How to size a building transformer for peak load without overspending

Sizing a building transformer for peak load is a critical decision for anyone balancing reliability, efficiency, construction constraints, and budget control. An undersized unit can overheat, create unacceptable voltage drop, restrict expansion, and turn routine switching events into operational risks. A transformer selected with excessive margin, however, ties up capital in capacity that may never be used and can operate inefficiently at persistently low loading.

The right rating is not simply the total nameplate load divided by a power factor. It is the result of understanding how the building will actually consume power: which loads run together, how long peak periods last, whether large motors start across the line, how nonlinear electronic loads affect heating, and whether future capacity is a defined requirement or merely an assumption.

A sound building transformer selection therefore begins with a load model, not with the largest standard kVA size available from a supplier.

Start with the load that will exist in operation, not the connected load on drawings

Electrical schedules often show connected loads for every circuit, panel, mechanical system, and tenant allowance. This is necessary for design coordination, but connected load is rarely the load that the transformer must supply at one time. Lighting may be fully energized during business hours, while electric heating, chillers, kitchen equipment, lifts, pumps, and process equipment follow different operating cycles.

The key figure is the maximum coincident demand: the highest realistic kVA expected at the transformer terminals during normal operation. This should be calculated from a load schedule that distinguishes between:

• continuous loads, such as base lighting, servers, ventilation, refrigeration, and process equipment;
• intermittent loads, including lifts, fire pumps during testing, kitchen equipment, and workshop machinery;
• seasonal loads, particularly cooling and heating systems;
• standby or alternate loads that are not intended to operate simultaneously;
• future tenant, production, or EV-charging allowances.

Demand factors and diversity factors are useful, but only when they reflect the actual building use. Applying a generic diversity percentage to an entire load list can produce a deceptively neat calculation and a poor field result. A mixed-use building, a data-intensive office, a cold-storage facility, and a residential tower may have similar connected loads but very different peak-demand patterns.

Where an existing site is being expanded, interval metering data is usually more valuable than any generic demand factor. At least 12 months of data is preferable because it captures seasonal peaks. The useful record is not merely monthly energy consumption in kWh; it is the highest measured kW or kVA demand, the time of occurrence, the power factor at peak, and the duration of that peak.

Convert demand into transformer kVA correctly

Transformers are rated in kVA, while many building loads are listed in kW. The basic conversion is:

Required kVA = Maximum demand in kW ÷ operating power factor

For example, if the calculated coincident demand is 800 kW and the expected power factor at peak is 0.90, the initial transformer demand is approximately 889 kVA. That does not automatically mean a 1,000 kVA transformer is the correct selection. The calculation still needs to account for the load profile, ambient conditions, harmonics, starting duty, utility requirements, and expansion plan.

Power factor should be treated as an operating input, not an optimistic design assumption. Capacitor banks may improve the site power factor, but their performance can vary with load level and switching arrangements. In installations with variable-speed drives, UPS systems, rectifiers, or significant harmonic content, conventional capacitor correction must also be coordinated carefully with harmonic studies and detuned-reactor requirements.

It is also important not to confuse power factor with transformer loading. A 1,000 kVA transformer can theoretically supply 1,000 kVA at its rated conditions. At a 0.8 power factor, that equates to 800 kW; at 0.95 power factor, it equates to 950 kW. The transformer thermal limit remains kVA-based.

Peak demand is not the same as a short electrical event

Some loads create very high current for a short time without requiring a substantially larger transformer. Motor starting is the most common example. A direct-on-line motor can draw several times its full-load current during acceleration. The relevant question is whether that starting current causes an excessive voltage dip at the motor terminals or disturbs other sensitive loads on the same system.

This is particularly important where a building transformer serves large chillers, pumps, compressors, crushers, elevators, or ventilation fans. The selection process should examine:

• motor rating and starting method;
• locked-rotor current or starting-current profile;
• transformer impedance;
• upstream network strength and available fault level;
• cable length and voltage drop to the motor control center;
• the effect on lighting, controls, IT systems, and other operating motors.

Increasing transformer kVA can reduce voltage dip, but it is not always the most economical answer. Soft starters, variable-frequency drives, sequential motor-start logic, or a revised mechanical operating sequence may solve the problem at lower overall cost. Conversely, selecting a larger transformer without checking fault-current implications can create a separate issue: switchgear and protective devices may need higher interrupting ratings.

Do not use “future growth” as a vague reason to oversize

Future capacity is one of the most common justifications for selecting the next transformer size up. In many projects, that is prudent. In others, it is simply an untested contingency that adds cost to the transformer, switchgear, cables, civil works, and protection system.

A better approach is to divide future demand into three categories.

Committed growth includes signed tenant requirements, approved production equipment, an expansion phase already funded, or a contractual EV-charging deployment. This load should normally be included in the transformer rating.

Probable growth may include a planned but not yet contracted expansion. It can justify physical provisions such as spare feeder ways, space for a second transformer, cable-duct capacity, and protection-panel expansion. Whether it justifies immediate transformer capacity depends on the cost and disruption of future replacement.

Speculative growth is a general expectation that the building “may need more power later.” This is usually better addressed through a staged distribution strategy than by purchasing large unused capacity on day one.

For example, if the current peak demand supports an 800 kVA calculation and a possible future load could add 250 kVA, the decision is not automatically a 1,250 kVA installation. A project may be better served by a 1,000 kVA unit with a clear load-management plan, or by two smaller transformers with one position reserved for later expansion. The answer depends on load criticality, outage tolerance, available footprint, utility connection conditions, and the economics of a future shutdown.

Evaluate loading over time, not only the calculated maximum

Transformer efficiency is not constant at every loading point. Core losses occur whenever the transformer is energized, while winding losses rise approximately with the square of current. A unit that is greatly oversized may carry unnecessary no-load losses continuously. A unit operated close to rating for long periods may have acceptable performance if its thermal design and ambient conditions support it, but it leaves less room for abnormal operating states and growth.

The useful target is not a universal loading percentage. A building with a stable, well-understood load and good maintenance access can reasonably operate closer to its selected rating than a facility with uncertain tenant demand, frequent overload episodes, or critical continuity requirements.

For most decisions, compare options using lifecycle cost rather than purchase price alone. The comparison should include:

• transformer purchase and transport cost;
• installation, foundation, fire protection, and enclosure requirements;
• no-load and load losses at the expected annual load profile;
• energy price and expected operating hours;
• maintenance access and outage costs;
• the cost of future expansion or replacement;
• risk exposure if a single transformer fails.

A lower-loss transformer can be commercially justified even when its purchase price is higher, particularly in facilities with high annual operating hours. However, loss figures must be compared on the same basis: rated voltage, tap position, cooling method, impedance, applicable standard, and stated temperature reference. Quoted losses without these conditions are not a reliable procurement comparison.

Harmonics can change the thermal picture

Modern buildings contain growing concentrations of nonlinear loads: UPS systems, LED drivers, data equipment, battery chargers, variable-speed drives, rectifiers, and EV chargers. These loads can introduce harmonic currents that increase eddy-current losses and heating in transformer windings, leads, and associated distribution equipment.

A conventional kVA calculation may therefore be insufficient where the nonlinear load share is high. The design should identify the expected harmonic spectrum and total demand distortion, then confirm whether a standard transformer rating is suitable or whether derating, enhanced thermal design, electrostatic shielding, a suitable vector group, or harmonic mitigation is required.

This is not a reason to specify special features automatically. It is a reason to obtain load data early. A transformer built for a heavily harmonic commercial or industrial load may differ materially from one serving predominantly linear lighting and motor loads, even when both have the same nominal kVA rating.

Transformer impedance is a system decision

Impedance is often treated as a supplier detail, but it directly affects both voltage regulation and short-circuit current. Lower impedance generally reduces voltage drop under load and can improve motor-starting performance, yet it raises downstream fault current. Higher impedance limits fault current but may worsen voltage dip and regulation.

The appropriate impedance must be coordinated with the low-voltage switchboard, busduct, feeder cables, protective-device ratings, and discrimination study. A transformer cannot be selected independently from the rest of the distribution system.

This is especially relevant when replacing an existing unit with a larger rating. A new transformer may have a different impedance percentage and significantly increase available fault current at the main switchboard. If the existing board’s short-circuit withstand rating is exceeded, the project scope can expand rapidly.

Choose the transformer type around the site, not a preference

The choice between oil-immersed and dry-type construction is often driven by location, fire strategy, access, environmental conditions, and local approval requirements rather than electrical rating alone. Dry-type units may be appropriate indoors where fire and spill considerations dominate. Oil-immersed units can offer strong performance and may be well suited to outdoor substations or dedicated transformer rooms designed with the necessary containment, ventilation, fire protection, and access arrangements.

For projects with medium-voltage distribution, a solution such as a 33kV Oil-Immersed Power Distribution Transformer should be assessed as part of the complete substation arrangement: primary voltage, secondary voltage, vector group, tap range, cooling designation, impedance, insulation level, losses, protection interfaces, and site environmental conditions all need to align with the network design.

Applicable technical requirements should be stated clearly in the procurement specification. IEC 60076 is widely used for power-transformer design and testing, but local grid codes, electrical installation rules, fire regulations, utility connection standards, and national requirements may impose additional conditions. The utility may also specify voltage regulation limits, metering arrangements, fault-level assumptions, protection coordination, or restrictions on parallel operation.

Build resilience separately from capacity

A frequent mistake is to assume that a larger single transformer provides reliability. It provides more capacity, but it remains a single point of failure. If continuity is critical, resilience should be evaluated through network architecture: dual transformers, sectionalized busbars, normally open ties, separate essential-load boards, standby generation, battery storage, or carefully defined load shedding.

Two transformers do not automatically provide better value. They increase equipment count, protection complexity, space requirements, and no-load losses if both remain energized at light load. But where outage costs are high, or where maintenance must be performed without shutting down the facility, a split-load arrangement can be more economical over the operating life of the asset.

The decision should distinguish clearly between “capacity required under normal operation” and “capacity required after a transformer outage.” Those are different design cases and should be documented separately.

A practical approval test before releasing the order

Before placing a transformer order, the electrical design should be able to answer a short set of operational questions without relying on broad safety margins. What is the calculated coincident peak kVA? Which assumptions create that result? What is the longest expected high-load period? Are seasonal conditions represented? Can large motors start without unacceptable voltage dip? Has harmonic loading been considered? Does the selected impedance keep fault duties within equipment ratings? Which future loads are committed, and which are only possible? What happens if the transformer is unavailable?

If these questions have clear, documented answers, the selected building transformer is much more likely to be neither undersized nor unnecessarily oversized. The objective is not to buy the largest practical unit. It is to select capacity that matches the real load, protects the distribution system, supports a credible growth path, and avoids paying for idle electrical infrastructure throughout the life of the building.