Energy-efficient transformers: annual savings of a 1000 kVA unit

Choosing energy-efficient transformers is a cost decision rather than a marketing one: the losses of a distribution transformer run for 8 760 hours a year over a service life that is typically 25 years or more, and every kilowatt-hour is paid by the substation owner. This article works through a 1000 kVA, 10/0.4 kV, 50 Hz example typical of distribution networks in Uzbekistan, comparing the S11 level with S13, the energy-efficiency grade 1 S22 (listed on this site as S-M-NX1) and the amorphous-alloy SH15, with every assumption stated so that you can substitute your own figures.

Where transformer losses come from

A transformer has two loss components, and they behave differently:

• No-load loss (P0) is the core loss from hysteresis and eddy currents. It is present whenever the HV side is energised, whatever the load. Over a year it is simply P0 × 8 760 h.
• Load loss (Pk) is the I²R and stray loss in the windings, guaranteed at rated current and corrected to the reference temperature (75 °C for oil-immersed units). It scales with the square of the load: at 50 % load it is a quarter of the rated value.

Annual energy loss is therefore:

E = (P0 + β² × Pk) × 8 760 h, where β is the load factor.

Note that β must be the RMS (root-mean-square) load factor of the load profile, not the arithmetic mean: a load that swings between 20 % and 80 % loses more than a steady 50 % load. The two components are also reduced by different means: a better core (higher-grade grain-oriented steel or amorphous ribbon) cuts no-load loss, while more conductor in the windings cuts load loss, at a cost in weight and money.

Loss values for a 1000 kVA, 10/0.4 kV unit

The product tables on this site for the S13 oil-immersed distribution transformer and the SH15 amorphous-alloy transformer give the following values at 1000 kVA (vector group Dyn11, 10 kV / 0.4 kV, short-circuit impedance 4.5 %). The S11 row is the catalogue reference value commonly quoted for the S11 level under GB/T 6451; it is not taken from this site. The grade 1 row is an assumption, explained below.

The S-M-NX1 grade 1 oil-immersed transformer is designed to energy-efficiency grade 1 of GB 20052-2020, but its page does not yet publish a loss table, so round figures in the range typical of grade 1 designs at this rating are used purely to illustrate the method; the guaranteed P0 and Pk stated in the offer and confirmed in the routine test report must replace them. The pattern is what matters: S13 and SH15 keep the S11 load loss and reduce only the no-load loss (by 28 % and 61 % respectively), whereas a grade 1 design reduces both components.

Worked example: annual savings of energy-efficient transformers

Assumptions: energised 8 760 h/year, RMS load factor β = 0.5, electricity at an assumed 0.06 USD/kWh. Actual tariffs in Uzbekistan depend on the consumer category and change over time; use the rate in your own supply contract.

• S11: (1 150 + 0.25 × 10 300) W = 3 725 W, or about 32 600 kWh/year.
• S13: 3 405 W, about 29 800 kWh/year. Saving versus S11: 2 800 kWh, about 170 USD/year.
• SH15: 3 025 W, about 26 500 kWh/year. Saving: 6 100 kWh, about 370 USD/year.
• S22 / S-M-NX1 (assumed values): 2 665 W, about 23 300 kWh/year. Saving: 9 300 kWh, about 560 USD/year.

The S13 and SH15 saving does not depend on the load: it is (1 150 W − P0) × 8 760 h whatever the load profile. The grade 1 saving does, because part of it comes from the windings. At β = 0.3 it falls to about 6 200 kWh (370 USD/year); at β = 0.7 it rises to about 13 900 kWh (830 USD/year).

Payback logic without quoting prices

The premium over S11 depends on rating, execution and the copper price at the time of the order, so no prices are quoted here. The arithmetic is simple: payback = price premium ÷ annual saving. At the assumed tariff and β = 0.5, every 1 000 USD of premium pays back in:

• about 6 years for S13;
• about 2.7 years for SH15;
• about 1.8 years for the grade 1 unit (assumed loss values).

Against a 25-year service life, a payback under five years is a comfortable case. The same arithmetic also shows when a higher grade is not justified, for example a standby transformer energised for a few weeks a year.

For a tender, the loss-capitalisation (total owning cost) method described in IEC 60076-20 is more robust: each bid is compared as TOC = price + A × P0 + B × Pk, where A and B are the value of one watt of no-load and load loss over the evaluation period. With 0.06 USD/kWh, 8 760 h and a 10-year horizon without discounting, A = 5.3 USD/W and B = A × β² = 1.3 USD/W at β = 0.5. Stating A and B in the enquiry lets each supplier optimise the design against your economics.

When amorphous makes sense

An amorphous core brings the no-load loss down to roughly 40 % of the S11 figure (450 W against 1 150 W at 1000 kVA), and the SH15 no-load current is half that of S13 (0.3 % against 0.6 %). Load loss is unchanged. Amorphous therefore wins where the transformer is energised most of the time but loaded lightly:

• PV plants. A PV step-up transformer is energised around the clock but loaded only in daylight; at night it sits at no-load. With the grade 1 values assumed above, amorphous and grade 1 give the same annual saving at β ≈ 0.3, and amorphous wins below that, which is typical for PV.
• Rural and seasonal loads such as irrigation pumping, cotton ginning or holiday-season retail in the regions of Uzbekistan, where the RMS load factor is low.
• Distribution box-type substations (KTP) sized for future growth, which run for years at a fraction of rated load.

Amorphous is not the first choice for heavily loaded industrial units: above the crossover load factor (about 0.3 with the values assumed here) the load-loss term dominates, and a grade 1 steel-core unit, which also reduces Pk, saves more. Two practical points: amorphous units are larger and heavier for the same rating, which must be checked against the transformer compartment of a box-type substation (KTP), and the ribbon is sensitive to mechanical stress, so handling follows the manufacturer's instructions.

Dry-type units follow the same logic

For indoor substations and PV inverter rooms the choice is between the SC(B)-NX1 grade 1 dry-type transformer (SCB18 level, GB 20052-2020 grade 1) and the SC(B)H15 amorphous-alloy dry-type transformer. The formula is identical; three details differ:

• Dry-type load losses are guaranteed at a higher reference temperature (120 °C for class F insulation, 145 °C for class H, against 75 °C for oil-immersed units), so compare offers at the same reference.
• For SC(B)H15 the site states a no-load loss 75 % below the GB/T 10228 reference value and a sound level 5 to 15 dB below the limit of the applicable noise standard (JB/T 10088), which matters in buildings.
• Forced-air cooling (AF) adds fan consumption; count it only for the hours the fans actually run.

Checklist for specifying an energy-efficient transformer

1. Obtain the guaranteed P0 and Pk from the offer, at the same reference temperature for every bid, and ask whether they are guaranteed maxima: IEC 60076-1 (GB/T 1094.1) allows a tolerance of +15 % on each component and +10 % on total losses unless a tighter figure is agreed.
2. Estimate the RMS load factor from the real load profile, not the installed capacity.
3. Calculate E = (P0 + β² × Pk) × 8 760 h and multiply by your contract tariff.
4. Divide the price premium by the annual saving, or state A and B factors in the tender.
5. Light or intermittent load and PV: amorphous (SH15, SC(B)H15). Steady heavy load: grade 1 (S-M-NX1, SC(B)-NX1). Standby duty: S11 or S13.
6. Ask for the routine test report with measured losses and the type-test certificates (see the certificates page).

Send us the rating, voltage combination, expected load profile and installation type (box-type substation, indoor room or PV field) and we will return a quotation with guaranteed loss values for the grades you want to compare. Use the contact page to request it.