Fluid mechanics of chemical injection: jet penetration, turbulent diffusion, Reynolds number regimes, static mixer selection, and dead-zone elimination for complete homogenisation.
A dosing stream injected into a pipe is a turbulent round jet. Its penetration depth, decay rate and the distance to homogeneity are governed by the momentum flux ratio and the turbulent Schmidt number.
When a chemical is injected into a flowing water stream, the injected fluid forms a round turbulent jet. The key parameter is the jet-to-crossflow momentum flux ratio J:
ρj = injected fluid density (kg/m³)Vj = injection nozzle velocity (m/s)ρc = crossflow (main pipe) density (kg/m³)Vc = crossflow velocity (m/s)For J >> 1 the jet penetrates to the opposite wall; for J << 1 it is deflected immediately downstream. In water-treatment dosing the reagent is dilute and densities are nearly equal, so the ratio simplifies to the velocity ratio squared: J ≈ (Vj/Vc)². Typical design targets are Vj = 1.5–3.0 × Vc, giving J = 2.25–9.
The centreline concentration decay of a turbulent round jet follows a power law. For a free jet the excess concentration decays as x−1; in a confined pipe the walls reflect the jet and the decay is faster, roughly x−1.5 once the jet has impinged. The mixing length Lm to achieve ±5% concentration uniformity is empirically:
Lm = axial distance from injection point to ±5% uniformity (m)D = pipe diameter (m)J = momentum flux ratio (dimensionless)DN200 pipe (D = 0.20 m), crossflow Q = 80 m³/h, Vc = 0.71 m/s. Injection nozzle diameter 12 mm, pump flow 150 L/h, Vj = 1.47 m/s.
Chemical transport in a turbulent pipe flow is governed by the interplay between turbulent advection and molecular diffusion, captured by the turbulent Schmidt number.
The turbulent Schmidt number Sct relates turbulent momentum diffusivity (νt) to turbulent mass diffusivity (Dt):
In pipe flows Sct is typically 0.5–0.9 for passive scalars in water. Lower values (0.3–0.5) are found in jets and high-shear regions where coherent structures enhance scalar transport. The turbulent mass diffusivity is therefore 10³–10&sup4; times larger than molecular diffusivity, which is why rapid mixing is achievable in centimetres rather than metres.
The Reynolds number of the main pipe flow determines the turbulence intensity and the eddy size distribution:
Re < 2,300: laminar — dosing is diffusion-limited and very slow; avoid.2,300 < Re < 4,000: transitional — mixing is unpredictable.Re > 4,000: turbulent — dosing jet mixes by turbulent entrainment; design here.For turbulent pipe flow the friction factor λ (Darcy–Weisbach) sets the wall shear stress and the turbulence intensity. The root-mean-square velocity fluctuation u' is roughly:
This turbulence level is what tears the dosing jet apart and folds it into the bulk flow. At Re = 50,000 (typical for DN150 at 1 m/s) the integral length scale of turbulence is roughly 0.05D (7.5 mm) and the Kolmogorov scale is ∼30 μm — three orders of magnitude smaller than the jet diameter, guaranteeing that viscous dissipation destroys concentration gradients at the microscale.
Static mixers reduce the mixing length by one to two orders of magnitude by forcing successive divisions and re-orientations of the fluid. The pressure drop and mixing quality trade-off is quantifiable.
A static mixer works by splitting, stretching and recombining the flow. Each element doubles the number of striations. After n elements the striation thickness is reduced by 2n. To achieve homogeneity the striation thickness must be reduced to the Batchelor scale lB:
lK = Kolmogorov microscale (m), lK = (ν³/ε)1/4Sc = molecular Schmidt number (ν/Dm), ∼500–1,000 for common dosing reagents in waterlB ≈ 1–5 μm for typical water-treatment pipe flowsThe pressure drop across a static mixer is higher than empty pipe. The multiplier K depends on mixer type:
180° twist elements alternating left/right. K ≈ 3–5 vs empty pipe. Good for low-viscosity fluids. 6 elements typically sufficient for ±5%.
Crossed-bar lattice. K ≈ 10–20. Very high mixing efficiency; 3–4 elements achieve ±1%. Preferred for viscous or gassing reagents.
K ≈ 5–8. Intermediate pressure drop. Effective for two-phase gas–liquid mixing (e.g. CO2 dissolution for pH).
High-shear static mixers can degrade polymer flocculants. For polymer dosing use low-shear injection with long straight-pipe mixing (L/D > 20) or specialised low-shear mixer elements.
DN100 pipe, Q = 25 m³/h, V = 0.88 m/s, water at 20°C. Six helical elements, K = 4. Empty-pipe λ ≈ 0.019.
Poor injection geometry creates stagnation regions where chemical accumulates or bypasses. CFD and tracer studies reveal the failure modes.
A tracer study with a pulse injection of dye or salt measures the residence time distribution (RTD). The dimensionless variance σθ² characterises the spread:
σθ² = 0: perfect plug flow (no mixing)σθ² = 1: perfect CSTR (complete mixing)0.05–0.15: good pipe mixing with static mixer> 0.3: significant dead zones or bypassing — redesign requiredCommon failure modes and fixes:
Injecting flush with the pipe wall creates a wall jet that hugs the boundary. 30–50% of chemical can remain near the wall for >10D. Fix: use a radial injection lance projecting to 0.3–0.5R, or inject at 45° into the flow.
Injecting immediately downstream of an elbow places the jet in the low-pressure separation zone on the outer radius. Fix: inject 3–5D downstream, or use the elbow's natural secondary flow as a mixer.
Pipe expansions, partially closed valves, or oversized pipe create low-Re zones where mixing stalls. Fix: size the injection pipe for Re > 4,000; avoid diameter changes within 10D of injection.
A lance with 3–4 radial orifices at 90° spacing injects into multiple quadrants simultaneously. At Qj,total = 0.5–2% of main flow, this achieves ±5% in 2–3D without a static mixer.
We model mixing with CFD, specify static mixers, and size lances for your pipe geometry, reagent and flow range.
Our expertise spans multiple industries with sector-specific water treatment solutions.