We migrated a 25 L pilot emulsion to 1,000 L/hr by replacing the tank impeller with a 7.5 kW inline rotor–stator plus a powder eductor, which cut blend time 35% but pushed D90 from 18 to 24 µm. Has anyone tuned tip speed or recirculation loop design to recover droplet size without driving up kWh/kg, and which head geometry scales cleanly?
I’d try a fine slotted/emulsor head and a small throttling valve to hold about 1.5–2 bar across the rotor–stator; that pulled D90 back near 18 µm for us at about 21 m/s with little change in kWh/kg. Where’s your eductor tee, because shifting it right up to the inlet and shortening the suction run helps the concentrate get one clean pass — like giving the droplets a second lap. @OP, if you’re on a single‑stage, the emulsor/multi‑tooth geometries from IKA/Silverson tend to scale cleanly: https://www.silverson.com/en/products/in-line-mixers/.
Coarse + multitooth heads in series recovered D90 for us; @OP try 3–4 turnovers, watch inlet NPSH?
What moved the needle for us was moving the oil addition into a side‑stream tee 150–200 mm upstream of the rotor and running it through a very short static mixer (3–5 elements) so the rotor sees a tight pre‑emulsion; that pulled D90 back to about 18 µm at about 19–20 m/s without a kWh/kg bump. Small caveat: keep the static mixer ΔP under about 0.3 bar and purge air, or you’ll trade droplet size for heat and foam, @OP.
On a dairy emulsion we tightened the size tail by adding a small vertical vent pot 200–300 mm upstream of the rotor–stator to de‑air the eductor stream; same tip speed, D90 dropped from about 24 to about 18 µm without extra power. Keep a bit of inlet head so you don’t flirt with cavitation — worth a quick try, @OP?
@simonB78’s NPSH point is spot on — add a small discharge throttle and hold about 1–1.5 bar after the rotor–stator; that recovered our droplet tail at the same tip speed by suppressing micro‑cavitation and nudging residence time. Watch a slight heat bump and seal load; what discharge pressure are you running now?
Try preloading the water with all the emulsifier and warming it to about 35–40°C so viscosity drops and adsorption is quick; that pulled our D90 back about 4–6 µm at the same power draw. If you can, run a two-stage head — coarse slot then fine tooth — ‘break then polish’ — it scaled cleanly for us, though you’ll want to watch the powder leg for any starvation. What’s your aqueous viscosity at process temperature?