DAF Saturator Design: Engineering the Air-Dissolution System

What does the saturator actually do?

Dissolved-air flotation works because air is far more soluble in water under pressure than at atmosphere. The saturator holds a fraction of the clarified effluent (the recycle, Q_r) at 4–6 bar while contacting it with compressed air. When this air-charged water is injected into the flotation tank through a nozzle and drops to atmospheric pressure, the excess dissolved air comes out of solution as 20–100 µm bubbles that attach to and lift the flocculated solids.

The design objective is deceptively simple: dissolve the maximum air mass into the smallest, cheapest vessel, reliably, across the plant's temperature range. Everything below serves that goal. For the wider unit context see our companion note on how a dissolved-air flotation unit works, and for hydraulic sizing of the tank itself see how to size a DAF system.

How much air can water actually dissolve? (Henry's law)

Air solubility is governed by Henry's law: the equilibrium mass of a gas dissolved in a liquid is proportional to its partial pressure. Because air is ~78% N₂ and ~21% O₂, the practical figure engineers use is the total air saturation concentration S_a at atmospheric pressure, which falls steeply with temperature.

The table below gives S_a for air in clean water. The key design consequence: water at 30 °C holds roughly 30% less air than at 5 °C, so a saturator sized only for winter will under-deliver in summer.

Values are for air (N₂+O₂+Ar) in equilibrium with the atmosphere and vary by a few percent between references; use site-specific data where high precision matters.

The available-air equation, term by term

The mass of air released per unit volume treated — the quantity that ultimately floats the solids — is captured by the available-air equation. It links the saturator conditions to the air-to-solids budget you set when sizing the flotation tank.

The (f·P − 1) group is the crux. At atmosphere (P = 1) it is zero — no pressure, no released air. It says that only the air dissolved above what the water would hold at 1 atm is available to precipitate as bubbles. Raising P or f lifts A almost linearly; raising temperature cuts S_a and therefore A. Some texts write the saturator term as S_a(f·P − 1) and treat η separately; the physics is identical.

Worked example 1: air mass delivered at 5 bar, 15 °C

Take a DAF treating Q = 80 m³/h with a 50% recycle, a packed saturator at 5 bar gauge (6 bar-a ≈ 5.9 atm) and water at 15 °C.

• Inputs: S_a = 26.3 mg/L (from the table at 15 °C); f = 0.9 (packed); P = 5.9 atm; Q_r/Q = 0.5.
• Saturator term: f·P − 1 = (0.9 × 5.9) − 1 = 5.31 − 1 = 4.31.
• Available air per unit feed: A = 26.3 × 4.31 × 0.5 = 56.7 mg/L of feed.
• Air mass rate: 56.7 g/m³ × 80 m³/h = 4,536 g/h ≈ 4.5 kg air/h.
• A/S check: if the feed carries 1,200 mg/L TSS, solids load = 80 m³/h × 1.2 kg/m³ = 96 kg/h. A/S = 4.5 / 96 = 0.047 kg air/kg solids — comfortably inside the usual 0.01–0.06 window.

Had the vessel been unpacked (f = 0.65) the saturator term would fall to (0.65 × 5.9) − 1 = 2.84, giving A = 37.4 mg/L and just 3.0 kg air/h — a 34% loss of float capacity from the same pressure and recycle. That single factor is why most modern high-rate DAFs use packed towers.

Packed vs unpacked (sparged) saturators

An unpacked saturator is simply a pressure vessel with an air space; recycle water sprays or sparges through it and air dissolves across the gas–liquid interface. Mass transfer is limited by the modest interfacial area, so equilibrium saturation is only ~60–70% for typical retention times. A packed saturator fills the contact zone with plastic media (Pall rings, saddles, structured packing) that shears the falling film into a large, renewing interfacial area, pushing saturation to ~90% in a smaller vessel.

The mass-transfer rate follows the two-film model, N = k_La · (C_s − C), so packing works by raising the interfacial area a — often 3–5× — far more than by changing k_L. The trade is fouling risk and cleaning access, covered below.

Sizing the vessel: retention time, hydraulic loading and diameter

Two rules size the pressure vessel. First, the recycle must spend enough time in contact to approach equilibrium — a retention time of 1–3 minutes on the liquid volume. Second, the hydraulic loading on the packing cross-section (typically 40–80 m³/m²·h) must be high enough to wet the media but low enough to avoid flooding.

The vessel is then detailed to a code (e.g. PD 5500 / ASME VIII) for the design pressure, with an air space above the liquid controlled on level.

Worked example 2: sizing the saturator vessel

Continuing example 1 (Q_r = 40 m³/h recycle):

• Liquid volume at 2 min retention: t = 2/60 = 0.0333 h; V = 40 × 0.0333 = 1.33 m³ of liquid hold-up.
• Add air space: allow ~25% freeboard for the compressed-air cushion → total shell volume ≈ 1.33 / 0.75 = 1.78 m³.
• Packing cross-section at L_H = 60 m³/m²·h: A_x = 40 / 60 = 0.667 m².
• Diameter: D = √(4 × 0.667 / π) = √0.849 = 0.92 m.
• Packed height for the liquid volume: a 0.667 m² column holding ~1.33 m³ implies ~2.0 m of wetted height, a sensible tower aspect ratio.

So a roughly 0.9 m diameter × 2.5–3 m tall packed vessel, rated for 6 bar-a, meets both the retention-time and hydraulic-loading constraints. If the arithmetic gave an over-tall, narrow column you would lower L_H; a squat, wide one means raising it.

Choosing the saturator pressure (and why it sets bubble size)

Pressure enters the available-air equation linearly, so more pressure means more dissolved and more precipitated air. But there are diminishing returns and a bubble-size penalty. Most DAFs run 4–6 bar (gauge). Going higher raises compressor energy and, importantly, changes the bubble population: the pressure drop across the injection nozzle controls the supersaturation ratio and hence nucleation rate.

Higher saturator pressure and a well-designed needle valve or WhiteWater™-type nozzle produce more, finer bubbles (nearer 20–40 µm), which give more bubble surface area per unit air and better floc capture. Too low a pressure yields coarser 100 µm+ bubbles that rise fast but capture poorly. The bubble rise velocity follows Stokes' law, v = g(ρ_w−ρ_b)d²/18μ, so halving bubble diameter cuts rise velocity four-fold — micro-bubbles stay in contact with flocs far longer, which is exactly what you want. Below the packing, verifying the bubble–floc contact zone with CFD analysis of the contact and separation zones can de-risk a difficult duty.

Air feed, level control, and fouling

Air feed. Two arrangements dominate. A dedicated compressor feeds air into the vessel head under level/pressure control — precise but with rotating plant to maintain. An eductor (venturi) on the recycle pump discharge entrains atmospheric air, dissolving it as the stream is pressurised — no compressor, but less independent control of the air:water ratio and generally suited to lower-duty units.

Level control. The vessel runs partly full: too high a liquid level floods the packing and carries under-saturated water; too low starves the nozzle of water and passes free gas that makes coarse bubbles. A level transmitter modulating the air-bleed or the recycle throttle holds the interface within a narrow band.

Scaling and fouling. Packing offers a large surface for CaCO₃ scale (hard waters, especially where CO₂ strips as pressure drops), biofilm and grease. Fouled media lose interfacial area, dropping f back toward the unpacked value and quietly starving the DAF of air. Mitigations: use clarified (not raw) water as recycle, design for CIP/acid-wash access, keep hydraulic loading high enough to self-scour, and monitor saturator differential pressure as an early fouling indicator. For fats-and-grease-heavy streams, pairing with the guidance in water-treatment equipment selection helps avoid recycling FOG back through the packing.

Worked example 3: summer vs winter air delivery

Same unit as example 1 (packed, 5 bar-g = 5.9 atm, Q_r/Q = 0.5), but compare a 5 °C winter recycle against a 30 °C summer recycle. Only S_a changes.

• Winter (5 °C): S_a = 34.3 mg/L. A = 34.3 × 4.31 × 0.5 = 73.9 mg/L → 5.9 kg air/h at 80 m³/h.
• Summer (30 °C): S_a = 20.8 mg/L. A = 20.8 × 4.31 × 0.5 = 44.8 mg/L → 3.6 kg air/h.
• Seasonal swing: summer delivers 44.8 / 73.9 = 61% of the winter air mass — a 39% collapse from temperature alone.

Design rule: size the saturator for the warmest expected recycle temperature so the air-to-solids ratio still meets target in summer. Here, to restore the winter air mass in summer you would raise the recycle ratio from 0.50 to about 0.50 × (73.9/44.8) = 0.82, or lift the saturator pressure — either way, plan the recycle pump and vessel for the hot-day case, not the annual average.

Saturator design sequence

1. Fix the air-to-solids target. From float tests, set the required A/S (0.01–0.06 kg air/kg solids) and hence the air mass the saturator must deliver at design flow.
2. Select pressure and vessel type. Choose 4–6 bar and a packed tower for f ≈ 0.9 unless the duty is light or the water heavily scaling.
3. Solve the available-air equation for recycle ratio. Use A = Sa·(f·P − 1)·η·(Qr/Q) at the warmest temperature to back-calculate the recycle flow that meets the A/S target.
4. Size the vessel. V = Qr·t with t = 1–3 min; set diameter from hydraulic loading Ax = Qr/LH (40–80 m³/m²·h); add ~25% air-space freeboard.
5. Detail air feed and level control. Specify compressor or eductor, a level transmitter modulating air bleed/recycle, and the injection nozzle for fine bubbles.
6. Guard against fouling. Use clarified recycle water, design CIP access, monitor saturator differential pressure, and re-check f over the maintenance interval.