Industrial Mixers and Agitators: Impeller Selection, Baffles and Scale-Up

Quick Answer

Industrial agitator selection starts with the mixing duty, not the tank. Blending low-viscosity liquids needs high flow and low shear — hydrofoil impellers at a power number around 0.3, drawing little power while moving large volumes. Gas dispersion and emulsification need high shear — Rushton disc turbines at a power number near 5, concentrating energy at the blade tips. Solids suspension sits between, usually served by a 45° pitched-blade turbine at a power number near 1.3. Baffles are not optional in an unbaffled vertical tank: four baffles at 90°, each T/12 wide (one-twelfth of tank diameter), set off the wall by about one-sixth of their width, convert useless swirling rotation into genuine top-to-bottom turnover. Impeller diameter is typically D/T = 0.3-0.4 for turbulent blending and 0.5-0.6 for viscous duty. Power draw follows P = Np · ρ · N³ · D⁵, so doubling impeller speed increases power eightfold and doubling diameter increases it thirty-twofold — which is why agitator motor sizing is unforgiving of guesswork. Scale-up holds a defined quantity constant: equal tip speed for shear-sensitive duties, equal power per unit volume for blending, and equal just-suspended speed for solids.

A paint plant in Turkey commissioned a new 30 m³ blending tank. The agitator was sized by a supplier who asked for the tank volume and the fluid viscosity and nothing else.

The mixer worked. The problem was that it worked as a very expensive stirrer. Pigment settled on the tank floor within an hour of stopping, and the batch had a visible shade gradient from top to bottom that took forty minutes of extra mixing to resolve, every batch, four batches a shift.

The tank had no baffles. The impeller was a flat-blade turbine mounted centrally. What they had built was a device that spun the entire tank contents around the shaft as a solid body — fast rotation, almost no axial movement, a deep vortex, and surprisingly little actual mixing. Add four baffles and change the impeller to a 45° pitched blade and the same motor achieved uniform blending in eight minutes.

Mixing is one of the few unit operations where the difference between a good design and a bad one is not 20%, it is a factor of five. And the difference often costs less than the paint in one batch.

What Mixing Actually Requires

Mixing is the conversion of shaft power into fluid motion of a useful kind. There are two kinds of useful motion and they trade off against each other:

Flow (pumping capacity) — the bulk volume the impeller moves per unit time. Flow carries material around the tank, turns the contents over, suspends solids, and equalises concentration and temperature.

Shear — the velocity gradient, concentrated near the blade tips. Shear breaks droplets and bubbles, disperses agglomerates, and creates interfacial area.

For a given power input, you choose flow or shear. You cannot maximise both.

Identify the duty before anything else. Everything downstream follows from it.

Impeller Types

Axial flow impellers

Discharge along the shaft axis, producing a top-to-bottom circulation loop.

Hydrofoil impellers (three or four cambered blades, airfoil profile) are the modern standard for flow-controlled duty.

• Power number Np ≈ 0.3
• Flow number Nq ≈ 0.55-0.65
• Highest flow per unit power of any impeller type
• Uses: blending, heat transfer, solids suspension of easily suspended solids, crystallisers

Marine propellers (three blades, cast or fabricated)

• Np ≈ 0.3-0.5
• Robust, inexpensive in small sizes
• Mostly in small tanks and portable mixers, and as side-entry units

Pitched-blade turbines (four to six flat blades at 45°)

• Np ≈ 1.3
• Produces both axial flow and useful shear — a genuine compromise impeller
• The default for solids suspension and for mixed-duty service
• Robust and cheap to fabricate in any alloy

Radial flow impellers

Discharge outward to the tank wall, creating two circulation loops — one above the impeller and one below.

Rushton disc turbine (six flat blades mounted on a disc)

• Np ≈ 5.0
• The classic high-shear impeller. The disc prevents gas from short-circuiting up the shaft, which is why it is the standard for gas dispersion.
• Uses: gas-liquid dispersion, fermentation, emulsification
• Power draw drops sharply once gas is introduced — "gassed power" can be 40-60% of ungassed power, which matters enormously for motor sizing and for what happens if the gas supply fails

Concave-blade and hollow-blade disc turbines (Smith turbine, BT-6 and similar)

• Np ≈ 2.5-4.0
• Hold gassed power much closer to ungassed power and handle higher gas rates before flooding
• The better choice in modern gas dispersion design

Flat-blade turbine (four to six flat vertical blades, no disc)

• Np ≈ 3.5-4.0
• General radial mixing; largely superseded by the above

Close-clearance impellers for viscous fluids

Above roughly 10,000-20,000 cP, turbulent mixing is no longer achievable and turbine impellers simply carve a cavern in the fluid while the rest sits still.

Helical ribbon — a ribbon following a helix close to the wall, pushing material down the centre and up the wall (or vice versa). Handles 100,000 to several million cP.

Anchor — a frame closely following the tank contour, often with wall scrapers. Excellent for heat transfer in viscous and fouling service; poor at top-to-bottom mixing on its own, so often combined with a coaxial high-speed inner impeller.

Coaxial systems — a slow outer anchor or ribbon for bulk movement plus a fast inner disperser for local shear. The standard for adhesives, sealants, high-viscosity polymers and thick pastes.

High-shear devices

Sawtooth disperser (Cowles blade) — a serrated disc running at 15-25 m/s tip speed. Creates intense local shear for powder wetting and pigment dispersion. Poor bulk flow, so almost always paired with a slow-speed sweep impeller.

Rotor-stator mixers — a rotor turning inside a slotted stator at very small clearance, generating shear rates in the 10⁵ s⁻¹ range. Used for emulsification, homogenisation and dispersing difficult powders. Available as in-tank batch units and as inline units.

Power Draw and the Cube Law

Power consumed by a turbulent impeller:

P = Np · ρ · N³ · D⁵

Where P is power (W), Np is the dimensionless power number, ρ is density (kg/m³), N is rotational speed (rev/s) and D is impeller diameter (m).

Two consequences dominate all agitator design:

Speed cubed. Increase speed by 26% and power rises 100%. A nominal 1,450 rpm motor on a 20:1 gearbox gives 72.5 rpm; if someone changes the gearbox to 15:1 to "mix a bit better," speed becomes 96.7 rpm and power demand rises by a factor of 2.37. The motor trips, or the shaft fails.

Diameter to the fifth. Increase impeller diameter 20% and power rises by 1.2⁵ = 2.49. Fitting a slightly larger impeller because the old one was worn is a common and destructive field modification.

Reynolds number for mixing determines regime:

Re = ρ · N · D² / μ

• Re > 10,000 — fully turbulent, Np is constant
• Re 10-10,000 — transitional, Np varies
• Re < 10 — laminar, Np is inversely proportional to Re

In the laminar region, power is proportional to N² and D³, not N³ and D⁵. Using turbulent correlations for a viscous fluid gives badly wrong answers.

Density and viscosity at the worst case, not the normal case. Power scales directly with density. An agitator sized for a 1,100 kg/m³ slurry that is sometimes run with a 1,450 kg/m³ batch is 32% overloaded in that batch. Size the motor for the heaviest and most viscous batch the tank will ever see, plus a margin — 15-20% over calculated power is normal practice, and more if the fluid properties are uncertain.

Baffles

In an unbaffled vertical tank with a centrally mounted impeller, the whole fluid mass rotates with the impeller. This is solid-body rotation — fast movement, no mixing. A vortex forms, deepens, and eventually reaches the impeller, at which point air is entrained and power draw collapses.

Standard baffle configuration:

• Four baffles, spaced 90° apart
• Width = T/12 (one-twelfth of tank diameter); T/10 is also used
• Off-wall clearance ≈ one-sixth of baffle width — around T/72. The gap prevents stagnant zones and solids build-up behind the baffle.
• Height: from near the tank bottom to just below the maximum liquid level
• In tanks with heating coils, the coils provide partial baffling, but dedicated baffles are still usually required

Baffles convert rotation into axial and radial flow. They also increase power draw substantially — a baffled tank draws considerably more power at the same speed than an unbaffled one, because the impeller is now doing useful work instead of spinning the contents.

When baffles are not needed:

• Off-centre mounting — a top-entry mixer mounted at about one-third tank radius from centre and angled 10-15° from vertical creates asymmetric flow that breaks up solid-body rotation. Common for small portable mixers and for tanks where baffles would interfere with cleaning. The penalty is a strong bending moment on the shaft.
• Viscous laminar mixing — no swirl develops; close-clearance impellers need no baffles.
• Side-entry mixers — mounted in the tank wall, angled horizontally, creating a rotational flow field deliberately. Standard for large storage tanks and for crude oil tank sludge control.
• Square or rectangular tanks — the corners provide baffling, though usually imperfectly.

Tank Geometry and Impeller Placement

Standard geometry for a baffled turbulent tank:

• Liquid height = tank diameter (H/T = 1) for a single impeller
• Impeller diameter D/T = 0.3-0.4 for turbulent blending; 0.4-0.5 for solids suspension; 0.5-0.6 for viscous duty
• Bottom clearance C = D or T/3 for a single impeller
• H/T up to about 1.2 is served by one impeller

Multiple impellers are needed when H/T exceeds roughly 1.2:

• Spacing of about one impeller diameter to 1.5 D between impellers
• A common arrangement is a hydrofoil at the top for flow and a disc turbine at the bottom for gas dispersion or shear
• In tall fermenters, three or four impellers on one shaft is standard

Dished or conical bottoms suspend solids better than flat bottoms, because flat bottoms have a stagnant annulus at the wall-to-floor corner. A dished bottom with a pitched-blade turbine at C = T/4 gives much better floor sweeping than a flat bottom.

Solids Suspension

The key parameter is N_js — the just-suspended speed, defined by Zwietering as the speed at which no particle remains stationary on the tank floor for more than one or two seconds.

Zwietering's correlation:

N_js = S · ν^0.1 · d_p^0.2 · (g·Δρ/ρ_L)^0.45 · X^0.13 / D^0.85

Where S is a geometry factor, ν is kinematic viscosity, d_p is particle diameter, Δρ is the solid-liquid density difference, X is solids loading as a mass percentage, and D is impeller diameter.

What the exponents tell you:

• D^-0.85 — a larger impeller suspends solids at substantially lower speed. Large-diameter, low-speed is the efficient way to suspend solids.
• Δρ^0.45 — dense solids are much harder to suspend. Suspending sand is a different problem from suspending plastic pellets.
• d_p^0.2 — particle size matters less than people expect.
• X^0.13 — loading matters remarkably little. Going from 5% to 30% solids raises N_js by only about 25%.

Three suspension levels, with very different power requirements:

1. On-bottom motion — particles move but may remain on the floor. Lowest power. Adequate for dissolution of fast-dissolving solids.
2. Just-suspended (N_js) — no particle stationary for more than 1-2 s. The usual design target. Achieves essentially all available solid-liquid mass transfer area.
3. Uniform suspension — concentration uniform throughout the tank. Requires roughly 2 × N_js, therefore about 8 × the power. Only necessary when withdrawing a representative slurry through a side nozzle or feeding downstream equipment that requires consistent solids concentration.

Specifying uniform suspension when just-suspended would do is one of the most expensive mistakes in agitator sizing, because of the cube law.

Gas Dispersion

For gas-liquid contacting — fermentation, oxidation, hydrogenation, wastewater aeration, chlorination — the impeller must break the gas stream into fine bubbles and distribute them.

Two failure regimes to design away from:

Flooding. Gas rate too high for the impeller speed; gas rises past the impeller in large bubbles without being dispersed. Mass transfer collapses. The flooding boundary depends on gas flow number and Froude number; disc turbines flood at lower gas rates than concave-blade designs.

Gassed power reduction. Gas cavities form behind the blades, reducing the effective blade area. For a Rushton turbine, gassed power can fall to 40-50% of ungassed power at high gas rates. Two implications:

• Mass transfer is lower than ungassed power would suggest
• If the gas supply stops while the agitator runs, power draw jumps back to ungassed value. A motor sized on gassed power will trip or burn. Always size the motor on ungassed power.

Concave-blade and hollow-blade turbines hold gassed power at 70-85% of ungassed and tolerate much higher gas rates before flooding, which is why they have largely replaced flat-blade Rushton turbines in new designs.

Gas is introduced through a sparge ring below the impeller, typically at 0.8-1.0 D diameter, with holes sized for sonic or near-sonic discharge to ensure even distribution across all holes.

Scale-Up

Scale-up is the hardest part of mixing, because not all parameters can be held constant at once. Choosing which to hold constant is the whole decision.

Practical rules:

• Shear-sensitive duty (cell culture, flocculation, crystal growth, polymer solutions): hold tip speed constant. Typical limits are 2-4 m/s for fragile crystals and flocs, 4-8 m/s for general blending, and up to 15-25 m/s for sawtooth dispersers where shear is the point.
• Gas-liquid mass transfer and general turbulent blending: hold power per unit volume constant. Typical values are 0.1-0.3 kW/m³ for mild blending, 0.3-1.0 kW/m³ for solids suspension, 1-2 kW/m³ for gas dispersion and 2-5 kW/m³ for difficult dispersion duty.
• Solids suspension: do not scale at all. Calculate N_js directly at full scale from the Zwietering correlation.
• Geometric similarity should be preserved wherever possible — same D/T, C/T, baffle configuration and impeller type. Scaling up with a different geometry introduces a second variable and the result becomes guesswork.

Blend time scales badly. At constant P/V, blend time increases roughly with the two-thirds power of the scale ratio. A 10× volume scale-up at constant P/V increases blend time by about 2.2×. If blend time is critical, this must be accounted for in batch cycle planning rather than discovered at commissioning.

Mechanical Design

The hydraulics decide the impeller; the mechanics decide whether the unit survives.

Shaft design must consider:

• Torsional stress from transmitted torque
• Bending stress from hydraulic imbalance — in a baffled tank the load on the impeller fluctuates as blades pass baffles, producing a significant cyclic bending moment at the shaft seal and bearings
• Critical speed — operating speed must stay below about 70% of first natural frequency, or above 130% if running supercritical. Running near critical speed destroys shafts and seals quickly. Long unsupported shafts in deep tanks are the common offender.
• Deflection at the impeller, which loads the seal and the bearings

Bottom steady bearing. For long shafts, a bearing in the tank bottom reduces deflection and raises critical speed dramatically. The penalty is a wearing part inside the process fluid, which is unacceptable in sterile, abrasive or high-purity applications. Cantilever (unsupported) shafts are preferred where the process allows, which is why shaft and drive design become critical.

Seals, where the shaft enters a closed vessel:

• Lip seals — low pressure, non-hazardous, atmospheric tanks
• Packed gland — low pressure, tolerant, leaks a little, cheap
• Single mechanical seal — general closed-vessel duty
• Double mechanical seal with barrier fluid — pressure vessels, toxic or sterile service, zero emission
• Magnetic drive — no shaft penetration at all; for highly toxic, sterile or high-pressure duty. Limited torque capacity and higher cost.

Drive train: motor, gearbox or belt reduction, drive shaft, and a rigid mounting arrangement — usually a beam or mounting plate on the tank nozzle. The nozzle and its reinforcement must carry the full dynamic bending moment, not just the static weight. Agitator nozzle design failures on thin-walled tanks are common.

Common Specification Mistakes

1. No baffles in a vertical tank with a centre-mounted impeller. You get solid-body rotation and a vortex instead of mixing. The paint plant above.

   Prevention: Fit four baffles at T/12 width with off-wall clearance, or mount off-centre and angled, or use a side-entry unit.

2. Specifying the impeller before the duty. A Rushton turbine on a simple blending duty wastes most of its power on shear nobody needs; a hydrofoil on a gas dispersion duty floods immediately.

   Prevention: Define the duty first — blending, suspension, dispersion, heat transfer, emulsification — and select the impeller family from it.

3. Motor sized on gassed power. When the gas supply trips, power jumps back to ungassed and the motor trips or burns.

   Prevention: Always size the motor on ungassed power with a 15-20% margin.

4. Specifying uniform suspension when just-suspended is adequate. Uniform suspension requires about 2 × N_js, therefore around 8 × the power.

   Prevention: Specify just-suspended unless a side withdrawal nozzle or downstream process genuinely requires uniform concentration.

5. Using turbulent power correlations for a viscous fluid. In laminar flow, power scales with N² and D³, and Np is a function of Reynolds number.

   Prevention: Calculate the mixing Reynolds number first and use the correct regime, and switch to close-clearance impellers above 10,000-20,000 cP.

6. No check on shaft critical speed. A long shaft in a deep tank running near its first natural frequency fails at the seal and the bearings within weeks.

   Prevention: Require the vendor to state first critical speed and demonstrate operating speed is below 70% or above 130% of it.

7. Scaling up by holding blend time constant. Power per volume rises so steeply with scale that this is normally impossible.

   Prevention: Hold tip speed for shear-sensitive duty, power per volume for mass transfer, and calculate N_js directly for solids. Accept and plan for longer blend time at scale.

8. Agitator nozzle not designed for dynamic bending moment. The nozzle and its reinforcement are sized for the static weight, then the hydraulic imbalance cracks the shell.

   Prevention: Give the tank designer the agitator's dynamic bending moment and torque, not just its mass.

9. Field-fitting a larger impeller or changing the gearbox ratio. Power rises with D⁵ and N³; small changes have enormous consequences.

   Prevention: Treat any change in impeller diameter or speed as a redesign requiring a new power calculation and a check of motor, gearbox, shaft and seal.

10. Sizing on nominal fluid properties, not worst case. The heaviest and most viscous batch is the one that matters.

    Prevention: Specify the full range of density, viscosity and solids loading, and size on the worst combination.

Supply from Kasko Makine

Kasko Demir Çelik Makine supplies agitators, mixers and the vessels they go into:

Top-entry agitators

• Hydrofoil, propeller, pitched-blade turbine, flat-blade turbine, Rushton and concave disc turbine impellers
• Single and multiple impeller shafts
• Cantilever and bottom-steady-bearing shaft designs
• Open-tank and closed-vessel configurations
• Gear-reduced and belt-reduced drives
• Materials: carbon steel, 304/316L stainless, duplex 2205, Hastelloy, titanium, rubber-lined and glass-lined options

Specialised mixers

• Side-entry mixers for storage tanks and sludge control
• Bottom-entry mixers for sterile and pharmaceutical duty
• Portable clamp-mount and drum mixers
• Helical ribbon and anchor agitators for viscous products
• Coaxial systems combining slow sweep with high-speed dispersion
• Sawtooth dispersers for pigment and powder work
• Rotor-stator high-shear mixers, batch and inline
• Static (inline) mixers for continuous blending in pipework
• Magnetic-drive agitators for sealed and sterile vessels

Seals and drives

• Lip seals, packed glands, single and double mechanical seals with barrier systems
• Electric motors in IE3, IE4 and IE5, including Ex-rated for hazardous areas
• Gearboxes, couplings and variable frequency drives

Vessels and tanks

• Agitated tanks and reactors in carbon steel, stainless and alloy
• Jacketed and half-pipe-coil vessels for heating and cooling
• Baffle fabrication and retrofit
• Agitator nozzle reinforcement design and fabrication

Engineering support

Send the tank dimensions (diameter, straight side, bottom type), fluid properties (density, viscosity, solids content and particle size, settling behaviour), the mixing duty, batch volume range and required blend or suspension criterion, and we will return an agitator selection with impeller type, diameter, speed, calculated power and motor rating — with the design basis stated so you can check it rather than take it on trust. For a scale-up from an existing or pilot unit, send the existing geometry and performance and we will state which parameter has been held constant.

Certification

Material certificates to EN 10204 3.1, welding procedure and welder qualification documentation to ISO 15614 and ISO 9606, surface finish records for sanitary duty, dye penetrant and radiographic examination where specified, shaft critical speed calculations, and ATEX documentation for hazardous-area drives.

Logistics

Standard portable and small top-entry mixers generally ship in 2-4 weeks. Engineered agitators typically 8-14 weeks. Agitated vessels and reactors 12-20 weeks depending on size, material and inspection requirements. Shipping from Istanbul by road to Europe and the Caucasus, and by sea to Gulf, African and Asian destinations.

Send your tank and fluid data and we will return an agitator selection with the calculation basis within four working days. Reach us at info@kaskomakine.com or WhatsApp +90 (537) 521 1399.

Continue Reading: Process Equipment Series

• Mechanical Seals: API 682 Types and Piping Plans
• Industrial Electric Motors: IE Classes and Enclosures
• Industrial Gearboxes and Couplings

Frequently Asked Questions

Q: How do I select an impeller for an industrial mixer?
A: Start from the duty, not the tank. Blending and heat transfer need flow, so use a hydrofoil impeller with a power number around 0.3. Solids suspension needs balanced flow and shear, served by a 45° pitched-blade turbine at a power number near 1.3. Gas dispersion and emulsification need shear, so use a disc turbine with a power number of 2.5 to 5. Above about 10,000 cP, switch to helical ribbon or anchor impellers.

Q: Do mixing tanks need baffles?
A: Any vertical tank with a centre-mounted impeller in turbulent flow needs baffles, or the whole contents rotate as a solid body with a vortex and almost no mixing. The standard configuration is four baffles at 90° spacing, each one-twelfth of tank diameter wide, set off the wall by about one-sixth of their width. Alternatives are off-centre angled mounting, a side-entry mixer, or close-clearance impellers in viscous service where no swirl develops.

Q: How does agitator power change with speed and impeller size?
A: Turbulent power follows P = Np · ρ · N³ · D⁵. Power rises with the cube of speed and the fifth power of diameter, so a 26% speed increase doubles power and a 20% larger impeller increases power by 2.5 times. This is why changing a gearbox ratio or fitting a slightly larger impeller in the field frequently trips the motor or fails the shaft.

Q: What is just-suspended speed in solids mixing?
A: Just-suspended speed, N(js), is the agitator speed at which no particle remains stationary on the tank floor for more than one or two seconds, defined by Zwietering's correlation. It is the usual design target because it achieves essentially all available solid-liquid mass transfer. Uniform suspension throughout the tank requires roughly twice that speed and therefore about eight times the power, so it should be specified only when genuinely required.

Q: Why does agitator power drop when gas is introduced?
A: Gas cavities form behind the impeller blades, reducing effective blade area and therefore power draw. A flat-blade Rushton turbine can fall to 40 to 50% of its ungassed power at high gas rates, while concave-blade designs hold 70 to 85%. The practical consequence is that the motor must always be sized on ungassed power, because if the gas supply trips while the agitator runs, power jumps straight back to the ungassed value.

Q: How do you scale up a mixing process?
A: Choose one parameter to hold constant. For shear-sensitive materials such as cells, flocs and growing crystals, hold tip speed constant. For gas-liquid mass transfer and general turbulent blending, hold power per unit volume constant. For solids suspension, do not scale at all — calculate just-suspended speed directly at full scale. Keep geometric similarity in D/T, clearance and baffling, and expect blend time to rise with scale.

Q: What tip speed should an agitator run at?
A: Typical limits are 2 to 4 m/s for shear-sensitive duty such as flocculation and crystal growth, 4 to 8 m/s for general blending and solids suspension, 8 to 12 m/s for gas dispersion with disc turbines, and 15 to 25 m/s for sawtooth dispersers where intense local shear is the objective. Tip speed is the parameter to hold constant when scaling up any duty where shear damage matters.