How to Design a Work Roll Cooling Header: Flat Fan Nozzles, Flow Density and Spray Angle

A work roll cooling header is a quality control device wearing a water pipe’s clothes. Every revolution, the surface is heated in the roll bite then passed under a spray header that must strip that heat back out, completely, evenly, at line speed. When the sprays are right, the roll runs cool, the thermal crown holds its shape, and strip profile and flatness stay on tolerance. When they are wrong, the roll fire-cracks and wears unevenly, and every defect transfers to the steel passing through the gap.

This guide covers the engineering behind a roll cooling nozzle spec: flow density in L/min per metre of barrel, spray angle and offset, overlap layout, and water quality and maintenance, with flow, pressure and angle ranges from vendor roll cooling literature and published mill studies.

Why the Roll Cooling Nozzle Is a Quality Control Device

Heat enters a work roll by four paths: conduction from the hot strip through the scale layer in the bite, radiation from the strip, friction in the arc of contact, and bearing and back-up roll friction. The first two dominate in hot rolling; in cold rolling, friction work in the bite leads. In a finishing stand the roll surface faces a strip at 900–1,100 °C, and model studies show peak surface temperatures around 500–550 °C immediately after the bite exit.

That cycle, heat up in the bite, quench under the header, repeats thousands of times per hour and is the root cause of three failure modes:

• Thermal fatigue. The outer layer expands and contracts with each revolution; model work on a hot strip mill finishing stand showed that adding exit-side spray rows cut predicted plastic strain per cycle, while removing a row raised both peak surface temperature and strain per cycle. Roll cooling decides how many cycles the surface survives before fire cracking starts.
• Hardness loss and abrasion. Above roughly 200 °C, the surface hardness of conventional roll materials falls steeply, and a soft surface wears fast against hot strip and back-up roll. Keeping the surface below that threshold is the difference between a campaign measured in thousands of tons and one in tens of thousands.
• Thermal crown distortion. The roll body expands most where it is hottest, the strip-width zone, so a poorly cooled roll grows a crown in the center of the barrel, changing the roll gap profile and showing up as strip profile and flatness deviation. The sprays are the main fast actuator for thermal crown control, which is why modern mills divide the barrel into valve-controlled zones.

None of this is optional: roll gap profile is the sum of ground crown, bending and thermal crown, and the thermal term is the one the operator can move during a campaign.

Why Flat Fans Beat Every Other Pattern on a Roll Barrel

Two spray families compete for roll cooling duty: flat fan (flat jet) and full cone. Full cones were common in older mills; flat fans now dominate new headers, for physical reasons.

A flat fan produces a thin, elongated sheet of liquid, an elliptical footprint on the roll surface, with a parabolic liquid distribution across its width. Arranged along a header, those parabolas sum into a very even combined distribution, exactly what a barrel needs. A full cone produces a round footprint with a light center unless it carries an internal vane insert, and covering a barrel with circles leaves either gaps or heavy overlap bands.

Three properties make flat fans the standard answer:

• Impact through the water film. Flat jet nozzles produce relatively large droplets and high impact per unit area. In a rolling mill the roll surface runs under a film of coolant, and the spray must penetrate it to make solid–liquid contact. Flat fans impinge through the water layer a fine full cone struggles to break.
• No internal inserts. The sheet is formed by the orifice itself; there are no vane cores to clog, so clogging risk on dirty mill water is structurally lower than for a full cone with an x-vane insert.
• Self-aligning mounting. Because the footprint is directional, orientation matters. Every major vendor sells self-aligning flat fan tips, from dovetail connections with a built-in offset angle to tips with locating keys machined into a box-type header plate. Alignment is designed into the hardware, not left to a fitter’s eye.

The one place narrow flat fans do a different job is descaling: the same geometry at very different duty: 150–250 bar, narrow 15–25° angles to concentrate impact, hardened orifice inserts for abrasive scale (table below). Mills do not interchange the two nozzles: the same tip cannot descale and cool well.

Duty	Typical pressure	Typical angle	Pattern priority
Descaling	150–250 bar	15–25°	Maximum impact, tight zone
Work roll cooling (hot mill)	8–15 bar header	40–65°	Even heat removal
Work roll cooling (cold mill)	3–8 bar header	40–65°	Even heat removal + lubricant film
Wide barrel / plate mill	8–15 bar	80–110°	Fewer rows, full-width coverage

Flow Density: The Number That Decides Cooling Power

The single most useful number in a roll cooling design is flow density: coolant flow per metre of roll barrel, in L/min·m. It is the practical proxy for the local heat transfer coefficient, which rises roughly as water flux to the power 0.65–0.75 (Mitsutsuka’s classic spray cooling relationship). Doubling the flow density therefore does not double the cooling. It raises the heat transfer coefficient by about 60% (2^0.7 ≈ 1.62). Beyond a certain flux, extra water buys little extra heat removal; the design problem becomes distribution and coverage rather than raw volume.

What do real mills run? Published operating figures for hot rolling work roll cooling sit in the range of 2,500–4,000 L/min total per roll, at header pressures of 0.8–1.5 MPa (8–15 bar) and water temperatures of 35–45 °C, with 95–100% of the barrel covered. Cold rolling mills run similar or higher totals, 3,000–6,000 L/min, at lower pressure, 0.3–0.8 MPa (3–8 bar), because the coolant is an emulsion at 2–5% concentration that must lay a lubricating film rather than blast scale. On a 1.8 m barrel, 3,000 L/min is roughly 1,700 L/min·m, a useful anchor for a wide finishing stand. Selective cooling zones commonly carry 150–400 L/min each; a widely quoted benchmark for a single flat spray test row is 40 L/min per nozzle at 5 bar, 60° spray angle.

Pressure converts to flow through the standard nozzle law: flow scales with the square root of pressure (Q ∝ √P). Raise a header from 5 to 8 bar and every nozzle delivers about 26% more flow (√(8/5) ≈ 1.26). Mills throttle cooling by pressure, valve-controlled zones, or nozzle pulsing. The caution: add rows or orifice capacity rather than cranking pressure alone. Very high spray pressures make water rebound off the roll, and the heat transfer gain collapses. Moderate pressure, high coverage is the design rule.

Water’s specific heat is 4.18 kJ/kg·K, so every litre that rises 10 °C carries away about 41.8 kJ; a header delivering 720 L/min with a 10 °C rise removes roughly 500 kW by sensible heating alone (12 kg/s × 4.18 × 10), and flash evaporation adds latent heat at 2.26 MJ/kg on top. That is why roll cooling flows run in hundreds of litres per minute: finishing stand work rolls absorb on the order of a megawatt each.

The Physics Under the Spray: Bite Heat Transfer, Surface Stress and Impact Pressure

Flow density is a proxy for cooling power; what the roll actually feels is heat flux and cyclic stress. Here is how those mechanisms work.

Heat transfer in the roll bite. In the arc of contact, heat flux into the roll is the product of a heat transfer coefficient and the strip-to-roll temperature difference (q = h·ΔT). Conduction through the scale layer dominates in hot rolling, so bite heat flux is set as much by scale thickness and strip temperature as by the header, which is why the first cooling rows sit as close to the bite exit as space allows. The coefficient is not a fixed material property: published work from the Heat Transfer and Fluid Flow Laboratory (heatlab.cz) shows it depends on impingement density, droplet size and velocity, and, critically, on roll surface temperature. Above the Leidenfrost temperature a stable steam film blankets the surface and heat transfer collapses: roughly 400 W/m²K in film boiling versus about 4,000 W/m²K in nucleate boiling below it. A point at 900 °C next to a point at 700 °C is therefore cooled about ten times more intensively, small temperature differences amplify into hot and cold bands. The same data explain why “more water is always better” fails: increasing water quantity can decrease cooling intensity as excess water rebounds instead of clinging and evaporating.

Because HTC cannot be predicted reliably for a new nozzle configuration, it is measured: heatlab’s rotating test benches, experimental rolls of 650 mm and 350 mm diameter, the larger fed by a 150 kW pump station, instrument the roll with subsurface thermocouples and solve the inverse heat conduction problem (IHCP) to reconstruct surface temperature and HTC histories around the circumference. Recent published tests of flat fan roll cooling on such a rig measured average HTCs of roughly 9,000–11,000 W/m²K in the nucleate boiling range at 0.3–0.8 MPa, an order of magnitude above film boiling, and a reminder that local flux must keep every point on the cool side of that cliff.

Thermal and mechanical surface stress. The bite-to-header cycle is a stress cycle. In the roll gap the surface layer is driven into compression by contact pressure and sudden heat input; in the cooling zone, contraction pulls it into tension. The cyclic load runs the surface layer around a stress–strain loop whose enclosed area is proportional to the deformation energy absorbed per revolution, the energy that nucleates fire cracks, with growth predominantly circumferential. These cyclic thermal stresses add to mechanical stresses from rolling force and bending; combined with oxidation and abrasion, they wear the surface out. Heatlab’s simulation software (SimRoll) tracks the resulting temperature field and thermal crown over a simulated campaign, which is why instrumented tests and simulation, not datasheet numbers, set the final header spec.

Impact pressure measurement. Impact pressure, the force per unit area a spray exerts on the surface, is the measured link between nozzle selection and film penetration. It is not the same as header pressure: it depends on standoff, droplet size and velocity, and the water layer the spray must cross. Benches measure impact pressure distribution across the footprint with a traversing sensor; heatlab’s bench records distributions through water layers of 5, 20 and 40 mm at a 200 mm standoff, showing how film depth reshapes what reaches the roll. The practical lesson is direct: a fan that looks right on paper can lose most of its impact crossing the coolant film, so spec impact pressure, and verify it on a bench, alongside flow density and angle. Flat fans are favored partly because their high impact per unit area survives film crossing better than fine full cones.

Spray Angle: Match the Fan to the Barrel and the Gap

The spray angle sets footprint width at a given standoff. For a flat fan the geometry is simple:

Footprint width = 2 × standoff distance × tan(θ/2)

At a 150 mm standoff, a 40° fan covers about 109 mm, a 60° fan about 173 mm, a 90° fan 300 mm, and a 110° fan about 429 mm. Halve the standoff and you halve the footprint, and roughly double the impact flux, as the same flow lands on a quarter of the area.

Practical selection rules for roll cooling headers (details in the spray angle guide):

• 15–25°: narrow, high-impact fans for descaling and spot cooling.
• 40–65°: the standard range for work roll cooling. A 60° fan at 100–150 mm standoff gives a 115–170 mm footprint, suiting a 75–120 mm pitch.
• 80–110°: wide fans for plate mill rolls or where space limits rows: fewer nozzles per row, lower local impact, coverage against flux.

Offset Angle and Header Rows: Where the Water Actually Hits

The spray angle sets footprint width; the offset angle, the spray axis relative to the header/roll axis, sets where on the roll circumference the water lands. It is measured, not stylistic: Lechler’s published roll cooling studies found the optimum at roughly 60° to the horizontal, and steepening toward 75° measurably reduced cooling effect, as spray height grows at the footprint’s outer edges, impact falls, and heat transfer suffers. Perpendicular impingement gives the most symmetrical footprint and the highest heat transfer coefficient, why top and bottom headers are arranged symmetrically, and why misalignment shows up as drive-side/operator-side cooling differences.

Headers are built as rows; a common layout is two: the row aimed at the roll gap carries basic cooling and lubrication, roughly 20–30% of the header flow, while the second carries 70–80% and is zoned for selective cooling. Mills run two to five rows depending on stand position and space. Row count is a fatigue-life decision, not a plumbing preference.

Two mechanical details make this repeatable in the mill:

• Built-in offset angles. Dovetail flat fan tips commonly carry an automatic offset, 5° on small series, 15° on the industrial standard series, so a nipple welded square to the header automatically produces the correct spray offset.
• Alignment verification. Alignment tips screw into the header in place of nozzles for fabrication, pressure testing, and blocking off positions, how mills mask roll edges or shut down zones without removing hardware.

Selective cooling takes the row concept further: in cold rolling especially, the barrel is divided into zones of 25–100 mm width, each fed by a valve-controlled flat fan nozzle, switched individually, by PLC, push button, or automatically from a shape metering roll, to correct asymmetric shape defects. Zone control is either level control (discrete flow levels per zone, e.g. 5, 10, 15 L/min) or pulse control (each nozzle on/off at a duty cycle). Same nozzle hardware; the difference is the valve strategy.

Overlap: Even Coverage Without Hot or Cool Bands

Uniform cooling is the entire point, and uniformity is decided at the overlap between adjacent fans. Two philosophies converge on the same goal:

• Parabolic summation. Because a flat fan’s distribution is parabolic, adjacent fans can be pitched so the parabolas add to a flat combined profile, the approach Lechler documents, arranging footprints “without interference” to avoid both gaps and double-strength bands.
• Overlapping patterns. Spraying Systems’ VeeJet practice deliberately overlaps adjacent fans, the pattern that produces uniform cooling across the barrel. Their Virtual Overlap Analysis (SprayScan VOA) software exists precisely because overlap geometry is where headers go wrong.

The two descriptions differ in wording but agree on the physics: the combined liquid distribution across the header must be flat. A valley leaves the roll hot; a peak runs it cold and the strip profile shows it. First-pass rule: pitch adjacent fans so each footprint overlaps its neighbor by roughly 20–30% of the footprint width, then verify the layout on paper or in a test lab before the header is welded. Edge handling matters as much as the body: where the strip does not cover the roll, cooling is rolled off deliberately. Edge zones are masked or de-rated so they do not over-cool the center. A cold-running edge shrinks, the roll profile goes concave, and strip edges lose contact.

Water Quality, Filtration and Clogging

Roll cooling water is not clean water: mill scale, oxides, descaling debris and dissolved hardness are all nozzle threats. A plugged nozzle does not fail silently. It prints a hot band on the strip, a coil rejection. The defense is layered:

• Free passage by design. Flat jet nozzles have no internal vanes, so clogging resistance is inherently better than insert-type nozzles. Specify unobstructed flow passages.
• Integral strainers. Many roll cooling flat fan tips come with built-in strainers (Spraying Systems’ H-VVL is the classic example), sized to catch particles before the orifice; on mill duty they are standard, not optional.
• Filtration upstream. Strainer mesh should be sized below the smallest nozzle orifice in the header. If the mill water system cannot hold that grade, the header needs its own strainer or filter basket.
• Materials. Hot water, scale and mill scale slurry erode soft materials. Abrasion-resistant stainless steel bodies and hardened orifice inserts are standard; brass and soft alloy tips are for clean-water service only.

Watch pressure, not appearance: rising header pressure at constant flow means strainers are loading; falling pressure at constant pump speed, or rising flow at constant pressure, means an orifice is eroding or a tip has blown out. Either way the mill pays in strip quality long before visual inspection finds it.

Maintenance: What Kills Roll Cooling Performance

Most roll cooling degradation is gradual, and most of it is not the nozzle’s fault:

Symptom	Root cause	Fix
Hot bands across the strip, repeatable position	Plugged or missing nozzles in one header position	Clean or replace tip; check strainer and upstream filter
Strip edges overcooled	Edge zones over-covered; wrong offset at barrel ends	Mask edge positions with alignment tips; de-rate edge zones
Flow rising at same header pressure	Orifice erosion from scale abrasion	Flow-check the header; replace tips on schedule
Drive side vs operator side cooling difference	Header misalignment or wrong nozzle orientation	Alignment check per vendor procedure (regular, not just at install)
Shorter roll campaigns, fire cracking	Insufficient rows or flux for the stand duty	Add rows, raise flow density, verify peak surface temperature in the model
Header pressure creeping up	Strainers loading with scale	Clean strainers; audit filtration grade

Orifice erosion deserves the sharpest eye. Flow through a flat fan slot scales with the square of its dimension. A 10% enlargement raises flow by roughly 21%, and with it the local cooling. Flow checks beat visual checks: measure flow and pressure against the design curve and replace tips when flow drifts more than 10–15% from spec, on schedule, not after the rejects start.

Alignment is the other recurring killer. Self-aligning tips remove most human error at install, but headers get knocked, replaced and rebuilt. A header rotated a few degrees changes every footprint on the roll. Vendor practice is explicit: check alignment regularly, not only at commissioning. The visible evidence, asymmetric cooling, strip steering problems in tandem mills, roll profile deviation, shows up long after the geometry drifted.

Spec Sheet: What to Send a Nozzle Supplier

A roll cooling nozzle recommendation is only as good as the duty data behind it. Send the supplier: barrel width and roll diameter, strip width, header pitch and standoff, available pressure and water temperature, target flow density in L/min·m, water quality, and line speed. With those inputs a supplier can specify spray angle, orifice size and material, and tell you when the real problem is row count, overlap or filtration rather than the nozzle itself.

Sizing a Header: Worked Example

A wide hot strip mill finishing stand work roll has a 1,800 mm barrel; the strip is 1,600 mm wide. The mill wants a flow density of 400 L/min·m, a figure at the upper end of the hot rolling band.

• Total header flow: 1.8 m × 400 L/min·m = 720 L/min.
• Nozzle count: at a 100 mm pitch, 18 nozzles per row → 40 L/min per nozzle, the 5 bar benchmark tip.
• Angle and standoff: 60° fan at 120 mm standoff gives a 139 mm footprint; pitched at 100 mm, overlap ≈ 28%, inside the 20–30% band.
• Rows and split: bite row 25% (180 L/min, ~10 L/min per nozzle); selective row 75% (540 L/min, ~30 L/min per nozzle), zoned into 10 zones of two nozzles (~60 L/min each).
• Pressure check: with only 8 bar at the header, the 40 L/min @ 5 bar nozzle delivers ~50 L/min, oversize; specify a 32 L/min @ 5 bar tip instead.

Every number is a starting point, not a guarantee. Heat load, water temperature and roll material move the target. But the chain is the point: flow density sets header flow, pitch sets nozzle count, angle and standoff set footprint and overlap, and the pressure law converts between supply conditions.

FAQ

Can the same flat fan nozzle be used for descaling and roll cooling? Generally no. Descaling needs a narrow-angle, high-impact pattern at 150 bar or more; roll cooling needs broader, even coverage at moderate pressure. Mills spec different tips for each duty.

What pressure should a work roll cooling header run at? Hot strip mill roll cooling headers commonly run 8–15 bar; cold rolling headers 3–8 bar on emulsion; fine-spray work roll cooling is quoted from 10–40 bar.

How do I know when to replace roll cooling nozzles? By flow, not by looks. Measure header flow and pressure against the design curve and replace tips when flow drifts more than 10–15% at the same pressure, or when a flow check shows uneven distribution.

Do roll cooling nozzles need strainers? On mill water, yes. Scale and debris are always present, and a plugged tip prints a hot band on the strip. Integral or header-mounted strainers with mesh below the smallest orifice are standard.

Flat fan or full cone for roll cooling? Flat fans are the preferred pattern for roll cooling: the elongated footprint matches the barrel, the parabolic distribution sums evenly, the large drops and high impact penetrate the water film, and the insert-free geometry resists clogging on mill water.

Roll Cooling Header Checklist

• Flow density set in L/min·m, matched to stand heat load, not copied from another mill
• Header pressure and water temperature confirmed at the stand, not at the pump
• Spray angle chosen for standoff and pitch; footprint and overlap calculated
• Adjacent fans overlapped 20–30% of footprint width; combined distribution verified flat
• Offset angle set near 60° to horizontal; nozzles perpendicular to the roll where possible
• Self-aligning tips (dovetail or keyed) with correct built-in offset for the header
• Rows sized: bite row ~20–30% of flow, selective row ~70–80%, zoned for crown control
• Strainers below the smallest orifice; materials rated for scale abrasion
• Flow-check schedule defined (10–15% drift = replace), alignment checked on a calendar

Summary

Roll cooling is a heat-transfer problem with a geometry answer. The flat fan is the right roll cooling nozzle because its footprint, distribution and impact match what a roll barrel needs, and the header is right when flow density, spray angle, offset, overlap and water quality line up. Start from the heat load, express it as L/min per metre of barrel, lay the fans out so the combined distribution is flat, and verify with flow checks. The same mill, upstream of the mill stand, fights a different battle: descaling headers on the run-out table shows how header geometry decides whether the scale comes off or the surface does. For the duty overview and the range that covers it, see the roll cooling application page, the flat fan nozzle product page, the cooling nozzle selection guide, or contact us with your barrel width and header drawing.