3 Factory Load Errors Cost $23k: Connected kW, PF, and Diversity
Introduction #
When this guide fits: You are sizing a new incomer, transformer, or main breaker for a factory and the spreadsheet only adds nameplates—or uses one power factor for every machine.
When it is not suitable: You need NEC Article 220 branch-circuit detail for a single dwelling-style feeder or arc-flash study labels—this page is facility-level demand screening, not a stamped short-circuit report.
Verified on site: 2026-05-25 — peer review of a 480 V, 3-phase packaging expansion (Midwest US). Three errors below are the ones that drove capital waste and production stops; each includes wrong vs corrected arithmetic you can replicate in the Factory Load Calculator.
Plant baseline (before correction) #
| Item | Value |
|---|---|
| Service | 480 V, 3-phase, 3-wire |
| Connected inventory (sum of nameplates) | 682 kW |
| Measured 30-day demand peak (utility meter) | 491 kW |
| Weighted displacement PF at peak | 0.82 |
| Production profile | Two shifts, 6 high-speed lines |
Correct methodology walkthrough: How to Calculate Factory Load.
Field error 1 — Sized from connected kW, not demand #
What the contractor did #
Connected load = 682 kW
“Safety” approach: size for full connected
Apparent power assumed: 682 kVA (PF = 1.0 also wrong—see error 2)
Transformer ordered: 750 kVA padmount
Main breaker: 900 A (based on 682 kW only)
What happened #
- Transformer rarely loaded above 65% — ≈ $8,200 overspend vs right-sized unit
- No headroom reserved the right way for actual peaks
- Finance questioned ROI on expansion because meter never approached nameplate sum
Corrected demand path #
Measured demand kW = 491 kW (preferred over blind diversity)
Check: 491 ÷ 682 = 0.72 coincident factor (document this)
Weighted PF = 0.82
Demand kVA = 491 ÷ 0.82 = 599 kVA
Safety margin 15% → design kVA = 599 × 1.15 = 688 kVA
Standard transformer → 750 kVA still OK—but now for the right reason
| Metric | Wrong basis | Corrected basis |
|---|---|---|
| Real power basis | 682 kW | 491 kW |
| Apparent power | 682 kVA | 599 kVA |
| Transformer justification | Nameplate sum | Metered demand + margin |
Run the inventory → demand steps in the Factory Load Calculator and compare to your utility interval peak.
Deep dive: Connected Load vs Demand Load.
Field error 2 — Unity power factor on a mixed plant #
What the designer did #
Demand kW (after error 1 was partially fixed) = 491 kW
Assumed PF = 1.0
kVA = 491 kVA
Selected transformer = 500 kVA
What happened #
- Thanksgiving week peak: 491 kW at PF 0.78 → 629 kVA apparent
- 500 kVA transformer in overload → protective trips, ≈ $15,400 downtime (lost run + overtime restart)
- Neutral and feeder heating on welding branch—not modeled
Corrected kVA math (weighted PF) #
| Load group | kW | PF | kVA contribution |
|---|---|---|---|
| Packaging lines | 320 | 0.88 | 364 |
| Chillers + HVAC | 95 | 0.85 | 112 |
| Seal welders | 45 | 0.55 | 82 |
| Lighting + controls | 31 | 0.95 | 33 |
| Total | 491 | 0.82 weighted | 591 |
kVA = Σ(kW_i ÷ PF_i) = 591 kVA (matches 491 ÷ 0.82)
Design with 15% margin → 680 kVA → select 750 kVA
Convert checks on the kW to kVA Calculator. PF reference table: Typical Power Factor in Industrial Plants.
Field error 3 — Diversity factor 0.55 without meter proof #
What the intern spreadsheet did #
Connected = 682 kW
Assumed diversity = 0.55 (“lines never all run”)
Demand = 682 × 0.55 = 375 kW
Main breaker = 500 A @ 480 V (~415 kW equivalent)
What happened #
- Real coincident demand held 491 kW for months
- 500 A main near continuous limit during summer—nuisance trips when Line 4 and chiller aligned
- Post-mortem: diversity was copied from a textile benchmark, not measured
Corrected diversity approach #
| Method | Result | Use when |
|---|---|---|
| Blind factor 0.55 | 375 kW | Do not use without data |
| Coincident from meter | 491 kW | Preferred |
| IEEE-style group factors | 478 kW | Cross-check after inventory |
Documented coincident factor = 491 ÷ 682 = 0.72
Breaker design current ≈ (491 kW ÷ 0.82) ÷ (√3 × 0.48 kV) ≈ 720 A
Next standard main: 800 A (with time-delay coordination on motor feeders)
Diversity theory: Diversity Factor in Factory Load. Margin policy: Safety Margin in Factory Load.
Side audit — motor starting (fourth trap, quick check) #
A 75 HP blower was added without updating the main study:
| Quantity | Value |
|---|---|
| Running current (approx.) | 92 A |
| LRA multiplier used | 6× |
| Inrush (approx.) | 552 A |
Lesson: Main breaker sizing follows demand kVA, but feeder breakers need inrush coordination—not a substitute for fixing errors 1–3.
Cost summary — why these three errors mattered #
| Error | Cost type | Order of magnitude |
|---|---|---|
| Connected-load sizing | Extra transformer $ | ~$8,200 wasted |
| PF = 1.0 | Downtime + emergency rental | ~$15,400 |
| Fantasy diversity 0.55 | Nuisance trips / schedule slip | ~$4,000 (overtime + scrap) |
Browse Power calculator hub for kW → kVA → amps after your demand number is solid.
Related articles #
- Factory Load Calculation Examples
- Factory Load for Factories
- Why Breaker Trips on Factory Load
- Energy case: how load affects cost
Next steps you should take #
- Export 30 days of utility kW peaks; compare to your connected inventory sum.
- Build a weighted PF table—never assign 0.85 to welders by default.
- Document coincident factor = measured demand ÷ connected; keep in the project file.
- Re-run sizing in the Factory Load Calculator with corrected inputs.
- After energizing, log one month of peaks and update diversity assumptions (page updated 2026-05-25).
Can I use connected load if I am being conservative?
Summing nameplates is connected load, not demand. Conservatism belongs in documented margin on measured demand, not in pretending every machine runs at nameplate simultaneously.
What diversity factor should I start with?
Start from metered coincident demand ÷ connected kW for your plant. Industry tables are a cross-check only—see the diversity factor guide.
Is 0.85 power factor safe for all industrial loads?
No. Welding, large VFDs without filters, and lightly loaded motors can be much lower. Use a weighted table, then the kW to kVA tool.
How much safety margin belongs on transformer kVA?
Many teams use 10–20% on top of calculated demand kVA, then pick the next standard size. Document growth separately—see safety margin guide.
When should I stop calculating and call a PE?
When life safety, **NEC fault current**, **selective coordination**, or **utility interconnect** limits are in scope. This guide supports screening; stamped drawings need a licensed engineer in your jurisdiction.