 + 60 = 180 m³/h.
- Separation area: As = (Q + Qr) / SLR = 180 / 8 = 22.5 m².
- Rise-velocity check: for a 60 µm aggregate (da = 6×10−5 m) of reduced bulk density ρa = 700 kg/m³ in water at ρw = 998 kg/m³, µ = 1.0×10−3 Pa·s: vb = 9.81 × (6×10−5)² × (998 − 700) / (18 × 1.0×10−3) = 5.8×10−4 m/s ≈ 2.1 m/h.
The check flags a subtlety: a lone 60 µm aggregate rises at only ~2 m/h, below the 8 m/h design rate — which is exactly why DAF works by aggregating multiple bubbles onto each floc (raising da and lowering ρa) rather than relying on single microbubbles. Attaching a few 40–60 µm bubbles to a floc easily lifts the effective rise velocity above 8–10 m/h. If the required rise rate cannot be met, either lower the SLR (larger tank) or improve bubble-floc contact upstream. To pick a defensible design SLR for a given effluent, cross-check against our DAF sizing methodology.
Worked example 2: sizing the contact zone
Using the same total flow of 180 m³/h (0.050 m³/s), size a contact zone for tc = 90 s at an upflow velocity of vup = 60 m/h.
- Required volume: Vc = tc × (Q + Qr) = 90 s × 0.050 m³/s = 4.5 m³.
- Contact-zone plan area: Ac = (Q + Qr) / vup = 180 / 60 = 3.0 m².
- Contact-zone depth: Hc = Vc / Ac = 4.5 / 3.0 = 1.5 m.
So the contact zone is a compact riser of ~3 m² and 1.5 m effective depth. Note it is roughly 13% of the 22.5 m² separation area from Example 1 — the reaction zone is deliberately small and fast, the clarification zone large and slow. The baffle over which flow passes from one to the other should be designed so the crest velocity does not shear the fresh aggregates apart.
How do you check cross-flow velocity and short-circuiting?
In the separation zone, clarified water travels horizontally toward the effluent laterals while aggregates rise vertically. The cross-flow (horizontal) velocity must stay low enough that the vertical rise dominates, otherwise aggregates are dragged toward the outlet before they reach the float.
where vcf = mean horizontal velocity (m/h), Ax = vertical cross-sectional area of flow = width × depth (m²). Keep vcf low (typically < 40–60 m/h) and comparable to or below the aggregate rise velocity. Hydraulic stability is judged by the densimetric Froude number Fr = v / √(g′H) and the flow Reynolds number Re = vRh/ν; here g′ = gΔρ/ρ is reduced gravity across the warm/cold or clean/laden density interface.
Check for Example 1: take a separation zone 3.0 m wide and 2.4 m deep, so Ax = 7.2 m². Then vcf = 180 / 7.2 = 25 m/h (6.9×10−3 m/s). Mean horizontal residence time across a 7.5 m long zone is 7.5 / 6.9×10−3 ≈ 1,090 s (~18 min), comfortably longer than the ~2–3 min an aggregate needs to rise 2.4 m at 8 m/h. The ratio vcf/vb at the design rise rate is ~25/8 ≈ 3 near the inlet but falls sharply as flow spreads, confirming the layout is rise-dominated away from the baffle.
The danger is that these mean velocities hide reality. A rising bubble cloud is lighter than the bulk; incoming recycle is often warmer. Both create buoyant density currents that short-circuit along the surface or plunge along the floor, so parts of the tank are dead while others carry many times the mean flux. A low densimetric Froude number (Fr « 1) signals a stratification-dominated regime prone to such currents. This is precisely why physical hand calculations are validated with computational fluid dynamics of the flotation basin before fabrication — the mean-flow arithmetic cannot see a density current, but a CFD model can.
How do the float beach and effluent laterals keep flow uniform?
Uniform flow is engineered at the two boundaries of the separation zone — the top (float removal) and the bottom (clarified-water collection).
- Float removal — weir/beach. Accumulated float is removed either by a slowly submerging beach (an inclined ramp over which a surface scraper pushes the float) or by intermittent flooding over a fixed weir. A gently sloped beach with a low, uniform scraper speed avoids drawing bulk water up with the float, keeping the surface hydraulics quiescent and the float solids concentration high (typically 2–5% dry solids).
- Effluent collection — perforated laterals. Clarified water is drawn off through a manifold of perforated pipes or a false floor spanning the full tank width. The collection must be hydraulically uniform: if one lateral or region draws disproportionate flow, it creates a localised upward velocity that exceeds the design rise rate and pulls aggregates down into the effluent.
The design rule for uniform draw-off is that head loss through the collector orifices should dominate the head loss along the lateral, so flow splits evenly regardless of position. A common target is that orifice (port) head loss is at least 3–4× the manifold friction loss, which holds the flow distribution to within a few percent across all ports. Undersized or unevenly spaced orifices are a frequent cause of the streaky, patchy clarification that CFD later diagnoses as a collection-induced short-circuit.
DAF tank internal design sequence
- Fix total hydraulic flow. Sum feed and recycle (Q + Qr). This total, not the feed alone, drives every zone calculation because the recycle leaves with the clarified effluent.
- Size the separation zone. Choose a design rise rate (SLR) — 5–12 m/h low-rate, 15–40 m/h high-rate — and set As = (Q + Qr) / SLR. Confirm SLR is below the aggregate rise velocity.
- Size the contact zone. Set contact-zone volume Vc from a 60–120 s contact time and plan area Ac from a 20–80 m/h upflow velocity; the depth follows as Vc / Ac.
- Check cross-flow velocity. Compute vcf = (Q + Qr) / Ax through the vertical cross-section and confirm it is low relative to the aggregate rise velocity, so separation is rise-dominated.
- Design flow-uniformity features. Specify a low-speed beach/weir for float removal and perforated effluent laterals sized so orifice head loss dominates manifold friction (roughly 3–4×).
- Validate hydraulics with CFD. Model density currents and short-circuiting, especially for high-rate designs with little hydraulic margin, before committing to fabrication.
Frequently asked questions
What is the difference between the contact zone and the separation zone in a DAF tank?
The contact (reaction) zone is a narrow, fast-upflow region where released microbubbles collide with and attach to coagulated flocs, forming buoyant aggregates in about 60–120 seconds. The separation (clarification) zone is a wide, quiescent basin where those aggregates rise to the float and clarified water is collected below. They are sized by different equations — kinetics for the first, rise-rate hydraulics for the second.
What surface loading rate should a DAF tank be designed for?
Conventional low-rate DAF is designed for a separation-zone surface loading (rise) rate of about 5–12 m/h, relying on discrete buoyant rise of aggregates. High-rate DAF operates at 15–40 m/h by exploiting hindered, stratified bulk rise of a dense bubble cloud, often with lamella inserts. The recycle flow counts toward the loading, so it raises the hydraulic load as well as the air supply.
Why does cross-flow velocity matter in a DAF separation zone?
Clarified water moves horizontally to the effluent laterals while aggregates rise vertically. If the horizontal (cross-flow) velocity is too high relative to the aggregate rise velocity, aggregates are dragged toward the outlet before reaching the float and escape with the effluent. Designers keep cross-flow velocity low — typically below 40–60 m/h and comparable to or under the rise velocity.
What causes short-circuiting in a DAF tank?
Short-circuiting is driven mainly by density currents: the rising bubble cloud is lighter than the bulk water and the warm recycle is less dense, so flow stratifies and travels along the surface or floor instead of filling the tank uniformly. Uneven effluent draw-off adds to it. A low densimetric Froude number signals a stratification-prone regime, which is why CFD is used to check high-rate designs.
How is float removed from a DAF and why does the method matter?
Float is removed either by a slow surface scraper pushing it up an inclined beach, or by intermittent flooding over a fixed weir. The method matters because aggressive or fast removal draws bulk water up with the float, disturbing the surface hydraulics and diluting the float. A gently sloped beach at low scraper speed keeps clarification quiescent and the float solids high, around 2–5% dry solids.
Do you need CFD to design a DAF tank?
Not for a first-pass hand calculation, which sizes both zones adequately from loading rates and contact time. But mean-flow arithmetic cannot detect density currents or collection-induced short-circuits, which control real performance — especially in compact high-rate tanks with little hydraulic margin. CFD models these effects before fabrication and is strongly advised for high-value or high-rate designs.
Sources & further reading
- Edzwald, J.K. & Haarhoff, J. — Dissolved Air Flotation for Water Clarification (AWWA/McGraw-Hill, 2012)
- Wang, Hung & Shammas — Flotation Technology, Handbook of Environmental Engineering Vol. 12 (Humana Press, 2010)
- Edzwald, J.K., Dissolved air flotation and me (review), Water Research — flotation theory and contact/separation zone design
- Crittenden et al., MWH's Water Treatment: Principles and Design — flotation
- Metcalf & Eddy, Wastewater Engineering: Treatment and Resource Recovery — flotation
- AWWA — Water Treatment Plant Design