Wastewater Treatment Plant Transformer: How to Size a 120,000 m³/d Plant, Why Aeration Blowers and H2S Decide the Specification

The answer first: size a wastewater plant's transformer from connected load × 0.62, then let aeration harmonics and H2S rewrite the specification

A 120,000 m³/day municipal wastewater treatment plant with 3,198 kW of connected load does not need 3,198 kW of transformer capacity. Apply a demand factor of 0.60–0.70 — 0.62 for a plant of this size — and the peak falls to 1,983 kW, which at 0.90 power factor is 2,203 kVA. That lands on two 2,500 kVA 11/0.415 kV units in a main-tie-main arrangement: 44% loaded in normal operation, 88% on loss of one unit, and 53.5 kA of fault current at each LV board so ordinary 65 kA switchgear works. Two loads then override the arithmetic. Aeration blowers are 36% of peak demand and, when driven by 6-pulse VFDs, inject 35–40% current THD that removes 10% of transformer capacity under IEEE C57.110 — eating exactly the margin your N+1 contingency depends on. And the substation sits in a wet, hydrogen-sulphide-laden, occasionally flooded environment where IP rating, corrosion class and hazardous-area distance matter more than kVA. Get those two right and the transformer runs 20 years at 1% of plant energy; get them wrong and you replace it in six.

Key takeaways

• Size on demand, not nameplate: 3,198 kW connected → 0.62 → 1,983 kW → 2,203 kVA at pf 0.90. Summing nameplates would have specified 3,550 kVA and bought 60% too much iron.
• Aeration is the whole design: 2 duty × 355 kW blowers = 710 kW = 36% of peak. Their VFD front ends decide K-factor, harmonic derating and whether you need 0.415 kV or 6.6 kV distribution.
• Two 2,500 kVA units give 44% normal loading and 88% single-unit contingency. Keep both energised and the tie open: total loss 13.7 kW. Collapsing to one unit to "save no-load loss" raises it to 19.7 kW, i.e. 52 MWh and about USD 5,200 a year thrown away.
• Harmonic derating is the hidden capacity loss: at 35% THDi a K-13 unit delivers only 90% of nameplate, so 2,250 kVA against a 2,203 kVA demand leaves 2% margin. Either specify 18-pulse or active front ends (THDi <5%, derating 0.98) or buy the next frame size.
• Losses are 87% of purchase price over life: capitalised at USD 6,300/kW no-load and USD 2,700/kW load, the two units carry USD 55,800 of evaluated loss against USD 64,000 of purchase price. S11 saves USD 12,800 up front and costs USD 12,800 more in 20-year total cost of ownership.

What electrical load does a 120,000 m³/day treatment plant actually present?

Wastewater treatment is a low-voltage, high-hours business: dozens of small drives, three or four large ones, and a load that never stops. The table below is a real design basis for a 450,000 population-equivalent plant using conventional activated sludge with tertiary disc filtration and anaerobic digestion. Two columns matter more than the rest — installed kW tells you what to count, the control column tells you what the transformer will actually see.

Note what is absent: no single load above 400 kW, no 6.6 kV or 11 kV motors in the base case, and no large rectifier. That is why a wastewater plant's transformer problem is not short-circuit strength or through-fault duty — it is harmonics, corrosion and redundancy. It is also why the plants are cheap to electrify and expensive to electrify well.

How do you size the transformer when flow, and therefore load, changes every hour?

The calculation is one line: design kVA = (Σ connected kW × demand factor) / power factor. The judgement sits entirely in the demand factor. A plant's peak arrives during the morning flow peak with aeration at high output, sludge dewatering running and tertiary backwash in progress — typically 1.5–1.7 times the daily average. Demand factor falls as plant size rises because diversity improves: a small plant may have three loads coinciding, a large one almost never has all fourteen.

For the 120,000 m³/day case: 3,198 kW × 0.62 = 1,983 kW; at 0.90 power factor (VFDs correct displacement factor to near unity at the front end) that is 2,203 kVA. Measured plants of this size run 0.60–0.75 load factor on the transformer, so a 2,500 kVA unit would sit at 88% — too close for a station that cannot be shut down. Hence two units. The same arithmetic across the usual size range:

The demand factor is not a fudge — it is the diversity you can defend in a design review. If the client's engineer insists on 0.85 "to be safe", ask which loads actually coincide: the answer is blowers plus one influent pump plus dewatering, and the rest is diversity. Oversizing by 35% is not conservative, it is a permanent efficiency penalty, because a 3,500 kVA unit idling at 60% load burns more no-load loss than the correctly sized one burns in total. The mechanics of that penalty are set out in how much you can actually load a transformer.

Why do aeration blowers decide the transformer specification?

Aeration consumes 45–60% of a treatment plant's electricity, and in this design 710 kW of the 1,983 kW peak — 36% — passes through two blower drives. Dissolved-oxygen control means those drives run at 55–75% speed most of the day and swing between 40% and 100% as influent ammonia load changes. That load profile is precisely what a VFD is for, and precisely what makes the transformer's life difficult: the blower is not a linear load, it is a current-source of 5th and 7th harmonics.

The energy argument is settled: a VFD at 60% flow draws about 55% of rated power against 85% for throttling, so on 710 kW of duty blowers the difference is roughly 210 kW continuous — about 1,800 MWh a year, or USD 180,000 at USD 0.10/kWh. No transformer specification choice comes close to that number. This is why the correct sequence is choose the drive first, then size the transformer around the drive's harmonics, not the other way round. The underlying mechanism, including why triplen harmonics heat a delta winding without appearing in the line current, is covered in transformer harmonics.

How much transformer capacity do VFD harmonics actually take away?

IEEE C57.110 defines the method: harmonic currents increase eddy-current loss in the windings, raise the hot-spot temperature, and therefore reduce the kVA a given frame can deliver at rated temperature rise. K-factor is the shorthand — K-4 for modest non-linear load, K-13 for 6-pulse drives, K-20 for heavy rectifier load. The practical consequence on a 2,500 kVA unit is not academic:

Read the last two columns together and the trap appears. The N+1 case — one transformer carrying the whole 2,203 kVA — was designed with 12% headroom at 88% loading. At 35% THDi the usable capacity drops to 2,250 kVA, so the contingency case is now 98% of derated capacity, and the hot-spot limit arrives before the nameplate limit. Two honest ways out: specify 5% line reactors plus 18-pulse drives and land at 8% THDi (0.96 derating, 2,400 kVA usable), or move to 3,150 kVA frames and accept the extra no-load loss. What you must not do is take the K-13 label as a substitute for the calculation — a K-13 rating states the harmonic capability at rated load, not at 110% emergency load. Selection logic and worked derating examples are in K-factor transformer selection.

Do not add power-factor correction capacitors on a VFD-fed wastewater board. The drive's front end already holds displacement factor near 0.97, and a capacitor bank will resonate with the transformer and line reactance somewhere between the 5th and 7th harmonic — typically at 6.2–6.8 times fundamental on a 2,500 kVA / 6.5% unit — amplifying exactly the harmonic you were trying to reduce. If the utility penalises reactive power, use a detuned reactor or an active filter, not plain capacitors.

Should the aeration blowers run at 0.415 kV or 6.6 kV?

At 355 kW the answer is genuinely marginal, and it changes the transformer count. On 0.415 kV each blower draws 565 A and needs two parallel 3 × 240 mm² copper feeders; on 6.6 kV it draws 35.5 A and one 70 mm² cable does the job — a 16-fold current reduction. But you then need a second transformation stage, a 6.6 kV switchboard and MV drives.

The decision rule that survives site reality: below 400 kW and within 150 m of the substation, stay on 0.415 kV — two transformer stages cost more in no-load loss and maintenance than the cable saves. Above 400 kW, or when four or more drives above 250 kW cluster at one end of a large site, the 6.6 kV bus pays for itself through cable and switchgear, and it takes 700 kW of load off the LV board so the main transformers can drop a frame size. Voltage-level trade-offs in general are worked through in choosing the right voltage class.

Two transformers at 50% or one at 100%? What N+1 really costs

A treatment plant that discharges to a river under a permit cannot stop aerating. That single sentence dictates redundancy, and the redundancy option you pick changes both capital cost and annual loss. All four options below are technically workable; only one is defensible for a 120,000 m³/day plant.

Now the operating detail almost everyone gets wrong. With the tie open and each transformer feeding its own board, both units stay energised and share the load at 44%, giving 5.2 kW of no-load loss plus 8.5 kW of load loss — 13.7 kW total. Operators sometimes de-energise one unit "to save no-load loss", which loads the survivor to 88%: 2.6 kW no-load plus 17.1 kW load loss = 19.7 kW. The 2.6 kW saved is outweighed by 8.5 kW of extra copper loss, a net waste of 5.9 kW, or 52 MWh and about USD 5,200 a year. Keep both energised; switch the lead unit monthly so hours stay balanced. If instead you close the tie and run them genuinely in parallel, fault current doubles to 107 kA and 125 kA switchgear becomes mandatory — the conditions that make parallel operation safe, and the circulating-current penalty when they are not met, are set out in transformer parallel operation.

What happens to the LV board when a 355 kW blower starts direct-on-line?

Nothing good. A 355 kW, 4-pole motor at 0.415 kV draws 565 A at full load and about 3,390 A at 6 × inrush. Against a 2,500 kVA transformer with 6.5% impedance — 3,478 A rated, 53.5 kA short-circuit — that inrush produces a voltage dip of roughly 6.3% at the board. Motors tolerate it; the UV reactor's electronic ballasts and the SCADA PLC do not, and the plant's own contactors may drop out.

The engineering point is that a VFD solves starting, harmonics and energy in one purchase, which is why the transformer specification should be written after the drive schedule is fixed. One caution that catches commissioning teams: energising the standby transformer after months de-energised produces an inrush of 8–12 times rated current for up to 100 ms, with a second harmonic component that can trip a differential relay set too sensitively. The mechanism and the relay settings that avoid it are described in transformer inrush current.

Which transformer survives a wet, H2S-laden, occasionally flooded basement?

Wastewater substations are chemically harsher than coastal ones. Hydrogen sulphide at 5–50 ppm plus condensing humidity attacks copper, silver and even the tin plating on busbars; the result is sulphide films that raise contact resistance and, on instrument terminals, open circuits. Add the risk of a digester or wet-well flood and the transformer type decision becomes an environmental one.

For an indoor substation in the blower house — the common layout — cast resin at IP54 with Class F insulation, 155 °C rated rise, is the defensible choice despite the 25–40% price premium: no oil containment, no fire wall, no oil testing regime, and encapsulated windings that ignore the H2S. If the substation must be outdoor, oil-immersed with a sealed tank and C5-M paint per ISO 12944 is cheaper and perfectly adequate. The full comparison matrix, including ester oils for fire-sensitive locations, is in oil-immersed vs dry-type transformers.

Where may the transformer sit relative to a classified wet well?

This is the clause that most often sends a packaged substation back to the drawing board. NFPA 820 classifies enclosed wet wells, grit chambers, primary clarifier influent channels and digester galleries as Class I Division 1 locations; IEC 60079-10-1 gives the equivalent Zone 1 or Zone 2 with a defined extent. A transformer — even a sealed dry type — is not rated for a Division 1 atmosphere, so it must sit outside the classified envelope, and the envelope is drawn by ventilation, not by distance alone.

Two practical rules follow. First, put the substation in a dedicated room with positive pressure from a clean-air intake, a gas detector at the door interlocked to the ventilation fan, and a raised plinth 300 mm above floor level so a 100 mm flood does not reach the plinth-mounted cable boxes. Second, do not rely on a 3 m or 5 m distance to substitute for the classification drawing — NFPA 820's envelopes are defined per room and per ventilation rate, and a review will ask for the drawing. This is also the point at which earthing philosophy gets fixed, and getting it wrong here is expensive later.

How do you connect a biogas CHP without back-feeding the utility?

Anaerobic digestion on this plant yields enough gas for two 150 kW_el CHP units — 300 kW, about 23% of the 1,289 kW average demand. The CHP is not a generator in the utility sense; it is a captive source that reduces import. That distinction sets the protection scheme and, critically, the earthing.

Three requirements apply. Reverse-power protection: a directional relay (ANSI 32) at the CHP breaker set to trip at 2–3% export, because most utilities prohibit export from a plant this size and a stuck contactor would otherwise energise the utility feeder during an outage. Anti-islanding: loss-of-mains detection, and if the utility does allow parallel operation, compliance with IEEE 1547 or the local grid code — not a generic over/under-voltage relay. Single neutral-earth bond: the CHP alternator is earthed at the plant neutral, so its star point must not be bonded to earth a second time. Two bonds create a parallel path through the earth grid, circulate third-harmonic and unbalanced current, and produce nuisance earth-fault trips that take weeks to diagnose. The correct arrangement — one bond, all exposed conductive parts on the equipotential bar, RCDs only where the earthing system requires them — is described in low-voltage grounding system design.

The transformer consequence is modest but real: the CHP reduces net import by 23%, so the transformers' load loss falls with the square of current while no-load loss stays fixed. At 15% average reduction in current, load loss drops 28% — about 2.4 kW across both units, USD 2,100 a year. That does not change the sizing, but it does change the loss evaluation: it makes the low-no-load-loss grade (S13 or amorphous) more valuable, not less.

What does a cheap transformer really cost over 20 years?

Capitalised loss evaluation is the only honest way to compare transformer offers. Using USD 6,300 per kW of no-load loss and USD 2,700 per kW of load loss — representative evaluation rates for a plant running 8,760 hours at 0.10 USD/kWh — the three standard Chinese loss grades on a 2,500 kVA, 11/0.415 kV unit separate clearly. Load loss is evaluated at the 44% operating point, which is what these units actually see.

Two conclusions, both counter-intuitive to a purchasing department. S11 is USD 5,120 cheaper than S13 to buy and USD 12,833 more expensive to own over 20 years; the S13 upcharge pays back in 1.9 years. Amorphous S15 beats S13 on total cost by only USD 2,380 over 20 years, with a 5.7-year payback on the extra USD 8,960 — worth specifying where the plant runs at high load factor and electricity is above USD 0.08/kWh, harder to justify where it is not. The loss-grade arithmetic and the temperature corrections behind it are in S11 vs S13 vs S20, and the wider method in transformer total cost of ownership. Note also that transformer losses are only 1.07% of plant energy — 120 MWh against 11,290 MWh — so the loss decision, worth about USD 12,800 over life, is an order of magnitude smaller than the blower decision, worth USD 180,000 a year. Buying a premium transformer while throttling blowers is the classic misallocation on these plants.

What belongs on the specification sheet?

A wastewater transformer datasheet is short, and every line on it should be traceable to something in this article. The following is the specification actually used for the 2 × 2,500 kVA case.

Two lines deserve emphasis because they are the ones omitted from generic datasheets. The harmonic derating calculation should be submitted by the manufacturer, not accepted as a K-factor label — it is the only document that proves the unit can carry 110% of demand on contingency. And the temperature-rise type test should be a type test, not a routine test; a 2,500 kVA unit specified at 60/65 K will be built to that limit, and only a type test proves it. Insist on the loss test certificate too: a transformer that misses its guaranteed P0 by 10% costs about USD 1,600 per kW of shortfall in capitalised terms.

Which mistakes appear most often in wastewater transformer procurement?

Six failures account for most of the problems seen on these projects, and all six are specification errors rather than manufacturing defects.

• Summing nameplates. 3,198 kW of connected load specified as 3,550 kVA instead of 2,203 kVA. The plant then runs at 60% load forever, paying no-load loss for capacity it never uses.
• Ignoring harmonic derating. A standard 2,500 kVA unit on 35% THDi is a 2,250 kVA unit, and the N+1 case lands at 98% of that. The failure mode is a hot spot, not a trip — it degrades insulation for years before the DGA test shows it.
• De-energising the standby transformer. A well-intentioned energy measure that adds 5.9 kW of copper loss and removes the automatic transfer capability in one action.
• Putting the substation beside the wet well. A zoning violation, a corrosion problem and a flood risk in one decision; the fix after commissioning costs more than the building.
• Two neutral-earth bonds at the CHP. Circulating current, spurious earth-fault trips, and weeks of investigation before someone opens the alternator star point.
• Buying on price without loss evaluation. The USD 5,120 saved on S11 returns as USD 12,833 of extra loss cost — a 150% penalty on the saving.

The sizing itself is a one-line calculation; the difficulty is that five different disciplines — process, electrical, fire, civil and utility interconnection — each add a constraint to the same transformer. Fixing the drive schedule first, then the harmonic calculation, then the redundancy case, then the environment, produces a specification that stays stable through design review. Doing it in the reverse order produces a datasheet that changes four times and a transformer that satisfies none of the reviews.

Sources / 资料来源

Source: IEC 60076-1 — Power transformers, Part 1: General.
Source: IEC 60076-2 — Temperature rise for liquid-immersed transformers.
Source: IEC 60076-3 — Insulation levels, dielectric tests and external clearances in air.
Source: IEC 60076-5 — Ability to withstand short circuit.
Source: IEC 60076-7 — Loading guide for oil-immersed power transformers.
Source: IEC 60076-11 — Dry-type transformers.
Source: IEC 60076-20 — Energy efficiency for transformers.
Source: IEC 60079-10-1 — Classification of areas: explosive gas atmospheres.
Source: IEC 60529 — Degrees of protection provided by enclosures (IP code).
Source: IEC 60364-4-41 — Low-voltage electrical installations: protection against electric shock.
Source: IEC 60034-30-1 — Efficiency classes of line-operated AC motors (IE1 to IE4).
Source: IEC 62271-200 — AC metal-enclosed switchgear and controlgear for rated voltages above 1 kV.
Source: NFPA 820 — Standard for Fire Protection in Wastewater Treatment and Collection Facilities (electrical area classification tables).
Source: NFPA 70 — National Electrical Code, Article 450 (transformers) and Article 250 (grounding).
Source: IEEE 519-2022 — Recommended practice and requirements for harmonic control in electric power systems.
Source: IEEE C57.110 — Recommended practice for establishing transformer capability when supplying non-sinusoidal load currents.
Source: IEEE C57.12.01 — General requirements for dry-type distribution and power transformers.
Source: IEEE 1547 — Standard for interconnection and interoperability of distributed energy resources.
Source: ISO 12944 — Corrosion protection of steel structures by protective paint systems (C5-M category).
Source: GB/T 6451 — Technical parameters and requirements for three-phase oil-immersed power transformers.
Source: Commission Regulation (EU) 2019/1783 — Ecodesign requirements for transformers (amending Regulation 548/2014).
Source: QDTB Engineering Team — oil-immersed and cast-resin transformer supply to municipal water, wastewater and industrial projects, 2015–2025 (demand factors, load factors, price bands, evaluation rates and lead times quoted here are industry experience values, not published standards).