Short answer: Size an industrial rack washer on peak hourly return rate, not daily total. The PTW-1900 handles up to 450 trays/hour at the Standard 6-minute cycle. If your peak return exceeds that, you need parallel machines — daily-average math will lie to you.
This article gives you the formulas, three worked examples (production bakery, hospital, airline catering), and the four bottleneck failure modes we see in the field when sizing math is wrong.
Why daily totals lie
A central commissary running 6,000 meals per day across two shifts (12 hours) produces an average of 500 trays/hour. On paper, a single PTW-1900 at 450 trays/hour standard is almost enough — buy two, you’re fine.
In reality, that 500 trays/hour average comes from a return curve that looks like a Bactrian camel:
- 7:00-8:30 AM: breakfast plating — 200 trays returned
- 9:00-11:00 AM: bowl-rinse, idle — 50 trays returned
- 11:30 AM-1:00 PM: lunch return — 900 trays in 90 minutes = 600/hr peak
- 1:00-4:00 PM: prep — 200 trays returned
- 5:30-7:30 PM: dinner return — 800 trays in 120 minutes = 400/hr
The peak — 600 trays/hour — exceeds what one PTW-1900 can do. Two machines isn’t enough buffer either; the dishwashing operation falls behind 150 trays in the lunch return window, then needs 30 minutes after lunch service to catch up. Prep cooks stand around waiting for clean GN pans. That’s a real operational failure mode.
The single most expensive procurement mistake in this category is sizing for average instead of peak.
The throughput formula
Peak hourly demand (trays/hr) = (peak meal count) × (trays per meal) / (peak window hours)
Machines required = ⌈ Peak hourly demand / Machine throughput at intended cycle ⌉Where:
- Peak meal count = the largest single-meal-period output your kitchen produces (typically lunch in commissary, dinner in fine dining, evening crew in factory)
- Trays per meal = total trays consumed per meal portion, including hot-line GN, plating trays, bowl-rinse, and bus-back
- Peak window = the wall-clock hours during which those trays return (almost always 60-120 minutes, not the full shift)
- Machine throughput = real-world cycle output, NOT the marketing number; PTW-1900 Standard = 450/hr, Heavy = 280/hr, Quick = 540/hr (light soil only)
The ceiling function ⌈ ⌉ matters — round up. A wash bay running 95% capacity at peak has no margin for one cycle of pan-load misalignment, one operator restroom break, or one cycle dump-and-refill of the wash tank.
Worked example 1: production bakery
A wholesale bakery produces 12,000 baguettes + 8,000 viennoiserie (croissants, brioche, pain au chocolat) + 4,000 sheet-cake portions per shift, two shifts/day, six days/week.
- Bread tins: 12,000 ÷ 4 baguettes/tin = 3,000 tin events
- Viennoiserie sheet pans: 8,000 ÷ 12 pastries/pan = 667 pan events
- Sheet cake pans: 4,000 ÷ 24 portions/pan = 167 pan events
- Bowl-rinse + tools: ~150 events
- Total per shift: ~4,000 events
Spread over an 8-hour shift, that’s a 500/hr average — but the bakery’s actual production cycle concentrates 4-batch viennoiserie returns into 2 hours (8:00-10:00 AM and 6:00-8:00 PM):
- Peak return = 2,500 events ÷ 2 hours = 1,250 events/hr
Single PTW-1900 capacity: 450/hr. Required: ⌈1,250 ÷ 450⌉ = 3 machines parallel for peak coverage.
That’s not what the daily-average math says. Daily-average says 500/hr ÷ 450 = 2 machines. The 3rd machine pays for itself in the avoided lunch-recovery production delay and the operator labour redeployed during off-peak.
Worked example 2: hospital central kitchen
A 600-bed teaching hospital serves 5,400 patient meals/day on standardized polycarbonate trays (1 tray per meal, plus GN pans in the production line).
- Patient trays returning to dish room: 5,400/day × 1 service cycle/meal = 5,400 tray events
- Daily total events including production GN + cook-chill carriers + utensils: ~7,200
Peak return windows:
- Breakfast return: 8:30-9:30 AM = 600 trays in 60 min = 600/hr
- Lunch return: 12:30-1:30 PM = 1,800 trays in 60 min = 1,800/hr (this is the binding constraint)
- Dinner return: 5:30-6:30 PM = 1,200 trays in 60 min = 1,200/hr
Required: ⌈1,800 ÷ 450⌉ = 4 machines parallel for lunch peak.
Most hospitals install 2 machines and live with a 30-45 minute meal-return overflow into nurse-station holding carts. Survivable but creates IPC concerns (uncleaned trays accumulating in ward areas). Hospitals that install 4 machines or use staggered ward-floor return schedules avoid the overflow entirely.
Worked example 3: airline catering hub
LSG-class catering center, 80,000 meals/day, processing inbound trolleys from arriving flights.
- Trolleys per day: ~3,000 (14-20 trolleys per widebody flight × ~180 flights)
- Daily total trolley events: 3,000
Peak inbound return: morning hub arrivals concentrate 800 trolleys in 4 hours = 200 trolleys/hour peak.
PTW-1900 trolley capacity: 1 trolley per 6-minute cycle = 10 trolleys/hour per machine.
Required: ⌈200 ÷ 10⌉ = 20 machines parallel. That’s a hub-class installation — typical at Emirates Flight Catering Dubai, LSG Frankfurt, Gate Gourmet HKG, SATS Singapore.
For a spoke catering center handling 20,000 meals/day, peak might be 50 trolleys/hour = 5 machines. Common configuration in regional hubs.
The four bottleneck failure modes
When wash bay throughput falls short of peak demand, the operation fails in one of four predictable ways:
| Failure mode | Symptom | Operational impact |
|---|---|---|
| 1. Pan-out | Prep cooks waiting for GN pans / sheet pans / utensils | 20-40% labour productivity loss in next-meal prep window |
| 2. Cart accumulation | Dirty carts/trolleys stacking in service aisle | Sanitation citation risk + Joint Commission EC findings |
| 3. Overrun into next service | Wash bay still finishing breakfast at lunch service | Cross-meal contamination risk; HACCP CCP breach |
| 4. Operator burnout | Wash bay staff turnover at 240%+/yr | Hiring cost USD 8K-15K per replacement |
The economics of pan-out alone — 20-40% labour loss in a 30-cook prep brigade for the 90 minutes after a lunch peak — costs more per year than a third wash machine.
Throughput at different cycle profiles
The PTW-1900 supports four stored PLC cycle profiles, each with different throughput:
| Cycle | Time | Trays/cycle | Trays/hour | Best for |
|---|---|---|---|---|
| Quick | 4 min | 45 | 540 | Light-soil dessert plates, room-service trays |
| Standard | 6 min | 45 | 450 | General-purpose wash, default cycle |
| Heavy | 9-12 min | 45 | 225-300 | Caramelized sugar, baked-on protein, rotisserie carbon |
| Allergen Flush | 90 sec | 45 | rinse-only | Between-batch segregation, no detergent |
Sizing must use the cycle profile that matches the dirtiest expected load. If 30% of your peak is rotisserie carriers needing Heavy cycle, your effective peak throughput is the weighted average:
Effective throughput = (0.30 × 280) + (0.70 × 450) = 399 trays/hrNot 450. Sizing on 450 against a 30%-Heavy load mix would shortcut you 11% on capacity.
How to gather your real return curve data
If you’re sizing for an existing operation, capture two weeks of wash-bay tally data:
- Place a clicker counter at the wash bay entrance
- Operators tick every cart/rack arrival per 15-minute interval
- Compute the 95th-percentile 60-minute window across the 2-week sample
- Use that as your peak demand for sizing
For new operations, derive peak from menu engineering:
- Lunch service: model 50% of daily output in 90 minutes
- Production runs: assume 60% of batch sheet-pan return in 2-hour blast-chill window
- Catering center: assume hub-flight-bank concentration of 40% of daily volume in 4-hour bank
Common sizing FAQ
Q: Can I get away with one machine if my peak is 480 trays/hour (5% over capacity)? A: Technically yes if you can stretch peak by 6-7 minutes. But a 5% overshoot leaves zero headroom for cycle-dump events (the wash tank dumps every 6 cycles), operator breaks, or maintenance. We recommend installing a second machine when peak demand exceeds 80% of single-machine throughput sustained — about 360 trays/hour.
Q: What’s the right cycle to size on if my soil profile varies through the day? A: Use the weighted average across your load mix (formula above). For most foodservice operations the right baseline is Standard cycle (450/hr); production plants with cooked-cooled GN should baseline on Heavy (280-300/hr).
Q: Does staggering meal service shifts help? A: Yes, often dramatically. Staggering lunch service across two 45-minute windows cuts peak return from 1,800 trays/hr to roughly 1,000 trays/hr — turning a 4-machine required size into a 3-machine adequate size.
Q: How much extra throughput buffer is best practice? A: 20-25% over your computed peak. Industrial operations rarely hit theoretical capacity due to maintenance windows, operator availability, and cycle-dump events. The marginal cost of one extra machine over the projected 15-year service life is small relative to throughput failure cost.
Q: Does the PTW-1900 cycle time change with chamber load? A: No. The PTW-1900 runs fixed-duration cycles. Under-loaded racks cycle the full 6 minutes; this is by design — it ensures consistent thermal dose (82°C dwell) regardless of load.
Related reading
- PTW-1900 full specifications
- How Roll-in Rack Washers Work
- Wash Cycle Time Optimization
- How to Choose an Industrial Rack Washer — foundational buyer guide
- Water Quality Requirements — inlet pressure/temperature affects achievable throughput
- Labor Savings from Automating Tray Cleaning — throughput drives labor displacement
- 36-Month ROI Analysis — companion economics article
- Request a site assessment — V-TAI engineers compute your throughput sizing for free