Maximum Demand Calculation — Diversity Factors Explained Step-by-Step

A young engineer once presented me with a transformer sizing calculation for a 200-unit apartment complex. Each unit had a 63A main breaker at 230V — a connected load of 14.5 kW per unit. His conclusion: the building needed a 200 × 14.5 = 2,900 kVA supply, requiring two 1,500 kVA transformers.

The actual maximum demand of that building, measured over 12 months after occupation, peaked at 680 kVA. He had over-estimated by 4.3× — recommending $180,000 worth of transformer capacity that would never be used.

The missing concept was diversity — the statistical certainty that not all loads operate simultaneously at their maximum rating.

Connected Load vs Maximum Demand

Connected load is the sum of all individual load ratings. It's the theoretical maximum if every appliance, light, and socket outlet in a building operated at full capacity simultaneously. This never happens.

Maximum demand (or After Diversity Maximum Demand — ADMD) is the highest actual load that the installation draws from the supply over a representative period. It accounts for the fact that:

• Not all circuits operate simultaneously
• Most loads operate below their rated capacity
• Load peaks from different circuits occur at different times
• Some loads are intermittent (motors, lifts, HVAC compressors cycle on and off)

Diversity factors are always less than 1.0 (often much less). The more loads in a system, the lower the diversity factor — because the probability of ALL loads operating simultaneously decreases with the number of loads.

How Diversity Factors Work

Consider a simple example: 10 identical apartments, each with a 10 kW connected load.

• The connected load is 100 kW
• Each apartment's peak demand occurs at different times (one is cooking while another is idle; one is running air conditioning while another is switched off)
• The measured maximum demand might be 45 kW — a diversity factor of 0.45

Now consider 100 identical apartments:

• Connected load: 1,000 kW
• Maximum demand: approximately 300 kW — diversity factor of 0.30

And 500 apartments:

• Connected load: 5,000 kW
• Maximum demand: approximately 1,200 kW — diversity factor of 0.24

Standard Methods

AS/NZS 3000 (Wiring Rules)

AS/NZS 3000, Section 2, Part C — Maximum demand assessment

AS/NZS 3000 provides specific tables for calculating maximum demand by load category. The method:

1. List all loads by category (lighting, socket outlets, cooking, heating, air conditioning, motors)
2. Apply the specific demand factor for each category:
   • Lighting: first 2 kVA at 100%, remainder at 50%
   • Socket outlets: first 20 at 10A each (100%), next 20 at 5A each, remainder at 2.5A each
   • Cooking: first appliance at 100%, each additional at a reducing percentage
3. Sum the diversified loads to get the maximum demand
4. Apply an overall diversity factor for the number of consumers (for multi-unit installations)

The diversity factors for multiple residential units in AS/NZS 3000:

BS 7671 (IET Wiring Regulations)

BS 7671, Appendix A, Table A1 — Current demand of points of utilisation

BS 7671 Appendix A provides guidance on assessing current demand. The approach is similar but expressed differently:

• Socket outlets: first point at rated current, additional points at 40% or less depending on usage
• Lighting: 66 VA per point (standard), or rated value for discharge lighting
• Cooking: 10A + 30% of remainder above 10A + 5A per additional appliance
• Space heating: full rated current (no diversity for heating — it all runs when it's cold)
• Motors: full-load current of largest motor + 40% of remaining motors

IEC 60364

IEC 60364-8-1, Clause 8.1 — Energy efficiency — assessment of demand

IEC 60364 takes a more general approach, requiring the designer to assess maximum demand based on "knowledge of the intended use of the installation." It references load profiles and diversity factors but doesn't provide specific tables — instead, national committees (who adopt IEC 60364 into their national standards) provide the detailed factors.

The Common Mistakes

Mistake 1: Adding Up Circuit Breaker Ratings

The most egregious error. A 20A MCB on a lighting circuit doesn't mean that circuit draws 20A. A typical residential lighting circuit draws 2–3A. The MCB is sized for protection, not to indicate the actual load.

Adding up all MCB ratings in a distribution board gives the breaking capacity sum, not the maximum demand. For a typical residential DB with a 63A main breaker and twenty 20A MCBs, the sum of MCB ratings is 400A — but the maximum demand is the 63A main breaker rating (which itself includes diversity).

Mistake 2: Applying Diversity to the Wrong Level

Diversity factors apply at each level of the distribution hierarchy, but you can't apply them cumulatively without understanding the basis.

Correct approach: Calculate the maximum demand at each distribution board using the appropriate diversity factors for the loads it serves. Then apply inter-board diversity to get the main switchboard demand.

Incorrect approach: Calculate the undiversified load at each board, sum them all, and apply a single blanket diversity factor. This ignores the different diversity characteristics of different load types.

Mistake 3: Using Generic Diversity Factors for Specialised Buildings

The diversity factors in standards are based on typical residential or commercial usage patterns. They don't apply to:

• Data centres: virtually 100% diversity factor (all equipment runs continuously)
• Industrial process plants: diversity depends entirely on the process — a batch plant has high diversity, a continuous process has low diversity
• Hospitals: lighting and HVAC have normal diversity, but medical equipment and critical services have near-zero diversity
• Mining operations: process plant diversity is typically 0.7–0.85 (much higher than commercial buildings because the equipment runs continuously during operation)

Mistake 4: Ignoring Future Growth

A maximum demand calculation is a snapshot. If the building is likely to add loads in future (EV charging, additional air conditioning, kitchen upgrades in apartments), the transformer and main cable should be sized with margin.

Good practice: add 20–30% growth allowance for commercial buildings, and specifically account for EV charging infrastructure in residential developments. Many jurisdictions now require EV-ready provisions for all new residential units.

The Cost of Getting It Wrong

Oversized Supply (Most Common)

• Transformer operates at 20–30% of capacity — poor power factor, higher losses, higher capital cost
• Switchboard rated for unnecessary fault level (larger transformer = higher fault current)
• Main cables oversized — wasted copper, larger cable routes, higher installation cost
• Typical cost penalty: 30–50% more than a correctly sized installation

Undersized Supply (Dangerous)

• Transformer overloads during peak periods — accelerated insulation ageing, potential failure
• Voltage drops below acceptable limits — equipment malfunctions, motor stalling
• Protection devices trip under normal load — service interruptions
• Main cables overheat — fire risk
• Typical resolution: emergency transformer upgrade ($50,000–200,000 depending on size), disruption to occupants

Worked Example: Small Commercial Building

Building: 3-storey office building, 2,000m² total

Overall diversity factor: 210/325 = 0.65

For transformer sizing (accounting for power factor of 0.85):

Without diversity analysis, the engineer would calculate 325/0.85 = 382 kVA → select 500 kVA. The 500 kVA transformer costs approximately $8,000–12,000 more than the 315 kVA, has higher no-load losses (running costs), and delivers higher fault current to the switchboard (potentially requiring a higher-rated board).

The Mining Context

At a large-scale mining operation, maximum demand for the processing plant was calculated using actual motor duty cycles from the process design. The SAG mill (6 MW), ball mills, flotation cells, thickeners, and pumps have well-defined operating profiles. The overall plant diversity factor was 0.78 — meaning the 45 MW connected load had a maximum demand of approximately 35 MW.

But during plant startup — when all conveyors, mills, and pumps start in sequence — the demand transiently exceeds the steady-state maximum by 40% due to motor starting currents. The power supply and emergency generation had to be sized for this starting sequence, not just the steady-state demand.

This illustrates a key point: maximum demand must consider both steady-state and transient conditions.

Practical Recommendations

1. Always perform a formal maximum demand assessment — never use the sum of MCB ratings or connected load as the design basis
2. Use the appropriate standard method for your jurisdiction (AS/NZS 3000 tables, BS 7671 Appendix A, or local national annex to IEC 60364)
3. Apply different diversity factors for different load types — don't use a blanket factor
4. Never apply diversity to heating or data centre loads — these run at or near full capacity when they're needed
5. Allow for future growth — 20–30% for commercial, specific provision for EV charging in residential
6. Validate against real data where possible — metered data from similar buildings is the best design basis

Related Resources

• Auckland CBD Crisis: Maximum Demand Calculation — How underestimated demand caused a city-wide power failure
• Maximum Demand: The 20-Unit Apartment Block — 14% gap in demand between standards
• Cable Sizing: The 50m Office Feeder — Cable sizing flows directly from demand calculation
• View all worked examples →