Case Study: Spray Drying Scale-Up, Nozzle Choice From Pilot to Production
Spray drying scale-up: how nozzle choice moves from pilot to production without losing particle size control, with sizing logic and published ranges.
Spray drying looks simple until the scale changes. A pilot rig turns liquid feed into powder at a few kilograms per hour, and the trial samples pass. At production scale the line misses the spec in the first week, and the search for a cause usually stops at the nozzle.
A process that runs on the pilot and fails in production usually fails at the nozzle first. The droplet size shifts because flow and pressure were scaled without the physics that binds them. The chamber wets out because a wider cone reaches the wall. Or the median particle size drifts out of band, and with it the fines and the moisture. Each is an atomisation fault before it is a drying fault.
Pilot units evaporate 5 to 25 kg/h of water. Production dryers run at hundreds to thousands of kg/h. This case is about that transfer. It follows a composite food ingredient and chemical processing project moving from a small pilot rig to a dryer sized for 300 to 600 kg/h of water evaporation. No single site was audited. The configuration is generic, the arithmetic is published nozzle physics, and the sizing logic transfers to your line.
Case Snapshot
The reference configuration below is the generic shape of a project of this class, and the starting point the sizing was built from.
| Item | Value |
|---|---|
| Industry | Food ingredient and chemical processing, composite and anonymized |
| Duty | Drying a heat-sensitive liquid feed to a free-flowing powder |
| Pilot scale | 5 to 25 kg/h water evaporation, single pressure swirl nozzle |
| Production target | 300 to 600 kg/h water evaporation |
| Feed | Pumpable slurry or solution at 40 to 55 percent solids |
| Nozzle path | Pressure swirl (hydraulic) nozzles, hollow cone |
| Product target | Median particle size and moisture spec held across the scale step |
This reference configuration is the generic shape of the scale-up, a composite of typical projects, not one audited site. Real projects differ in feed, solids, and tower. What transfers is the nozzle logic, which is geometry and published flow physics rather than plant specific findings.
The Challenge
The natural first move is to scale the nozzle that worked on the pilot: multiply the flow by thirty and order one larger unit. That move fails on published nozzle physics. For a pressure swirl nozzle at fixed pressure, the median droplet size grows as the orifice grows. Doubling flow at constant pressure does not keep the particle size, because the larger orifice makes a coarser spray. Published trends tie the median droplet to nozzle capacity raised to a power often cited between a fifth and a half. A thirty fold capacity jump can push the median up by a factor of two or more before the dryer has done anything wrong.
The droplet matters because the droplet becomes the particle. Each drop dries into one granule and keeps the size class set in the first milliseconds of atomisation. Smaller droplets dry faster, since surface area rises relative to volume as the drop shrinks. Published references treat drying time as scaling roughly with the square of droplet diameter. A median up by a third therefore stretches drying time, pushes moisture up, or forces hotter air than the feed allows.
The chamber is the second constraint. A pressure swirl nozzle makes a hollow cone, and the cone widens as it falls. The geometry is plain trigonometry: a cone with an included angle of 60 degrees gains about 0.58 m of radius per metre of fall, since tan 30 degrees is about 0.58. In a narrow tower the cone reaches the wall quickly, and wet droplets stick, build a deposit, and shed as off-spec lumps. Published tower guidance keeps the cone clear of the wall, coupling cone angle to chamber diameter and nozzle position.
Drying air is the third partner. Fine droplets follow the air stream; coarse droplets fly on their own momentum and reach the wall regardless of the air. In a co-current tower the air must carry the spray long enough for drops to dry, and published references work with dwell times measured in tens of seconds. Uneven air at the top head strands droplets in slow zones, where they fall wet and start a deposit. The nozzle cannot fix an air problem, but the air cannot fix a droplet problem either.
Pilot data do not transfer linearly, and most reasons have nothing to do with the nozzle. The pilot tower loses more heat through its skin per kilogram of water. Residence time, air distribution, and feed delivery change with scale. The nozzle is the one component whose behaviour can be held constant, because droplet size is set by pressure, flow, and geometry, not by dryer size. That makes the nozzle the only reliable handle in the scale-up.
The customer spec closes the argument. Powder is sold on median particle size, fines fraction, and moisture, and each lot is tested against all three. The sizing therefore had to hold the median inside the band across a thirty fold flow step, which is a nozzle problem with a nozzle answer.
The Solution
The transfer logic runs in one order: hold the droplet, then the air, then the chamber. The droplet is held by keeping pressure and per-nozzle flow near the pilot values and scaling the number of nozzles instead of the orifice size. Scaling count rather than capacity is published practice on co-current towers, where several pressure nozzles share the top head at their design points.
The flow rule explains why pressure cannot span the gap alone. Flow through a pressure nozzle scales with the square root of pressure, so doubling pressure raises flow by about 41 percent. Pressure is a tuning dial inside a band, not a volume control across a decade. Push it too high and wear accelerates; drop it too low and the swirl weakens, the cone collapses, and the droplet spectrum widens. Published turndown guidance for swirl atomisation is narrow, commonly two to three to one, because quality depends on pressure rather than on a valve.
Worked example. The sizing arithmetic for the production step ran as follows.
| Quantity | Pilot rig | Production dryer |
|---|---|---|
| Water evaporation | 15 kg/h | 450 kg/h |
| Feed at 50 percent solids | 30 kg/h | 900 kg/h |
| Feed flow near 1,000 kg/m³ | 0.5 L/min | 15 L/min |
| Nozzles in service | 1 | 6 |
| Flow per nozzle | 0.5 L/min at 60 bar | 2.5 L/min at 150 bar |
The naive route was one nozzle at 15 L/min and 60 bar, which would move the median well outside the band under the published trends above. The sized route kept six nozzles, each rated for 2.5 L/min at 150 bar. At the pilot pressure of 60 bar, that unit carries about 1.6 L/min, roughly three times the pilot nozzle, and published trends put the coarsening near a 30 percent rise in the median. The pressure step from 60 to 150 bar refines the spray by roughly 20 to 30 percent at fixed flow. The two effects roughly cancel, so the design aim was a production median in the middle of the band, with the pilot product at the fine end so the drift has somewhere to go.
The count is a tower decision. Six nozzles fit this class of top head in published installation practice, with a spare position for growth, and each gets its own feed line where the manifold allows. A starved nozzle runs off its design point and changes its droplet size.
Where the target median sits below what a pressure nozzle holds reliably, or the feed is so heat sensitive that the drying air must stay cool, the published route is a two-fluid air atomizing nozzle. Compressed air shears the feed into droplets far finer than a pressure nozzle makes at the same feed rate, at the price of the most expensive utility on a drying line. That trade belongs to fine powders and small batches, covered with the air atomizing range.
The chamber check comes last, and it is a drawing exercise. Cone angle and nozzle position must keep the spray clear of the wall at the production fall height, per published guidance relating the clear distance to tower diameter. A cone that fits the pilot tower can still wet a production tower of different proportions, so the angle follows the production geometry. The same check applies when a dryer is re-nozzled rather than built new, the more common case.
The Engineering Behind the Choice
A pressure swirl nozzle is three parts in a line: tangential inlets, a swirl chamber, and the orifice. Feed enters tangentially and spins. The chamber turns pressure into rotation, and the rotating liquid exits the orifice as a thin conical sheet that breaks into droplets. The swirl makes the hollow cone, and it ties the median droplet to pressure, flow, and liquid properties such as viscosity and surface tension.
Geometry matters as much as pressure. The ratio of swirl chamber to orifice, the inlet size, and the internal finish all shift the droplet spectrum at fixed flow and pressure, which is why a swirl nozzle is sized as a system rather than drilled to a flow. A plain drilled hole does not atomise; it jets and dries as one lump. The droplet control in this case came from pairing the orifice class with the swirl chamber for the production flow, not from enlarging the pilot drawing.
The choice between pressure swirl and two-fluid atomisation is a four way trade: droplet size, turndown, air cost, and wear.
| Criterion | Pressure swirl (hydraulic) | Two fluid (air atomizing) |
|---|---|---|
| Typical published droplet range | Tens to just over one hundred microns at drying pressures | Roughly ten to sixty microns, air dependent |
| Turndown at constant droplet quality | Narrow, two to three to one | Wide, liquid and air can be trimmed together |
| Operating cost | Pump pressure only | Compressed air, the dominant utility on the line |
| Wear exposure | Orifice and swirl chamber, manageable with hard inserts | Fine liquid passages, plus air erosion where abrasive dust returns |
Materials follow the feed. 316L is the standard body for food ingredient and chemical duty. Abrasive feeds such as mineral slurries, brines, and ceramic slips erode soft stainless quickly, so the same nozzle family takes tungsten carbide or ceramic orifice inserts. Published hardness tables put tungsten carbide at roughly ten times the hardness of 316L: the body stays stainless for cleanability, and the insert takes the wear.
Wear shows up as drift before failure. A worn orifice at fixed pressure passes more flow, since the effective opening is larger, and the larger opening coarsens the spray. The median climbs, the moisture follows, and the lot drifts toward the edge of the band. Practical published practice treats flow at fixed pressure as the wear meter: log it, and pull the nozzle when flow drifts a few percent above baseline.
Between campaigns the dryer cone and ducting need a washdown, and that duty uses its own nozzles. Full cone spray nozzles cover the cone interior evenly for rinsing, sized separately from the atomisers: a cleaning nozzle is chosen for coverage, an atomiser for droplet size.
The Results
No site audit numbers appear in this table. The figures come from published reference ranges for spray drying and for the nozzle classes discussed here. They describe what the equipment delivers when the sizing logic above is followed, and they are ranges, not promises for any single plant.
| Metric | Typical published result | Why |
|---|---|---|
| Median particle size control | Inside published tolerance bands for pressure swirl nozzles when pressure and per-nozzle flow hold near the design point | The median is set at atomisation, so holding pressure and flow holds the median |
| Fines fraction | Reduced by a two-fluid path where the spec demands a tighter droplet distribution | Air atomizing gives a narrower droplet spread, so at a matched median the fine tail is smaller |
| Energy per kilogram of water evaporated | 4.5 to 6.5 MJ/kg for spray drying | Latent heat of water dominates, so the atomiser changes the powder, not the thermodynamics |
| Throughput flexibility | Roughly 20 to 30 percent above design by raising pressure inside the operating band | Flow scales with the square root of pressure, so a modest pressure rise buys flow |
| Campaign stability | Held by flow logging at fixed pressure | A worn orifice raises flow and coarsens the spray, so flow drift flags the nozzle before the lot drifts |
These are published ranges, not measured site results; no single plant was audited for this account. The honest use of the ranges is as the acceptance band for a spray test at the production flow before the tower is committed. The pilot product at the fine end of the band and a tested production median in the middle are the two anchors of the transfer.
Why This Case Matters
The logic transfers because the physics does not change with the dryer. Fix the droplet first, then the air, then the chamber. That order collapses most of the catalogue before flow is discussed, and it is the same order the spray drying nozzle guide walks through for any drying duty.
The pilot has a real job in this order. It is not there to prove the exact production powder. It fixes the droplet class, the solids, and the air temperature, and it should run toward the fine end of the spec so production drift lands in the middle of the band. Where the production median must match the pilot exactly, the answer is a two-fluid path or a per-nozzle flow close to the pilot value; the droplet size guide sets out that trade.
Scale-up arithmetic is cheap before steel is cut. The square root flow rule, the capacity exponent, and the cone trigonometry above take minutes on a calculator and rule out the dead ends. The remaining choices are nozzle count, pressure class, and the two-fluid alternative. The atomization nozzle comparison and the particle size guide for pressure spray drying cover the neighbouring routes, including the spinning disc atomiser for very large towers and coarser powders.
What to send for a sizing, in one list:
- Feed solids and viscosity at the atomisation temperature, because both set the droplet size at a given pressure.
- Target median particle size, allowed fines fraction, and moisture limit, because they define the band.
- Evaporation rate at pilot and production, because the ratio sets the nozzle count.
- Available feed pressure and pump curve, plus compressed air capacity if a two-fluid path is under review.
- Tower diameter and top head layout if the dryer exists, because the cone must clear the wall.
- Whether the feed is abrasive, because that decides the orifice material.
Each item changes the answer, so a complete list returns a sized configuration in one round. Send it through the contact page, and BoreJet comes back with the nozzle class, count, pressure, and swirl geometry, sized to your numbers rather than guessed from a catalogue. The atomising range for this duty lives on the air atomizing page, and the dryer cone washdown on the full cone page.
Your Duty May Differ
Send the Vessel Drawing and the Pump Curve.
We size the head class, the count and the operating point for your tank, not a catalogue guess. A coverage test on the actual vessel beats any estimate.