A distribution transformer is sized on measured maximum demand plus a defined growth allowance, not on connected load and not on a round number chosen for comfort. Connected load is the arithmetic sum of every nameplate on site. Maximum demand is the highest coincident kVA the site actually draws, and on an industrial plant it sits well below connected load because loads do not all run together. Getting the gap between those two numbers wrong is the most expensive error in an LT distribution design, and it costs money in both directions.

TL;DR

• Size on measured maximum demand plus a stated growth allowance, never on connected load.

• IS 1180 (Part 1): 2014, Amendment 4, caps total losses at both 50 percent and 100 percent load, so the standard itself assumes part-load running is normal.

• Oversizing raises no-load loss, which runs 8,760 hours a year regardless of production. Undersizing raises load loss, which rises with the square of loading. Which one costs you depends on load factor: at IS 1180 Level 2 ceilings a 500 kVA unit beats an 800 kVA unit below about 240 kVA average demand and loses above it.

• Demand no-load loss and load loss as two separately guaranteed figures. A single total-loss number cannot be converted into an annual energy cost.

• The largest direct-on-line motor often forces a bigger transformer than the load does, through starting voltage dip.

• IS 2026 (Part 2): 2010 cuts permitted temperature rise K for K when site ambient exceeds 40 degrees C monthly average or 32 degrees C yearly average, and that derating must appear on the rating plate.

What kVA rating does a distribution transformer actually need?

The rating is site maximum demand in kVA divided by the target loading you intend to run at, rounded up to the next standard rating, then checked against motor starting and site ambient. For most industrial plants the sensible target is 70 to 85 percent loading at present-day peak, which leaves headroom for growth without parking a large fixed loss on the bill for a decade.

Six steps get you there:

1. Establish maximum demand, not connected load. Use twelve months of DISCOM billing data where the site exists, or a load schedule with demand factors where it does not.

2. Convert kW to kVA at the real power factor. A 300 kW demand at 0.85 power factor is 353 kVA. Corrected to 0.98 by an APFC panel it is 306 kVA, and that difference alone moves you a full rating step.

3. Add a stated growth allowance, as a number and a horizon, for example 15 percent over five years, so it can be reviewed rather than inherited.

4. Check the largest motor start against the resulting kVA. This step reverses more sizing decisions than any other.

5. Apply site derating for ambient temperature and altitude under IS 2026 (Part 2).

6. Round up to a standard rating. IS 1180 (Part 1) standard three-phase ratings run 16, 25, 63, 100, 160 and 200 kVA, then 250, 315, 400, 500, 630, 800, 1000, 1250, 1600, 2000 and 2500 kVA. Non-standard ratings cost more, take longer and complicate spares.

New India Electricals Ltd supplies distribution transformers from 15 to 2500 kVA, at 11 kV, 22 kV and 33 kV primary, with secondary voltages of 433 V, 550 V or 660 V, vector group Dyn11, continuous duty, copper or aluminium wound, built to IS 2026, and BS 171.

Why connected load is the wrong number to size a transformer on

Connected load overstates demand because plant loads are not coincident. Two factors close the gap and they are not the same thing. Demand factor is the ratio of a load group’s maximum demand to its connected load, and it accounts for equipment running below nameplate or intermittently. Diversity factor is the ratio of the sum of individual group maximum demands to the coincident maximum demand at the incomer, and it accounts for groups peaking at different times.

The worked example below is illustrative. Replace the factors with your own measured data before they enter a specification.

Sum of group demands is 340 kW. Apply a diversity factor of 1.25 at the incomer and coincident demand is 272 kW, which is 286 kVA at 0.95 power factor. Sizing on the 560 kW connected load would have pointed at 630 or 800 kVA. Sizing on demand points at 400 kVA running at 72 percent loading, with room to grow. Write down which factor you used for each group and why. A load schedule with unexplained factors is a guess with a table around it.

What does oversizing a distribution transformer actually cost per year?

Oversizing costs money through no-load loss, which is fixed, present whenever the transformer is energised, and therefore accumulates over 8,760 hours a year regardless of production. Load loss runs the other way: it rises with the square of loading, so a lightly loaded transformer has very little of it. Whether a larger unit costs you or saves you depends entirely on load factor.

IS 1180 (Part 1): 2014 as amended in March 2021 caps maximum total losses at both 50 percent load and 100 percent load for every rating and every efficiency level, across five energy efficiency levels corresponding to 1-star through 5-star labelled transformers as prescribed by the Bureau of Energy Efficiency. A standard that publishes a 50 percent loading figure is telling you part-load running is the normal case.

Take the Level 2 ceilings for two adjacent ratings. A 500 kVA unit is capped at 1,430 W total at half load and 4,100 W at full load. An 800 kVA unit is capped at 2,147 W and 5,838 W. Solve those pairs for their fixed and variable components and compare across a range of average demands:

The crossover sits at roughly 240 kVA, which is 48 percent of the smaller transformer’s rating. Below it the smaller unit wins on energy. Above it the larger unit wins, because its lower current density more than repays the extra core loss. Oversizing is not automatically an energy penalty, and undersizing is not automatically an energy saving. A single-shift plant with long idle periods pays for oversizing. A three-shift plant at steady high load is often better off one rating up.

Which leads to the most useful line in this article. A total-loss figure cannot be converted into an annual energy cost, because you cannot separate the fixed component from the variable one without both numbers. Demand no-load loss and load loss as two separately guaranteed values, with tolerance and reference temperature stated, and reject bids quoting only a total. Capitalisation of losses is meaningless without them.

Why the largest motor decides the rating more often than the load does

The largest direct-on-line motor frequently forces a larger transformer than the steady-state load requires, because starting current produces a voltage dip across the transformer impedance that drops out contactors on unrelated feeders. The working rule New India Electricals Ltd applies when reviewing an LT scheme is that the largest DOL motor should not exceed roughly 15 to 20 percent of transformer kVA. On a 630 kVA transformer that caps the largest DOL start at about 94 to 126 kVA of motor input, which is a 90 kW motor and no more.

A 90 kW motor drawing roughly 110 kVA at full load pulls six to eight times that on a DOL start, and every contactor coil on the board sees the resulting dip at the same instant. IS 1180 fixes impedance by rating: 4.5 percent up to 630 kVA, 5.0 percent from 800 to 1250 kVA, and 6.25 percent from 1600 to 2500 kVA. Higher impedance limits fault current but deepens the starting dip, so the two requirements pull against each other.

Three ways out, and only one of them involves buying a bigger transformer:

• Change the starting method. Star-delta cuts starting current to roughly a third, a soft starter to between 2.5 and 4 times full-load current. Both cut starting torque too, so a loaded conveyor or a positive-displacement pump may stall. A VFD is the only method that reduces current while increasing torque, because it reduces frequency rather than voltage.

• Stagger the starts. Sequencing large drives through the PLC so no two heavy starts coincide removes the problem at no capital cost, where the process tolerates it.

• Go up one rating. Sometimes correct, often the lazy answer, and always the one that adds fixed loss for the life of the asset.

What cold-load pickup does to the diversity you designed for

Diversity is a design-time assumption that fails at the moment the system is most stressed. Cold-load pickup is the loading imposed on a transformer when supply is restored after an outage, recognised as a distinct loading condition in IEEE Std C57.91-2011, Annex F. When power returns, thermostatic loads that had drifted apart are back in phase, motors restart together, and the diversity factor you designed to briefly approaches unity.

This matters far more in Indian industrial estates and in the African markets New India Electricals Ltd supplies than in networks with rare outages. A plant seeing several interruptions a week runs the cold-load pickup case several times a week. Three consequences follow:

• The transformer sees demand well above its design maximum for minutes at a time, which is thermally survivable but electrically disruptive.

• Protection settings tuned to the diversified load nuisance-trip on restoration. The usual field response is to raise the settings until the tripping stops, which quietly removes the protection.

• Restart sequencing is the fix, not a bigger transformer. Staged reclosure through the PLC or a timed restart schedule in the MCC restores diversity artificially at no capital cost.

How ambient temperature and altitude cut the rating you can actually use

IS 2026 (Part 2): 2010 sets normal temperature-rise limits for oil-immersed transformers at 50 degrees C top oil rise and 55 degrees C average winding rise by resistance for naturally cooled and forced non-directed units with Class A insulation. The standard states plainly that these limits are modified from the IEC values as per Indian conditions, and it permits no plus tolerance on them.

Site conditions then bite. IS 2026 (Part 2) requires that site temperature should not exceed 40 degrees C as a monthly average of the hottest month, nor 32 degrees C as a yearly average. If either limit is exceeded, every specified temperature-rise limit is reduced by the same amount as the excess. A site whose hottest month averages 45 degrees C forces a 5 K reduction, paid for either in a physically larger core and winding at the same kVA or in reduced usable output from a standard unit.

Altitude stacks on top. Where the site is above 1,000 m and the test factory is not, permitted average winding rise falls by 1 K per 400 m of excess altitude on a naturally cooled unit, and 1 K per 250 m on a forced-cooled one. A plant on the Zambian Copperbelt at roughly 1,300 m and a plant in Bengaluru are not buying the same transformer at the same kVA.

The reduced limits must be shown on the rating plate, so a buyer can verify at delivery whether the derating was engineered or merely discussed. Indoor installation carries its own penalty: IEC 60076-7 treats a transformer in an enclosure as seeing an extra temperature rise of roughly half the rise of the air inside it, which makes substation room ventilation a sizing input rather than a civil afterthought.

Which specification lines decide whether the transformer fits the load

Six lines carry most of the sizing consequence.

The tap changer line matters most on export projects. On a fluctuating supply an on-load tap changer holds the LT bus inside tolerance without operator intervention, and the minus 21 percent range exists precisely because some networks sag that far. Off-circuit taps need a shutdown to change and suit stable supplies.

Sizing also has to respect the hierarchy above it. On a hot strip and cold rolling mill project in Nepal, New India Electricals Ltd supplied 25/31.5 MVA 66/11 kV power transformers alongside a 4 MVA 11 kV to 433 V unit, with 66 kV control and relay panels and outdoor substation equipment. The distribution transformer rating there was set by the LT load it fed, but the 11 kV bus arrangement above it fixed the fault level it had to survive. Sizing a distribution transformer without the upstream single line diagram in front of you is guesswork.

A sizing checklist to run against any tender

1. Confirm maximum demand from twelve months of billing data, or from a load schedule with stated demand and diversity factors for every group.

2. Convert to kVA at the power factor you will actually run at, after APFC correction, and state the growth allowance as a percentage and a horizon.

3. Check the largest DOL motor against 15 to 20 percent of the candidate kVA, and revise the starting method before revising the transformer.

4. Confirm ambient at the installation location, not outdoor shade temperature, apply the IS 2026 (Part 2) reduction, and add the altitude reduction above 1,000 m.

5. Confirm the derating will be marked on the rating plate.

6. Demand no-load loss and load loss as separate guaranteed figures, with tolerance and reference temperature.

7. State the IS 1180 energy efficiency level and corresponding BEE star rating in the purchase order, and check the secondary voltage against the actual motor voltage on site.

8. Size the LV panel, busbar and incomer on transformer full-load current, not on the load. A 630 kVA unit at 433 V delivers 840 A; a 400 kVA unit delivers 533 A.

9. Confirm the cold-load pickup case and agree a restart sequence before commissioning.

   Frequently asked questions

   How do I calculate the kVA rating of a distribution transformer?

   Divide site maximum demand in kVA by the target loading, then round up to the next standard IS 1180 rating. A measured maximum demand of 286 kVA at a 75 percent target gives 381 kVA, which rounds up to a standard 400 kVA unit. Maximum demand comes from twelve months of DISCOM billing data where the site exists, or from a load schedule with demand factors applied per group and a diversity factor applied at the incomer where it does not. Convert kW to kVA at the power factor you will actually operate at after correction, because moving from 0.85 to 0.98 shifts the answer by more than a full rating step.

   What is the difference between demand factor and diversity factor?

   Demand factor is the ratio of a load group’s maximum demand to its connected load, and it is always less than or equal to one. Diversity factor is the ratio of the sum of individual group maximum demands to the coincident maximum demand at the supply point, and it is always greater than or equal to one. Demand factor accounts for equipment running below nameplate or intermittently; diversity factor accounts for load groups peaking at different times of day. Both reduce the rating you need, and applying only one of them is the most common cause of oversizing on industrial projects.

   Is a bigger transformer always less efficient?

   No. A larger transformer carries higher no-load loss, which runs continuously, but lower load loss at the same delivered kVA because it operates at a lower fraction of rated current. Which effect dominates depends on load factor. Comparing IS 1180 Level 2 loss ceilings, an 800 kVA unit loses more than a 500 kVA unit below roughly 240 kVA average demand and less above it. A single-shift plant with long idle hours pays for oversizing; a continuous-process plant at steady high load often benefits from one rating up. Settling it needs no-load loss and load loss as separate guaranteed figures.

   What does IS 1180 specify for distribution transformers?

   IS 1180 (Part 1): 2014 covers oil-immersed distribution transformers up to and including 2500 kVA, 33 kV. It sets standard kVA ratings, fixes impedance by rating, specifies Dyn11 as the default vector group with Dyn1 permitted as an alternative, and caps maximum total losses at both 50 percent and 100 percent load. Amendment 4 of March 2021 extended it to five energy efficiency levels, Level 1 through Level 5, corresponding to 1-star through 5-star labelled transformers as prescribed by the Bureau of Energy Efficiency. State the required level explicitly in the purchase order.

   How much do high ambient temperature and altitude reduce a transformer’s rating?

   IS 2026 (Part 2): 2010 sets normal limits of 50 degrees C top oil rise and 55 degrees C average winding rise by resistance for naturally cooled oil-immersed units with Class A insulation, referenced to a site whose hottest month averages no more than 40 degrees C and whose yearly average is no more than 32 degrees C. Where either limit is exceeded, every temperature-rise limit is reduced by exactly the amount of the excess. Above 1,000 m the permitted average winding rise falls by a further 1 K per 400 m on a naturally cooled unit, and 1 K per 250 m on a forced-cooled one. Both reductions must be recorded on the rating plate.

   What happens if a distribution transformer is undersized?

   An undersized transformer runs at elevated winding temperature, and insulation ageing accelerates sharply with hot-spot temperature. Under the loading model in IEC 60076-7, adopted in India as IS 2026 (Part 7), the relative ageing rate for non-thermally upgraded paper doubles for roughly every 6 K rise in hot-spot temperature above the 98 degrees C reference. A transformer running 12 K hot ages at about four times the normal rate. It will not trip and it will not alarm. It will fail years early, and the failure is usually blamed on quality rather than on a sizing decision taken years before.
   Speak to New India Electricals Ltd before the rating is fixed rather than after the tender closes. The sizing conversation is short, and it is considerably cheaper than the alternative.

   ^[1][2][3][4]