Case Study: Conveyor Belt Spray Wash and Carry-Back Retrofit
A transfer point dropped manual hose-down for a recirculating belt spray wash. Fresh water fell from 6.75 to 0.94 cubic metres a day.
Carry-back is the fines that ride a belt back to the tail and drop where nobody planned. One aggregate transfer point ran a single carbide scraper and a hand hose. A 10 mm layer under a 1,200 mm belt is enough to shift tracking and start edge damage.
Published belt maintenance guidance lists carry-back and pulley build-up among the main causes of mistracking, edge damage, and splice fatigue. On this duty, conveyor belt cleaning is a mass balance problem before it is a nozzle problem. This case covers a belt conveyor spray wash retrofit at one transfer point in North America.
Case Snapshot
The figures below are the generic shape of this class of conveyor spray wash retrofit, not a record from one plant.
| Item | Value |
|---|---|
| Asset washed | 1,200 mm fabric belt on a 14 m transfer conveyor |
| Throughput | about 200 t/h of crushed aggregate and recycle fines |
| Belt speed | 1.5 m/s |
| Before | One primary carbide scraper plus manual hose-down |
| After | Four wash bars, three nozzles per bar, twelve nozzles |
| Nozzle | 110 degree flat fan, 2.5 mm equivalent orifice |
| Wash point | 4.0 bar at the nozzle, 4.4 bar at the pump |
| Water | Sump recirculation, 80 percent recovery target |
A 1,200 mm belt at 1.5 m/s moves 1.8 square metres of surface per second. That rate and the 0.03 kg each square metre sheds per pass drive every number here.
The Challenge
The transfer point dropped crushed aggregate from a 900 mm chute onto a 1,200 mm fabric belt. The belt ran 14 m to the head pulley at 1.5 m/s, and throughput held near 200 t/h across two shifts.
Control was one primary cleaner with a carbide blade, set behind the head pulley. The blade had been backed off twice to protect a repaired splice. Lower blade pressure let a wet film pass under the tip on every revolution.
That film became build-up. Fines dropped at the tail pulley, the snub pulley ahead of it, and along the leading 3 m of the return strand. Housekeeping ran to a hose and a shovel three times a day.
The wash was that same hand hose: 15 minutes per shift for three shifts at roughly 150 L/min. That is 6.75 cubic metres of mains water a day, once-through, leaving the tower as a slurry.
Respirable dust was the other pressure. It had to stay inside occupational exposure limits during live loading, so conveyor dust control was already a live subject on site.
Carry-Back Material and Where It Bonds
Belt carry-back is not one material. Three families behave differently under a wash, and the mix decides what the water has to do.
| Carry-back family | Where it bonds | What lifts it |
|---|---|---|
| Dry dust under 1 mm | Loose on the cover, upper return strand | Air knife, or no water at all |
| Damp fines under 5 mm | Cover texture behind the load zone | Pre-wet, then a light blade |
| Clay-bearing recycle fines | Cover texture plus a dried shell | Wet wash at full bar pressure |
Adhesion comes from moisture and pressure. At the load zone the falling stream presses fines into the cover. At the head pulley, belt tension presses them in for the whole wrap.
Dry dust releases on its own, so a conveyor air knife suits a dry, non-clay duty. Damp fines need a pre-wet before a blade can lift them. Clay-bearing recycle fines pack into the cover texture and dry into a shell a scraper skates over.
The mass balance comes before nozzle choice. At 200 t/h and 0.1 percent carry-back the belt returns 0.2 t/h, which is 0.056 kg/s. Published carry-back figures for bulk belts run 0.05 to 0.5 percent of throughput, so 0.1 percent is mid-band.
Divide by the surface rate. At 1.2 m wide and 1.5 m/s the belt presents 1.8 square metres per second. The load is 0.056 / 1.8 = 0.031 kg per square metre per pass.
Buildup at Pulleys and Load-Bearing Idlers
Carry-back does not stay spread. It collects wherever geometry traps it, and the traps repeat on almost every bulk conveyor.
| Build-up site | Why it collects there | Damage it causes |
|---|---|---|
| Tail pulley face | Belt wraps under tension, fines are ironed on | Local belt lift, mistracking, edge wear |
| Snub pulley | Low turning load, no self-cleaning | Bump under the belt each pass |
| Return idlers in the leading 3 m | Fines fall straight after discharge | Out of round shell, splice hammering |
A rubber-lagged tail pulley here runs near 400 mm diameter. A 6 mm layer of compacted fines lifts the belt locally by 6 mm, and the stiffer belt rides high and dips either side.
Idlers matter differently. A 5 mm shell build-up on a 100 mm idler turns rolling contact into a hammer blow. A 14 m conveyor carries about 30 m of belt plus wrap, so the splice passes a given idler every 20 seconds at 1.5 m/s. That is about 180 bumps an hour.
Scraper Wear Versus a Wet Wash
A scraper is a wear part working against a wear part. Extra blade pressure buys carry-back removal and spends belt cover.
Published field ranges for primary cleaner blades in abrasive aggregate service put blade life at roughly 4 to 12 weeks. Backing the blade off buys blade life and spends everything else.
A tensioned scraper also pulls against the belt, so the drive carries the friction. The real cost lands on the splice and the cover.
Wetting the fines lowers adhesion, so the blade works against a loosened layer instead of a bonded one. The water strips, and the blade wipes.
Droplet Momentum and Penetration
A fan nozzle fires a coherent liquid sheet, not a mist. The published rule of thumb for a flat fan holds the sheet for roughly 50 to 100 orifice diameters. A 2.5 mm orifice holds its sheet from about 125 mm out to 250 mm, so the 180 mm standoff sits inside the coherent zone.
Exit velocity is set by pressure alone. For water at 4.0 bar, 2 x 400,000 / 1,000 = 800, and the square root of 800 is 28.3 m/s.
Droplet size decides whether that momentum survives. Published data for hydraulic flat fans at 4 bar puts the volume median diameter near 150 to 300 micron for this orifice class.
Take a 200 micron droplet. Its Stokes relaxation time is 1,000 x (0.0002)^2 / (18 x 0.000018) = 0.123 s. Transit over 180 mm at 28.3 m/s takes 0.18 / 28.3 = 0.0064 s, so it lands at near 95 percent of exit velocity. Higher drag makes 80 to 95 percent the working band instead.
That repeat coverage is what separates belt conveyor spray washers from a hand hose.
The Solution
The retrofit put four wash bars on the conveyor and a recirculating sump under it. Each bar carries three spray bar nozzles at a 0.45 m pitch. The nozzle is a hardened stainless body with a tungsten carbide insert on a 2.5 mm equivalent orifice.
| Bar | Position | Duty | Nozzles |
|---|---|---|---|
| A | Return strand, 1.2 m before the tail pulley | Pre-wet to soften the film | 3 |
| B | Return strand, 0.4 m before the tail pulley | Full pressure strip of carry-back | 3 |
| C | After the discharge chute, carrying strand | Stop fines keying into the cover | 3 |
| D | Head pulley approach | Rinse the belt before the scraper | 3 |
The order matters more than the count. Bar A wets the film so bar B can strip it. Bar C stops new fines from bonding, and bar D hands the scraper a wet surface. That order is why a conveyor belt cleaning system beats a scraper alone.
Header sizing followed from total flow. Twelve nozzles at 8.125 L/min give 97.5 L/min, or 1.625 litres per second. Holding velocity under 3 m/s needs 1.625 / 3 = 541 square mm, a bore near 26 mm. A DN32 header then runs the span at 1.7 m/s.
Wetted parts are 316 stainless, with tungsten carbide inserts at a published 4,000 to 12,000 hours in this class. Ceramic lasts longer on straight abrasion, often quoted at 8,000 to 20,000 hours, but it chips on impact.
Branch nozzles are 3/8 inch and header connections are 1 inch. The rules that match BSP to NPT sit in the nozzle thread size guide.
Spray Bar Geometry and Coverage Math
One piece of trigonometry sets every bar dimension. A fan of total angle theta at standoff d paints a band of width w, where w = 2 x d x tan(theta / 2).
Work it for the chosen nozzle. The angle is 110 degrees, so half the angle is 55 and its tangent is 1.428. At a 180 mm standoff, w = 2 x 0.180 x 1.428 = 0.514 m. One flat fan spray nozzle covers a 514 mm band.
Overlap sets the pitch. Published practice for a uniform band is 10 to 15 percent. Take 12 percent, so the pitch is 0.514 x 0.88 = 0.452 m, rounded to 0.45 m.
Check the coverage. Three nozzles at a 0.45 m pitch span 0.90 m between outer centres. One half width at each end gives a wetted band of 0.90 + 0.514 = 1.414 m. The 1,200 mm belt is covered, with about 107 mm of overhang.
A narrower angle needs a closer bar. A 65 degree fan at 180 mm paints only 2 x 0.180 x tan(32.5) = 0.229 m. Covering 1.2 m would need seven of them. The spray angle guide tabulates that trade.
| Geometry step | Working figure | How it was set |
|---|---|---|
| Total fan angle | 110 degrees | Chosen for a short standoff band |
| Standoff | 180 mm | Inside the coherent sheet zone |
| Band width | 0.514 m | 2 x 0.180 x tan(55 degrees) |
| Overlap | 12 percent | Middle of the 10 to 15 percent band |
| Pitch | 0.45 m | 0.514 x 0.88, rounded |
| Nozzles per bar | 3 | 1.2 m belt at a 0.45 m pitch |
| Wetted width | 1.414 m | 0.90 m span plus one full band |
Nozzle Pattern Choice for Abrasive Slurry
Recirculated sump water decides the pattern, because any solid that passes the screen has to pass the nozzle. A spiral nozzle has the largest free passage of the common families and shrugs off coarse grit. That is why bar A can run on sump water while bar B takes the cleanest stream available.
A full cone fills its circle evenly, which suits wetting rather than stripping. On a belt the useful shape is a thin band, so a flat fan nozzle wins wherever the water is clean enough. Under a chute lip a full cone nozzle spreads the same flow over a round footprint.
The flat fan versus full cone comparison shows why edge definition matters on a moving surface. Hollow cone nozzles are the wrong family on this band duty.
Engineering Notes and Worked Arithmetic
Five pieces of arithmetic carried the design. All five check with a tape measure, a gauge, and a flow meter.
Orifice flow at pressure. For water, Q (L/min) is about 0.65 x d (mm) squared x the square root of P (bar). Solve for the required 8.13 L/min at 4.0 bar. The square root of 4.0 is 2.0, so 0.65 x d squared x 2.0 = 8.13. Then d squared = 8.13 / 1.3 = 6.25, giving d = 2.5 mm.
Check against ideal exit velocity. Ideal flow is area times velocity. Area is 0.7854 x 2.5 squared = 4.909 square mm, or 4.909e-6 square metres. At 28.3 m/s that gives 4.909e-6 x 28.3 = 1.388e-4 cubic metres per second, or 8.33 L/min, against 8.13 L/min from the 0.65 coefficient.
Flow at a changed pressure. Flow through a fixed orifice follows the square root of pressure. Taking the same nozzles from 4.0 to 5.0 bar gives 97.5 x sqrt(5.0 / 4.0) = 97.5 x 1.118 = 109.0 L/min. That is 12 percent more water for a 25 percent pressure rise. The flow rate calculation guide covers the relation in gpm and psi.
Fan coverage width. w = 2 x 0.180 x tan(55) = 0.514 m, with 12 percent overlap giving a 0.45 m pitch, as worked above.
Impact against standoff. Inside its coherent sheet, a fan jet lands at an impact pressure falling roughly with the reciprocal of standoff. The ceiling is line pressure, because no jet exceeds its own stagnation pressure. Published readings for industrial water spray nozzles at 4 bar land between 0.5 and 1.0 bar at 1 m. A 180 mm standoff is far inside that.
| Quantity | Substitution | Result |
|---|---|---|
| Nozzle flow | 0.65 x 6.25 x 2.0 | 8.13 L/min |
| Ideal flow | 4.909e-6 x 28.3 x 60,000 | 8.33 L/min |
| Ideal exit velocity | sqrt(2 x 400,000 / 1,000) | 28.3 m/s |
| System flow | 8.125 x 12 | 97.5 L/min |
| System flow in gpm | 97.5 / 3.785 | 25.76 gpm |
| System flow in m3/h | 97.5 / 16.67 | 5.85 m3/h |
| Pump pressure in psi | 4.4 x 14.5 | 63.8 psi |
Pump Duty and Water Volume Limits
Pressure costs power faster than it costs water. Pump hydraulic power is kW = Q (L/min) x P (bar) / 612.
At the pump the flow is 97.5 L/min and the pressure is 4.4 bar, because 0.4 bar is spent in the header and valves. Hydraulic power is 97.5 x 4.4 / 612 = 429 / 612 = 0.70 kW. At 0.65 pump efficiency the shaft load is 0.70 / 0.65 = 1.08 kW, so a 1.5 kW motor covers the duty.
Now test the tempting move. Push nozzle pressure from 4.0 to 5.0 bar and flow rises to 109.0 L/min. Hydraulic power becomes 109.0 x 5.4 / 612 = 0.96 kW against 0.70 kW. The P to the power 1.5 rule agrees.
The tower had a 12 mm feed line and a plant already running two process hoses, so 109 L/min of fresh draw was never available. The system ran 97.5 L/min from a sump at 4.0 bar instead.
| Item | Working figure | Check |
|---|---|---|
| Nozzle pressure | 4.0 bar | 58 psi at 14.5 psi per bar |
| Pump discharge | 4.4 bar | 0.4 bar header loss allowed |
| System flow at the pump | 97.5 L/min | 12 x 8.13 |
| Header velocity | 1.7 m/s | DN32, 35 mm bore |
| Branch velocity | 1.3 m/s | DN20, 24.4 L/min |
| Hydraulic power | 0.70 kW | 97.5 x 4.4 / 612 |
| Motor selected | 1.5 kW | Margin for a dirty filter |
Water Recovery and Effluent per Shift
The sump is what turns spray bars into a conveyor belt cleaning system a permit office will accept.
Gross spray per day follows from the duty cycle. Twelve nozzles at 97.5 L/min for a 60 second window use 97.5 L. At 48 windows a day that is 48 x 97.5 = 4,680 L, or 4.68 cubic metres.
Recovery is a target, not a measurement. For a sump with a settling zone and a coarse screen, the working figure is 80 percent. That returns 4.68 x 0.80 = 3.74 cubic metres a day.
The rest is the loss line. Carry-off on the wet belt takes about 0.42 cubic metres a day, evaporation and drift take 0.24, and sump purge takes 0.28. Those three add to 0.94, the fresh makeup the plant pulls.
Fresh water falls to 0.94 cubic metres a day, but liquid effluent is only 0.33. Most of the loss leaves on the belt or to the air. Material guidance for these loops sits in the industrial water spray nozzle reference.
| Loop element | Per day | Note |
|---|---|---|
| Gross spray volume | 4.68 m3 | 48 windows x 97.5 L |
| Recovered to sump | 3.74 m3 | 80 percent recovery target |
| Carry-off on the belt | 0.42 m3 | Leaves with the product |
| Evaporation and drift | 0.24 m3 | Open transfer point |
| Sump purge for solids | 0.28 m3 | Keeps the fines fraction down |
| Fresh makeup | 0.94 m3 | The three losses combined |
| Liquid effluent to drain | 0.33 m3 | Purge plus a periodic dump |
The Results
No plant audit figures appear below. These are working outcomes for this class of retrofit, built from the arithmetic above and held inside published bands.
| Metric | Before | After | Basis |
|---|---|---|---|
| Fresh water per day | 6.75 m3 | 0.94 m3 | Hose once per shift against sump makeup |
| Liquid effluent per day | 6.75 m3 | 0.33 m3 | Once-through against purge plus dump |
| Carry-back on the return strand | Unquantified film | 0.031 kg per square metre per pass | 0.1 percent of 200 t/h over 1.8 m2/s |
| Applied water per square metre | No set band | 0.90 L, about 0.9 mm film | 1.625 L/s over 1.8 m2/s |
| Water to solids ratio | No measurement | about 29 to 1 by mass | 0.90 kg against 0.031 kg |
| Manual wash labour | 45 min per day | 15 min per day | Inspection and top-up |
| Nozzles replaced per year | Not applicable | 4 to 8 of 12 | 5 percent flow rise trigger |
The fresh water cut is 86 percent, from 6.75 to 0.94 cubic metres a day. The effluent cut is 95 percent, because most of the recovered water stays in the loop.
The water to solids ratio is the number to keep. At about 29 to 1 by mass, the wash delivers roughly 29 kg of water per kilogram of fines. Published flushing guidance puts the practical band at 10 to 1 to 50 to 1.
Nozzle wear is the slow cost. Published data puts the flow rise at end of insert life near 5 to 10 percent, and 5 percent is the trigger here. Under a 5 percent flow rise, 8.125 x 1.05 = 8.53 L/min, a rise of 0.41 L/min. Across twelve nozzles that is 4.9 L/min for no extra impact.
Flow scales with diameter squared. A 5 percent flow rise means the orifice grew by the square root of 1.05, or 1.025, taking 2.5 mm to 2.56 mm. That is a wear of 0.06 mm.
Reading the Numbers Against Published Ranges
The comparison below is the check a maintenance planner should run. Nothing here is presented as measured fact.
| Figure in this case | Published class range | Verdict |
|---|---|---|
| Carry-back at 0.1 percent of throughput | 0.05 to 0.5 percent | Mid-band, conservative |
| Blade life at 4 to 12 weeks | Field range, abrasive service | Consistent |
| Sheet breakup at 50 to 100 diameters | Published rule of thumb | Standoff inside the band |
| Droplet VMD 150 to 300 micron at 4 bar | Published for hydraulic flat fans | Consistent |
| Carbide insert life 4,000 to 12,000 h | Published abrasive band | Middle of the band |
| Ceramic insert life 8,000 to 20,000 h | Published abrasion band | Longer, but brittle |
| Nozzle change at 5 percent flow rise | 5 to 10 percent trigger | Conservative end |
| Water to solids ratio about 29 to 1 | 10 to 1 to 50 to 1, flushing | Comfortable, not extreme |
Two of these need a caveat. Recovery percentage is a design target, and a sump with a poor settling zone sits under 70 percent. Insert life depends on solids hardness, so a plant handling silica sand lands at the low end of every band.
Dust control draws on the same water. A transfer point that already runs a dust suppression system starts with a wet belt, which helps carry-back and can hurt housekeeping. Where the product must stay dry, the move is dust suppression nozzles aimed at the air, as the wide angle nozzle guide sets out.
Why This Matters
The transfer point is where a conveyor spends the least and loses the most. Carry-back starts there, build-up collects there, and the scraper works hardest there. A wash bar on the return strand removes the film before it bonds.
The transferable asset is the order of the arithmetic, not the parts list. Surface rate from belt width and speed. Carry-back mass from throughput and a published percentage. The same variables set the sizing for conveyor spray systems on any belt width.
The decision checklist from this retrofit is short.
- Work the carry-back mass balance before quoting a nozzle.
- Set the standoff inside the coherent sheet length, not at the longest reach.
- Size the band from 2 x d x tan(theta / 2) and hold 10 to 15 percent overlap.
- Pre-wet ahead of the strip bar, and let the scraper finish a wet surface.
- Report recovery and effluent as two separate numbers, because they are.
Water decides most conveyor belt cleaning budgets. A site that can recirculate from a sump can afford a wash bar at 4 bar that a once-through site cannot. A site that cannot recirculate should still count the 6.75 cubic metres a day it already spends.
Most belt cleaning systems are judged on the run rather than the ledger. Predictable carry-back control is a conveyor that keeps its tracking, its splice, and its housekeeping hours.
The Final Call
BoreJet supplies spray bars and does the sizing arithmetic behind them: orifice flow, band geometry, overlap, pump duty, and sump ledger. The family used here sits in the flat fan nozzle range, and the pattern logic is covered by the spray patterns explained guide.
Send the belt speed, the belt width, the throughput, and the carry-back figure if you have one to the contact page. The reply comes back with nozzle count, orifice size, bar positions, and the water ratio your duty needs.
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.