District Heating & CHP Transformer: How to Size a Heat-Network Substation, Why the Pump Load and the Summer Minimum Set the Rating, and What Makes a CHP Step-Up Transformer Different

The answer first: a district heating substation transformer is sized by its pumps and its summer, not by its heat output

A district heating (DH) transformer almost never carries the heat itself — it carries the pumps, valve actuators, controls, ventilation and lighting that move the heat. That is why a 10 MW heat substation typically needs only a 160–250 kVA transformer, while a 25–50 MW combined heat and power (CHP) generator behind the same network needs a 40–75 MVA step-up transformer. Size the substation unit from connected pump and auxiliary load at 0.85–0.90 power factor with 20–25% spare, then check the summer minimum: when space heating stops and only domestic hot water runs, average loading can fall below 15% of rating and no-load loss becomes 70–85% of the transformer's annual loss energy. Size the CHP step-up unit on generator MVA at rated power factor, then add the two duties that separate it from a wind or solar GSU — heat-led load-following with daily start-stops, and a tap range wide enough to hold the busbar while electrical output swings with heat demand. Fix those four numbers (pump load, summer minimum, MVA, tap range) and the rest of the specification follows from them.

Key takeaways

• Two transformer families, two sizing rules: a network substation transformer is sized on pump and auxiliary kW (typically 100–1,250 kVA at 10 or 20/0.4 kV), a CHP step-up transformer on generator MVA at its rated power factor (40–75 MVA at 110 kV for 25–50 MW units).
• Flow, not heat, sets the electrical load: flow (m³/h) ≈ 0.88 × Q (kW) / ΔT (K) for water at 60–80 °C, and pump shaft power rises with flow cubed in a fixed pipe network — so raising ΔT from 20 K to 40 K cuts pumping energy to roughly one eighth of its original value.
• Low-temperature (4th generation) networks and large heat pumps move the transformer into harmonic territory: VFD-driven pumps and compressor drives with 6-pulse front ends draw 30–45% current THD, which is a transformer heating problem, not just an EMC problem.
• Indoor and basement locations push the choice to cast-resin dry-type to IEC 60076-11 — but the room then has to remove roughly 1,000 m³/h of air per 10 kW of transformer loss for every 10 K of allowable air rise.
• Cost is secondary to loss: at 1,500–2,500 equivalent full-load hours, a 1,000 kVA unit's no-load loss runs 8,760 hours a year while its load loss runs far fewer, so the low-P0 grade (S15/S20 or amorphous) pays back faster in a heat substation than in a continuously loaded industrial plant.

Where do transformers sit in a district heating system? Four duties, four specifications

A district heating scheme is an electrical system wrapped around a hydraulic one: heat is generated at a plant or CHP unit, distributed as pressurised hot water through pre-insulated pipes to EN 13941 design rules, and delivered through substations that transfer heat to each building's own secondary circuit. Transformers appear at four different points in that chain, and treating them as one commodity item is the most common source of both overspend and under-specification.

The first two are power transformers in the conventional sense. The last two are distribution transformers with an unusual load shape: a small, highly variable electrical load whose magnitude is set by hydraulic design rather than by building floor area. That is why a heat substation's transformer looks oversized on paper when the heat output is quoted in MW, and why quoting kVA from MW directly is the first mistake in most DH tenders.

For a network substation, the electrical load is dominated by two circulating pumps (one duty, one standby), a make-up water pump, motorised control valves, a heat meter and control panel, plus room ventilation and lighting. A design that follows the five-step logic in our transformer sizing and budget formula but starts from the wrong load will be wrong at every later step, so it is worth doing the hydraulic calculation first.

How do you size a heat-network substation transformer in five steps?

The sequence is fixed: heat load to flow, flow to pump power, pump power to connected load, connected load to kVA, kVA to a standard rating with the summer minimum checked.

Step 1 — design heat load and temperature difference. Take the substation's design heat load Q (kW) including domestic hot water, and the design primary-side temperature difference ΔT. For water at 60–80 °C, volumetric flow is m³/h ≈ 0.88 × Q / ΔT. A 10 MW substation on a 90/55 °C network (ΔT = 35 K) therefore circulates about 250 m³/h; the same load on a 70/40 °C secondary circuit (ΔT = 30 K) circulates about 290 m³/h.

Step 2 — pump head and shaft power. The duty pump must overcome the substation's own pressure drop (typically 50–150 kPa) plus its share of the network differential pressure, which in an extended network is 600–1,200 kPa. Shaft power is P (kW) = flow (m³/s) × Δp (kPa) / efficiency, with pump and motor efficiency around 0.70–0.80 for a well-selected unit. The 250 m³/h example against 800 kPa at 0.75 efficiency is 74 kW at the shaft.

Step 3 — add the auxiliaries. Make-up water pump 2–5 kW, valve actuators 1–3 kW, control and instrumentation 3–10 kW, ventilation and lighting 3–8 kW, plus any sump pump, pressurisation unit or trace heating. Add the standby pump's motor only if the scheme allows both to run (it normally does not) — but the transformer must still ride through a duty/standby changeover without nuisance tripping.

Step 4 — demand factor and power factor. Apply a demand factor of 0.7–0.85 to the connected load (pumps rarely run at design duty all year) and divide by the operating power factor. VFD-fed pumps with a DC-link run at 0.95–0.98 displacement power factor; DOL motors run at 0.80–0.87. The difference is worth a transformer size step.

Step 5 — round up and check. Choose the next standard rating, verify the LV short-circuit current against switchgear rating using the method in our short-circuit current calculation guide, and check the summer-minimum loading and the noise limit before freezing the specification.

Two sanity checks decide between 125 and 160 kVA in practice. First, the largest motor: a 75 kW DOL-started pump draws 6–7.5 times rated current for 1–3 seconds, and the resulting voltage dip must stay within the limits agreed with the distribution network operator — see inrush current and starting dip. Second, the future connection: most DH substations later acquire a small heat pump, a booster or an EV charger in the plant room, and 25% spare is cheaper than a second transformer.

Why does the summer minimum matter more than the winter peak for operating cost?

In a heating-dominated network the transformer's load follows the heat load, and the heat load follows the weather. A substation that runs at 60–75% of its transformer rating on the coldest week may run at 8–15% in August, when only domestic hot water and distribution losses remain. Because no-load loss is constant for all 8,760 hours while load loss falls with the square of loading, the annual loss energy is dominated by the light-load months.

Take a 1,000 kVA unit with no-load loss P₀ = 1.5 kW and load loss P_k = 11 kW at 75 °C. At an annual average loading of 15%, load-loss energy is 11 kW × 0.15² × 8,760 h ≈ 2,170 kWh against 13,140 kWh of no-load energy — 86% of the total. Even at an average loading of 40%, the split is roughly 64/36 in favour of no-load loss. This inverts the usual industrial buying logic: in a factory the load-loss grade (S11 vs S13 vs S20) dominates the payback, while in a heat substation the no-load grade dominates it.

The commercial consequence is straightforward: for a DH substation, specify the lowest no-load loss grade available in the chosen technology and accept a slightly worse load-loss figure if the price difference demands it. The comparison method is set out in S11 vs S13 vs S20 loss grades, and the whole-life view in transformer total cost of ownership.

There is also an oversizing penalty that appears as a voltage effect rather than a loss effect. A lightly loaded transformer with a fixed tap holds the LV busbar above nominal, because its regulation is only 1–2% at 15% load; if the network's 10 kV busbar is itself at +5% in summer, the 400 V side can sit at 410–415 V and push pump motors and electronics toward their upper voltage limit. Off-load tap changers set for winter can be one step out in summer — see voltage regulation and OLTC versus off-circuit tapping.

What does a real heat-network load profile look like over a year?

District heating load is usually described by a duration curve: the number of hours per year in which the load equals or exceeds a given value. For a northern-European or northern-Chinese city, the curve is steep at the top and long at the bottom — the peak may only occur for 50–200 hours, while 40–50% of annual heat is delivered below 40% of peak load.

Three consequences follow for the electrical design. First, transformer sizing on the 100-hour peak is correct for thermal capability but wrong for efficiency, which is why the no-load argument above matters. Second, pump selection should be based on the duration curve, not on the design point: a pump optimised purely for design flow will run far from its best efficiency point for most of the year, and an electronically controlled (variable speed) pump set will save more energy than any transformer upgrade. Third, the substation's electrical design must survive daily cycling in winter and long idle periods in summer without moisture problems — a transformer that sits at 10% load for months does not dry itself out.

For biomass, waste-to-energy and industrial CHP plants, the electrical side has the opposite profile: the turbine runs at high load for 6,000–8,000 hours a year, and the transformer is closer to base-load duty. The same network therefore contains both a heavily loaded GSU and dozens of lightly loaded substation transformers, and a single procurement specification for both is a compromise that suits neither.

How much does raising the network temperature difference cut electrical load?

Temperature difference is the most powerful and least expensive lever in heat-network electrical design, because it reduces flow and flow is what the pumps and their transformers are sized for. For a fixed heat load, flow is inversely proportional to ΔT: doubling ΔT halves the flow.

Those figures hold the pump head constant, which is the conservative case. In a real network the head is set by pipe friction and rises roughly with the square of flow, so pump power falls with the cube of flow: halving the flow cuts shaft power to about one eighth, not one half. That is the physical basis of the “4th generation district heating” argument for lower supply temperatures — although lower temperatures usually reduce ΔT as well, so the two effects must be evaluated together for each network.

For the transformer, the effect is indirect but real: smaller flows mean smaller pump motors, smaller starting currents and a smaller LV switchboard, which together can move a substation from 250 kVA to 160 kVA. If the network design is still open when the transformer is ordered, it is worth calculating both cases before fixing the rating.

When should a heat-network substation use a dry-type transformer instead of an oil-immersed one?

Most DH substations sit in basements, plant rooms attached to residential blocks, or buried chambers under pavements — locations where a mineral-oil transformer introduces a fire load, an oil-containment requirement and an insurance discussion. Cast-resin dry-type transformers to IEC 60076-11 remove the liquid, reduce the fire load to that of the resin and insulation system, and are the default choice for indoor building substations in most European and Asian markets. The trade-offs are real and worth stating plainly.

The choice is usually settled by the building code rather than by engineering preference: if the substation is inside a residential or commercial building, dry-type is often the only option that avoids a fire-rated compartment. Where the substation is a standalone chamber with outdoor access, a hermetically sealed oil unit is cheaper, more efficient at large ratings and easier to keep dry — the sealed tank removes the breather and the oil's moisture path entirely. The comparison of sealing systems is covered in hermetically sealed versus conservator, and the liquid alternatives in mineral versus ester versus silicone oil.

What does an indoor dry-type transformer need from the plant room?

A dry-type transformer converts its losses into room air, so the room is part of the thermal design. A 1,000 kVA cast-resin unit with 1.5 kW no-load and 11 kW load loss dissipates roughly 8–10 kW at typical DH loading; removing that heat with a 10 K air rise needs about 3,000 m³/h of ventilation air, and with a 15 K rise about 2,000 m³/h. Undersized ventilation derates the transformer silently: winding temperature rises, insulation ages faster, and the unit reaches its alarm threshold on the hottest week of the year rather than at commissioning.

Three details are worth writing into the specification. Specify the winding temperature rise at the intended cooling class (ΔT 100 K average for a naturally cooled Class F unit is common; forced cooling buys 25–40% more capacity). Specify IP protection and the enclosure type separately from the transformer, because a cast-resin core is IP00 and only the enclosure provides IP31 or IP54. And specify anti-condensation heaters with a humidistat for chambers below grade: cast resin is hygroscopic at the surface, and a cold, idle transformer in a damp chamber is where partial discharge starts.

What noise limits apply when the transformer sits inside a residential building?

Noise is often the constraint that decides between two otherwise equal offers. Environmental limits for building services in residential areas are commonly 45–55 dB(A) at the facade by day and 40–45 dB(A) at night, measured at the nearest dwelling; a 1,000 kVA dry-type transformer at 58–63 dB(A) at 1 m cannot meet that without attenuation. Oil units of the same rating are typically 3–5 dB(A) quieter because the tank walls and oil absorb and damp core vibration.

Values are typical ranges from manufacturer data at rated voltage and no load, measured to IEC 60076-10; the guaranteed figure must be a type-tested sound power or sound pressure level stated in the contract, not a catalogue approximation. Mitigation follows the physics: resilient mounts to break structure-borne paths, flexible connections on busbars and cables, an acoustic enclosure with a rated insertion loss, and distance — every doubling of distance reduces free-field sound pressure by about 6 dB(A), while inside a plant room the reduction is closer to 3–5 dB(A). The engineering detail is in transformer noise level.

How is a CHP step-up transformer different from a wind or solar GSU?

The hardware looks similar — a generator transformer at 10–15 kV on the LV side and 110–220 kV on the HV side — but the duty is not. A renewable GSU follows an inverter that is usually operated at unity power factor and at a resource-driven output; a CHP GSU follows a synchronous generator whose output is set by heat demand, and that difference propagates into five specification items.

Heat-led operation means the electrical output is a by-product. In an extraction-condensing unit, electrical output falls as extraction steam rises: on a cold morning the machine may make 30 MW of heat and 12 MW of electricity, and on a mild afternoon 5 MW of heat and 28 MW of electricity. The GSU must therefore be rated for the maximum electrical case and specified for a load profile that changes daily, not seasonally.

Start-stop count is high. A heat-led CHP unit in a municipal network may start and stop once or twice a day through the heating season and run continuously in mid-winter — 200–400 starts a year is not unusual. Thermal cycling is a loading-guide problem under IEC 60076-7 and load factor and overload capability, and it is also a mechanical problem for the tap changer and bushings.

Tap range is wider than a renewable GSU's. Because the HV busbar must be held while the machine's reactive output and the network's voltage swing, a CHP GSU commonly needs ±8 × 1.25% or ±10 × 1.5% on-load tapping, compared with ±8 × 1.25% or a fixed ratio for a solar unit tied to a stiff busbar. The tap changer normally sits at the HV neutral end, where insulation is cheapest.

Load rejection is the design overvoltage case. A generator breaker opening at full load leaves the GSU carrying the station's auxiliary load through a rapidly collapsing voltage, and the HV bus can rise 15–30% before excitation control responds. Over-flux capability of 1.10 pu continuous and 1.25 pu for one minute, plus HV surge arresters and an over-excitation limiter, belong in the specification.

Auxiliary supply comes from the GSU, not the grid. Unit auxiliaries — feed pumps, fans, controls — are usually tapped from the GSU LV side or from a separate unit auxiliary transformer, so the GSU must carry the auxiliary load when the generator is offline but the plant is still running. That case, not the generator rating, often sets the LV winding's continuous current in a black-start-capable plant.

What is the power-to-heat ratio and why does it set the CHP transformer rating?

The power-to-heat ratio (P/H) is the electrical output per unit of useful heat, and it determines how much generator MVA sits behind a given heat load. Back-pressure turbines have the lowest ratio because all steam exhausts to the network; extraction-condensing machines can swing between heat and power; gas engines and gas turbines sit at the other end.

Ratios are typical operating values from municipal and industrial CHP practice, not catalogue constants; the actual figure comes from the turbine heat balance at the design point. The practical rule is to obtain the heat balance for three cases — design heat, design power and minimum stable load — and to size the GSU on the highest generator MVA of the three, then add 5–10% for voltage and cooling margin. A GSU sized only on the nameplate electrical rating will be undersized in the condensing mode that the plant uses in summer, when heat demand disappears and electricity prices are highest.

How should a CHP step-up transformer be specified for heat-led operation?

A specification that a manufacturer can build without interpretation has eight lines in it. The voltage ratio and vector group (YNd11 is the default above 60 MVA; Dyn11 for smaller units), the rated MVA at a stated cooling class, the impedance with a tolerance band, the tap range and tap-changer type, the over-flux capability, the short-circuit withstand, the sound level and the test list.

Impedance of 10–14% is the usual band for a 25–75 MVA GSU at 110–220 kV. Lower impedance improves voltage regulation and reactive capability but raises the HV fault level and the cost of the HV circuit breaker; higher impedance reduces fault duty but increases voltage drop and reactive consumption under load. The value should come from the grid study, not from a preferred factory standard — and it must be stated with the standard tolerance (±7.5% of the declared value on the principal tapping, or ±10% where the grid code allows).

Cooling follows the rating split: ONAN for the base rating, ONAF for the forced rating, and OFAF or ODWF only where the site cannot accommodate the radiator area or the noise limit demands lower fan speed. A 75 MVA unit with ONAN/ONAF at 67/75 MVA gives the plant a continuous 67 MVA capability with all fans lost — a useful redundancy for a heat-led machine that must keep running in a cold snap. Cooling classes, ducting and fan-loss penalties are covered in transformer cooling methods and temperature rise.

Finally, specify the condition-monitoring set: two Pt100 winding sensors per phase plus top-oil and bottom-oil sensors, a Buchholz relay and pressure relief device on conservator units, an oil-level indicator, and a DGA sample point. Baseline DGA at the factory and again at commissioning gives the reference against which later results are judged — the interpretation method is in dissolved gas analysis, and the energisation checks in commissioning tests.

What changes at a large heat pump or electric boiler site?

Large electrically driven heat pumps are now the fastest-growing load on northern district heating networks, because they convert cheap or surplus electricity into heat at a coefficient of performance (COP) of 3–4. A 10 MW heat pump therefore draws 2.5–3.3 MW of electricity, which is a different order of magnitude from the pumps in a conventional substation and changes the transformer's role from auxiliary supply to main process supply.

Three consequences dominate. The compressor is VFD-driven, so the transformer sees a harmonic spectrum with a 6-pulse front end (35–45% current THD) or a 12-pulse or active front-end configuration (5–12%). The drive starts and stops with heat demand, so the transformer is cycled rather than base-loaded. And the heat pump's electrical load is seasonal in the opposite direction to the network's: it runs hardest in cold weather, when the network is also at peak, which raises the substation's maximum demand exactly when the CHP unit is already at full output.

Sizing follows the same logic as a motor load but with the compressor's service factor included: a 10 MW heat pump at COP 3.2 needs about 3.1 MW of shaft power, drawing 3.3–3.5 MVA at 0.90–0.95 pf through the drive, so a 4–5 MVA transformer at 6.6 or 10 kV is the usual answer, with the harmonic loss allowance built in. Electric boilers are simpler and larger: a 20 MW electrode boiler draws 20 MW at unity power factor and is usually fed at 10–33 kV through a dedicated 25 MVA transformer, with the electrode boiler's own step control providing the load management.

What harmonic limits apply, and how do you specify a transformer for VFD pumps?

Harmonics matter in DH substations for two reasons: the transformer heats up more than the RMS current suggests, and the distribution network operator has a contractual limit at the point of connection. The limits are usually taken from IEEE 519 at the PCC (voltage THD 5% and individual harmonics 3% below 69 kV, with current limits scaled by the short-circuit ratio) or from IEC 61000-3-12 for small installations and IEC 61800-3 for the drive's own emission class.

For the transformer itself, the relevant question is how much extra loss and heating the harmonic spectrum causes. A K-factor or factor-K rating (defined in UL 1561 and in IEC 60076-11 for dry-type units) expresses the transformer's ability to supply non-sinusoidal load without exceeding its rated temperature rise; K-4, K-9 and K-13 are the common steps. In practice, a DH substation with one or two 6-pulse VFD pumps reaching 35–40% current THD needs K-9 or better, or a transformer derated to about 85% of its nameplate. The physical mechanisms are set out in transformer harmonics.

There are cheaper fixes than a K-rated transformer, and they should be evaluated first: specify an active front end or a 12-pulse drive with a phase-shifting transformer, add a passive filter or a line reactor, or use several small drives with staggered pulses instead of one large one. Where the harmonic spectrum is known from the drive vendor, ask for a harmonic loss calculation rather than a K-factor label — the calculation is what the factory needs in order to guarantee the temperature rise.

How do you protect a transformer in a buried or humid substation chamber?

The environment, not the load, destroys most DH substation transformers. Chambers below grade are damp, poorly ventilated and occasionally flooded; coastal networks add chloride; and a transformer that idles at 10% load for eight months never gets warm enough to drive moisture out of its insulation.

Four measures cover most of the risk. Specify a sealed oil unit (hermetically sealed tank with no breather, or a conservator with a membrane-type conservator) so that the oil never contacts ambient air. Specify anti-condensation heaters and a humidistat on dry-type units, plus IP54 enclosures with cable entry from below and sealed glands. Specify the corrosion class explicitly — C4 or C5 per ISO 12944 for coastal and buried chambers, with a stated coating system and dry-film thickness. And specify a sump pump with alarm, plus a flood-level mark on the tank, so that a chamber flood is detected before the cable boxes are submerged.

Protection settings follow the same principle as any distribution transformer — overcurrent and earth-fault on the HV side, temperature trip on the LV side — but the DH case adds two specifics: the transformer must ride through pump changeover and motor re-acceleration without tripping on inrush, and the temperature trip should be set from the winding sensor rather than the top-oil sensor, because a lightly loaded unit's oil stays cool while the winding of a VFD-fed load can run hot. Device selection is covered in transformer protection devices.

What tests and documents should you require before shipment?

Routine tests for every unit: winding resistance on all taps, voltage ratio and vector group, no-load loss and current, load loss and impedance at 75 °C reference, applied and induced overvoltage, insulation resistance, and partial discharge for units above the relevant voltage class. For dry-type units to IEC 60076-11, the partial discharge measurement is a routine test and its limit should be stated in the contract.

Type tests, if not already certified for the design: temperature rise at the declared cooling class, lightning impulse, and short-circuit withstand. For a CHP GSU the short-circuit type test is worth insisting on, because the generator-fed fault level is set by the machine's subtransient reactance and is not limited by the grid. Special tests where the duty justifies them: sound level to IEC 60076-10, harmonic loss measurement, frequency response analysis (FRA) as a transport baseline, and a dissipation-factor and capacitance measurement on bushings to IEC 60137.

The document set matters as much as the tests: routine test certificate, type-test reports, general arrangement drawing with weights and lifting points, nameplate data to IEC 60076-1, tap-changer instructions, oil test certificate for oil units, and the factory's loading calculation. Reading the nameplate against the order is the last line of defence — the fields and their meaning are explained in transformer nameplate.

What does a district heating transformer cost?

Indicative 2026 FOB ranges from Chinese manufacturing, based on QDTB project experience. They assume standard loss grades, copper windings, Dyn11, off-circuit tapping, and exclude freight, duties, switchgear and installation. Prices move with copper and grain-oriented steel, so treat them as order-of-magnitude figures for budgeting rather than quotations.

For a like-for-like comparison, add the cost of losses over the intended life. The method in total cost of ownership typically shifts the choice one loss grade upward for DH substations, and the per-kVA structure of distribution transformer pricing is broken down in distribution transformer price breakdown. For a first-order estimate of the whole substation, the power system calculators take a heat or electrical load through to a transformer rating and an indicative budget.

What are the five specification mistakes most common in district heating substations?

1. Sizing the transformer from the heat rating. A 10 MW substation does not need a 10 MVA transformer; it needs 100–250 kVA for its pumps and auxiliaries. Quoting kVA from MW overstates the requirement by a factor of forty.

2. Ignoring the summer minimum. The transformer is specified for the winter peak and then operated at 8–15% load for eight months. The result is an avoidable no-load loss bill and a busbar that sits above nominal voltage all summer.

3. Choosing dry-type by default without checking the room. Cast resin solves the fire-load problem and creates a ventilation problem. A basement room with no louvre area and no mechanical extract will derate the unit silently; the ventilation calculation belongs in the same document as the transformer rating.

4. Treating VFD harmonics as an EMC issue only. Harmonic current heats the transformer, its neutral and its cables. A 6-pulse pump set at 40% THD on a standard transformer is a temperature-rise problem that appears as a nuisance trip two summers later.

5. Specifying the CHP GSU on the nameplate rating alone. The heat balance, the start count and the tap range determine the real duty. A GSU with ±2 × 2.5% tapping on a heat-led unit will run out of tap in the first winter and be operated at a voltage the machine was not designed for.

Frequently asked questions

Can a standard oil-immersed distribution transformer be installed in a basement?

Usually not without a fire-rated compartment, an oil-retention pit and the approval of the local fire authority. Most building codes treat a mineral-oil transformer above a few hundred kVA as a fire load that must be separated from escape routes and dwellings, which is why cast-resin dry-type units dominate indoor DH substations. Where the room is a dedicated, fire-separated chamber with outdoor access, a hermetically sealed oil unit is acceptable in many jurisdictions and is cheaper and quieter.

10 kV or 20 kV for a heat network substation?

20 kV halves the current for the same power, which halves the cable loss and roughly halves the voltage drop over a long network; it also raises the transformer's insulation level and the switchgear cost. For networks with substations more than 3–5 km from the primary station, or with a large number of small substations, 20 kV is usually the better whole-life choice; for compact city-centre networks at 10 kV, staying at 10 kV is straightforward. The transformer itself is not the deciding cost — the cable and switchgear are.

Does a heat-led CHP need a larger transformer than a power-led one?

At the same heat output, a heat-led machine usually needs a smaller transformer, because its electrical output is lower by design. The transformer becomes larger when the plant is specified to export power in summer condensing mode, which is the case that sets the maximum generator MVA and therefore the GSU rating. Specify the GSU for the highest MVA the machine can produce in any mode, not for the heating-season average.

How much spare capacity should a DH substation transformer have?

20–25% above the calculated demand is the usual allowance, and it should be justified by a named future load — a third pump, a booster, a small heat pump or an EV charger in the plant room. Beyond about 40% spare, the no-load loss of the oversized unit costs more than a second transformer would, and the summer voltage rise becomes a nuisance.

Is an amorphous-core transformer worth it in a heat substation?

More often than in a factory, because no-load loss dominates the annual loss energy at the low average loading typical of DH substations. An amorphous core cuts no-load loss by 60–75% compared with a conventional grain-oriented core, which in a 1,000 kVA unit at 15% average loading can remove 8,000–10,000 kWh a year. The premium is 15–25% of the transformer price, and the payback in a high-tariff market is typically 3–6 years. The comparison method is in amorphous core transformers.

Sources / 资料来源

Source: IEC 60076-1 — Power transformers, Part 1: General.
Source: IEC 60076-2 — Power transformers, Part 2: Temperature rise for liquid-immersed transformers.
Source: IEC 60076-3 — Power transformers, Part 3: Insulation levels, dielectric tests and external clearances in air.
Source: IEC 60076-5 — Power transformers, Part 5: Ability to withstand short circuit.
Source: IEC 60076-7 — Power transformers, Part 7: Loading guide for mineral-oil-immersed power transformers.
Source: IEC 60076-10 — Power transformers, Part 10: Determination of sound levels.
Source: IEC 60076-11 — Power transformers, Part 11: Dry-type transformers.
Source: IEC 60076-18 — Power transformers, Part 18: Measurement of frequency response.
Source: IEC 60214-1 — Tap-changers, Part 1: Performance requirements and test methods.
Source: IEC 60137 — Insulated bushings for alternating voltages above 1 000 V.
Source: IEC 60034-1 — Rotating electrical machines, Part 1: Rating and performance.
Source: IEC 61800-3 — Adjustable speed electrical power drive systems, Part 3: EMC requirements.
Source: IEC 61000-3-12 — Limits for harmonic currents produced by equipment connected to public low-voltage systems.
Source: IEEE 519-2022 — Recommended practice and requirements for harmonic control in electric power systems.
Source: IEEE C57.91 — Guide for loading mineral-oil-immersed transformers and step-voltage regulators.
Source: IEEE C57.116 — Guide for transformers directly connected to generators.
Source: EN 13941-1 — District heating pipes: design and installation of thermal insulated bonded single and twin pipe systems for directly buried hot water networks.
Source: EN 50464-1 / EN 50588-1 — Three-phase oil-immersed distribution transformers, 50 Hz, with highest voltage for equipment not exceeding 36 kV.
Source: ISO 12944 — Paints and varnishes: corrosion protection of steel structures by protective paint systems.
Source: UL 1561 — Dry-type general purpose and power transformers (K-factor rating).
Source: GB 20052-2020 — Minimum allowable values of energy efficiency and energy efficiency grades for power transformers.
Source: GB/T 6451-2015 — Technical parameters and requirements for three-phase oil-immersed power transformers.
Source: Lund et al., “4th Generation District Heating (4GDH): Integrating smart thermal grids into future sustainable energy systems”, Energy 68, 2014.
Source: Euroheat & Power — country statistics on district heating market shares.
Source: QDTB engineering experience — heat substation pump and auxiliary load tables, CHP power-to-heat ratios, typical sound pressure ranges, dry-type ventilation practice, and indicative FOB price bands. Values described as typical or indicative are ranges from district heating and CHP projects, not published standards.