DAF Nozzle Design: The Pressure-Release Device That Makes the Bubbles

Why does the pressure-release nozzle set bubble size?

In a dissolved-air flotation unit, air is dissolved into a recycle stream at 4–6 bar in the saturator, but no bubbles exist there — the air is in solution. Bubbles only form when that stream is throttled back to atmospheric pressure across the pressure-release device. Henry's law fixes how much air is dissolved; the sudden pressure drop makes the water massively supersaturated, and the excess air comes out of solution as a cloud of micro-bubbles.

The physics that matters is nucleation kinetics. A rapid, high-shear pressure release creates a very high local supersaturation ratio over a very short distance, favouring a large number of small nuclei — the desired ~30–50 µm bubbles. A slow or low-shear release lets fewer nuclei grow larger and coalesce, producing coarse 100–1000 µm bubbles that rise too fast, shear the floc, and short-circuit the DAF contact zone instead of gently lifting solids. In other words, the nozzle — not the saturator — is the component that governs bubble-size distribution.

The mechanism links back to the physics covered in DAF bubble dynamics and the upstream duty set by DAF saturator design: the saturator decides how much air is available, the nozzle decides what that air becomes.

What controls bubble size at the nozzle — pressure-drop rate and shear

Two coupled effects at the orifice control the final bubble-size distribution:

• Rate of pressure drop (dP/dt). The faster the saturated water crosses from saturator pressure to atmospheric, the higher the transient supersaturation and the more nuclei form simultaneously. High dP/dt ⇒ many small bubbles.
• Shear rate at release. Intense turbulent shear immediately downstream of the orifice breaks up incipient bubbles before they can grow and coalesce. High shear ⇒ tighter, smaller distribution.

This is why a well-designed fixed orifice or needle valve outperforms a simple slow-opening gate: it maximises both dP/dt and ε at the release point. It also explains why over-large nozzles (low velocity, low shear) produce visibly milky-but-coarse white water that floats poorly.

A useful dimensionless check is the Weber number, We = ρ·v²·d_b / σ, which compares disruptive inertial force to the stabilising surface tension. Bubbles break up while We exceeds a critical value (~1–4); the maximum stable bubble size is reached when We falls to critical. Because We scales with v², the ~20 m/s release jet keeps We high enough to shear nascent bubbles down into the target band before they detach and rise. Once in the quiescent contact zone, shear collapses, We drops below critical, and the surviving 40 µm bubbles are stable — exactly the outcome DAF needs.

How do you size the nozzle orifice area from the recycle flow?

Flow through the nozzle is turbulent orifice flow. Neglecting the small atmospheric-side and elevation terms, the discharge through a single orifice is:

Worked example 1 — number of nozzles. Size the pressure-release nozzles for a DAF whose recycle stream is Q_r = 36 m³/h at a saturator gauge pressure giving ΔP = 5 bar = 5×10⁵ Pa. Use C_d = 0.62.

• Convert flow: Q_r = 36 / 3600 = 0.0100 m³/s.
• Release velocity term: √(2·ΔP/ρ) = √(2·5×10⁵ / 1000) = √1000 = 31.6 m/s.
• Required total area: A_total = Q_r / (C_d·31.6) = 0.0100 / (0.62·31.6) = 0.0100 / 19.6 = 5.1×10⁻⁴ m² (≈510 mm²).
• Pick a nozzle bore: for a 3 mm orifice, a₁ = π(0.003)²/4 = 7.07×10⁻⁶ m² (7.07 mm²).
• Number of nozzles: N = A_total / a₁ = 5.1×10⁻⁴ / 7.07×10⁻⁶ = 72 → use ~72 nozzles of 3 mm bore.

Choosing a larger 4 mm bore instead gives a₁ = 12.6 mm², so N = 510 / 12.6 ≈ 41 nozzles — fewer, but each passes more flow at lower per-orifice velocity, which trades manufacturing simplicity against shear intensity.

How do you check velocity and shear at the nozzle?

Sizing the area is not enough — the per-orifice velocity must be high enough to generate the shear that keeps bubbles fine, without being so high that erosion or cavitation dominate. The orifice velocity is simply the release-velocity term scaled by C_d:

Worked example 2 — velocity and Reynolds check for the 3 mm nozzle above at ΔP = 5 bar:

• Orifice velocity: v = 0.62 × 31.6 = 19.6 m/s through each 3 mm bore.
• Jet Reynolds number: Re = (1000 × 19.6 × 0.003) / 1.0×10⁻³ = 58,800 / 0.001... = 5.9×10⁴.
• Verdict: Re ≈ 59,000 >> 10⁴, so the jet is well into the turbulent regime — good bubble break-up and high ε. Velocity ≈20 m/s is in the normal 15–25 m/s window; below ~10 m/s bubbles coarsen, above ~30 m/s erosion of the orifice edge accelerates.

Because the design tool becomes CFD once you need the actual local dissipation field, CFD of the nozzle jet and contact zone is used on demanding duties to confirm that shear is high where bubbles nucleate and that the white water spreads evenly rather than jetting to one wall.

What happens to nozzle count as you change ΔP?

Because Q ∝ √ΔP, halving the pressure drop does not halve the flow per nozzle — it reduces it only by √0.5 ≈ 0.71. So a lower saturator pressure needs more nozzles to pass the same recycle, and each releases at lower velocity (weaker shear, coarser bubbles). The table below holds Q_r = 36 m³/h, C_d = 0.62 and a 3 mm bore, and solves N and v at each ΔP.

The trend is the design lesson: higher ΔP gives fewer, higher-shear nozzles and finer bubbles, at the cost of more compressor/pump energy in the saturator. Most industrial DAF units settle at 4–6 bar as the economic optimum between bubble quality and energy.

How are nozzles spaced across the contact zone?

Even white-water distribution matters as much as bubble size. If nozzles are clustered, part of the contact zone is starved of bubbles and floc passes through unfloated, while over-aerated regions generate turbulence that shears the float. Nozzles are therefore mounted on a distribution manifold spanning the full width of the contact zone, typically one nozzle every 75–200 mm of header length, sized so the manifold friction loss is small (<5%) relative to ΔP so every nozzle sees near-identical pressure.

Worked example 3 — spacing. Distribute the 72 nozzles from Example 1 across a contact zone 3.0 m wide fed by two headers (36 nozzles each): pitch = 3000 mm / 36 = ~83 mm centres, comfortably inside the 75–200 mm guideline. The header cross-section is then sized so header velocity stays below ~1 m/s, keeping the pressure profile flat along its length so the first and last nozzle deliver the same jet velocity and hence the same bubble size.

Needle valves, fixed orifices or proprietary WRc-type nozzles?

Three device families dominate. The choice trades adjustability, bubble quality, clogging resistance and cost.

Materials follow the erosion/clogging risk: stainless steel (316), duplex, or hard ceramic/tungsten-carbide inserts at the orifice throat where the ~20 m/s jet and any grit cause wear. Because narrow bores block on fibres and struvite, a coarse strainer on the recycle line and periodic clean-in-place are standard. Needle valves give the most flexibility for R&D and variable flows; fixed orifices are cheapest and most robust; WRc-pattern nozzles give the best bubble quality per unit and are the usual choice where flotation performance is guaranteed.

Common nozzle design and operation mistakes

• Over-sizing the orifice. Low jet velocity means low shear and coarse, coalesced bubbles — the single most common cause of poor float despite adequate dissolved air.
• Running the saturator too low. Dropping below ~3–4 bar to save energy collapses dP/dt and coarsens bubbles; the white water looks grey and rises in slugs.
• Uneven spacing / oversized headers. Header friction >5% of ΔP makes end nozzles run slower, so bubble size varies across the tank.
• No recycle strainer. Fibres, grit and scale block fine bores, driving flow to the open nozzles and destroying distribution.
• Wrong throat material. Plain mild steel or soft brass erodes, the bore widens, velocity falls and bubbles coarsen within months.

How to size DAF pressure-release nozzles

1. Fix the recycle flow and pressure drop. Take Qr from the saturator/air-to-solids sizing and set ΔP as the saturator absolute pressure minus atmospheric (typically 4–6 bar).
2. Compute the total orifice area. Apply A_total = Qr / (Cd·√(2ΔP/ρ)) with Cd ≈ 0.6–0.65 for a sharp orifice, using SI units.
3. Choose a bore and get the nozzle count. Pick a 2–4 mm bore, compute single-orifice area a1 = πd²/4, then N = A_total / a1 and round up.
4. Check jet velocity and Reynolds number. Verify v = Cd·√(2ΔP/ρ) sits in ~15–25 m/s and Re > 10⁴ so the release is turbulent and bubbles stay fine.
5. Set the spacing across the contact zone. Distribute nozzles at 75–200 mm centres on a manifold sized so header loss is <5% of ΔP for even white water.
6. Select device type and materials. Choose needle valve, fixed orifice or WRc-type nozzle; specify 316/duplex or ceramic throats and a recycle strainer.