Current IntelPower LibraryShop main store

How-to

Estimating transformer energization voltage sag from source impedance

Power quality and harmonics Published 8 min read

Short answer

Screen the dip in pu with |ΔV|pu = k × St ÷ Ssc, using k = 10–12 times transformer full-load current at about 0.1 power factor. A 50 MVA transformer on a 500 MVA short-circuit source screens at 1.0–1.2 pu, which invalidates the fixed-current method and calls for a transient study; a 5 MVA unit on the same bus screens at 0.10–0.12 pu, a usable bound.

On this page
  1. What is the quickest way to estimate transformer energization voltage sag?
  2. Where do the 10–12 times full-load current and 0.1 power factor assumptions come from?
  3. How do I turn short-circuit MVA into source impedance?
  4. How do I calculate the inrush current in amps?
  5. How much sag does a 50 MVA transformer cause on a 500 MVA system?
  6. Can motor-starting software model transformer inrush?
  7. When does the shortcut stop working and a study become mandatory?
  8. How do I verify the sag by measurement at energization?
  9. Questions asked

Screen before you simulate: the transformer energization voltage sag, first cycle, worst phase, is k × St ÷ Ssc in pu — St the transformer MVA, Ssc the short-circuit MVA at the energizing bus, k = 10–12 times full-load current drawn at about 0.1 power factor for about 0.1 s. A 50 MVA transformer on a 500 MVA source screens at 1.0–1.2 pu: not a sag figure but a flag that the fixed-current shortcut has left its range. A 5 MVA unit on the same bus screens at 0.10–0.12 pu, a usable bound. What follows is the derivation, the arithmetic with units, the decision line, and the measurement that closes the estimate out.

What is the quickest way to estimate transformer energization voltage sag? #

Multiply the transformer MVA by 10–12 and divide by the short-circuit MVA at the bus: |ΔV|pu = k × St ÷ Ssc is the per-unit voltage-drop magnitude on the worst phase during the first cycles of inrush — the number to compute before anything else.

The shortcut is Ohm's law on a common base. Full-load current: IFL = St ÷ (√3 × VLL). Short-circuit current at the same bus: Isc = Ssc ÷ (√3 × VLL). Per-phase source impedance: |Zs| = VLL² ÷ Ssc. The drop from an inrush of k × IFL through that impedance, divided by the phase voltage VLL ÷ √3, collapses to k × St ÷ Ssc. Note what the bare ratio means: St ÷ Ssc is the full-load drop, and the inrush multiplies it by k.

Run the screen in this order:

  1. Confirm Ssc at the exact bus and switching lineup that will energize the transformer.
  2. Put it on the transformer base: Zsource,pu = St ÷ Ssc.
  3. Pick k = 12 for a worst-phase screen, k = 10 for a typical close.
  4. Multiply: |ΔV|pu = k × St ÷ Ssc.
  5. Read the result against the decision table later in this article.

Where do the 10–12 times full-load current and 0.1 power factor assumptions come from? #

They are a screening convention, not a nameplate value: 10–12 × rated current at about 0.1 power factor for roughly the first 0.1 s reproduces a worst-phase inrush closely enough for a dip estimate when no core-saturation data exists.

What actually sets the peak is the point-on-wave closing angle, the residual flux left in the core from the last de-energization, the saturation characteristic, the winding connection and the pole scatter of the switching device. That is why the convention is a range, not a number.

Two cautions travel with it. The highest inrush usually lands on one phase while the other two stay milder, so apply the screen as a single-phase event on the worst phase, not split three ways. And the current is harmonic-rich, which the RMS phasor screen ignores; if your problem is protection rather than bus voltage — fuses, breaker settings, differential restraint — read estimating transformer inrush current for protection coordination before touching settings.

How do I turn short-circuit MVA into source impedance? #

Divide line-to-line voltage squared by short-circuit volt-amperes — |Zs| = VLL² ÷ Ssc, ohms per phase — or, for the per-unit value on the transformer base, divide transformer MVA by short-circuit MVA: 50 ÷ 500 = 0.1 pu.

Use the short-circuit level for the actual bus and the actual lineup, including any parallel feeds, cables and upstream transformers that will be in service at the moment of closing. If your figure applies at another voltage level, refer it through each intervening transformer by the square of its turns ratio and add the series impedances on one base; the utility-side arithmetic is spelled out in converting utility per-unit impedance from fault MVA.

A parallel source raises Ssc and shrinks the dip; long cable runs or local generation lower it. The per-unit result itself does not care about voltage — only the amperes change with the kV.

How do I calculate the inrush current in amps? #

Full-load current is St ÷ (√3 × VLL) and the screening inrush is 10–12 times it on the energized winding: a 50 MVA transformer at 115 kV carries 251 A full load, so the k = 12 screen injects 3,012 A.

You need the amps for the things the pu figure cannot do: setting a measurement range that survives the peak, and putting a real number in front of anyone who asks what the breaker between the source and the transformer will see. Compute them on the winding you actually energize — primary side, using that winding's kV.

How much sag does a 50 MVA transformer cause on a 500 MVA system? #

The screen returns 1.0–1.2 pu of drop magnitude — a 100–120 percent sag, which is physically impossible — so the honest deliverable is not a percentage: this source strength puts the fixed-current method out of range, and the project goes to a transient study.

Worked example, with every input stated. Illustrative case: St = 50 MVA, energizing bus at 115 kV, Ssc = 500 MVA at that bus, k = 12.

  • IFL = 50,000,000 VA ÷ (1.732 × 115,000 V) = 251 A
  • Screening inrush: 12 × 251 A = 3,012 A
  • Source impedance: |Zs| = (115,000 V)² ÷ 500,000,000 VA = 26.45 Ω per phase
  • Phase voltage: 115,000 V ÷ 1.732 = 66,395 V
  • Drop magnitude: 3,012 A × 26.45 Ω = 79,670 V; 79,670 ÷ 66,395 = 1.20 pu
  • Shortcut check: 12 × 50 ÷ 500 = 1.20 pu

A drop of 1.0–1.2 pu cannot be read as a sag: the bus cannot fall to zero, and long before that the depressed bus voltage chokes the inrush producing the drop — current and voltage settle somewhere between the two extremes. That interaction is outside a fixed-current screen, so the deliverable here is a transformer-energization study in an electromagnetic-transient program.

For contrast, a 5 MVA unit on the same 500 MVA bus screens at 12 × 5 ÷ 500 = 0.12 pu: worst-phase retained voltage bounded at 0.88 pu for about a tenth of a second. That number you can take to the ride-through check for contactors, drives and control-power supplies.

Can motor-starting software model transformer inrush? #

Only as a fixed-current injection: if the package lets you impose a constant current at 0.1 power factor for 0.1 s on one phase, it reproduces this screen exactly; its motor acceleration model answers a different physical problem.

The dip arithmetic in motor-starting studies is the same phasor trick — a large reactive load on a source impedance — demonstrated for a motor in motor starting voltage dip calculation on an 11 kV bus. But a transformer is not a motor: the acceleration model carries no core saturation, no residual flux, no closing-angle sensitivity, no pole scatter and none of the harmonics that dominate the first cycles. With this heavily lagging current against a high-X/R source, the phasor subtraction lands within a few degrees of simple magnitude subtraction, so the fixed-current run adds little beyond the shortcut.

When does the shortcut stop working and a study become mandatory? #

Commit at the number: above 0.5 pu the coupling the screen ignores is already wide enough to matter, and at 1.0 pu and above the method is meaningless — the 50 MVA case, at 1.2 pu, goes straight to an electromagnetic-transient study.

Screen result k × St ÷ Ssc What the number is worth Decision
0.3 pu or below Dependable worst-phase upper bound Compare against load ride-through; no further analysis if margins hold
0.3 to below 1.0 pu Overstates the dip; voltage–current coupling grows through the band Order the study when the bound touches load ride-through, and above 0.5 pu regardless
1.0 pu and above Method invalid Transient study before planning the energization; the 50 MVA example sits here

The study must carry the transformer saturation curve, a spread of residual-flux states, per-pole closing times rather than one simultaneous event, and all three phases — then compare both the depressed RMS fundamental and the distorted instantaneous waveform against what the connected equipment tolerates. If it shows an unacceptable dip, the fixes are a stiffer switching configuration, controlled point-on-wave closing, or pre-insertion resistors; the study sizes them, the screen cannot.

How do I verify the sag by measurement at energization? #

Capture all three bus voltages and the energizing current with a recorder fast enough to keep the first-cycle waveform, then read the minimum per-phase RMS voltage, the dip duration and the peak current, and compare them with the screen.

IEEE Std 1159-2019, the recommended practice for monitoring electric power quality, is the reference for classifying sag magnitude and duration; confirm reporting requirements with the equipment documentation and the authority having jurisdiction. Make every connection with the bus de-energized, locked out and verified absent of voltage — energized work is outside this procedure and governed by NFPA 70E, 2024 edition.

  1. Log the lineup: source configuration, parallel feeds in or out, connected loads, and how long the transformer stood de-energized.
  2. Fit the recorder to the de-energized, verified bus, with a sampling rate that resolves the first half-cycle rather than RMS trends alone.
  3. Energize and capture the closing instant of each pole; expect different depths on repeat operations, because closing angle and residual flux vary.
  4. Read minimum RMS voltage per phase, dip duration and peak current, and keep the phases separate — averaging erases the controlling phase.
  5. Walk the loads afterwards: contactor dropout, drive undervoltage records, control-power ride-through, relay event logs.

If the worst measured phase sits at or under the screen bound and the loads survived, the screen is confirmed and you schedule mitigation when the margin is thin. If a load dropped out, correct the source strength or the switching — not the protection settings.

Questions asked

Do I need the source X and R separately for the screen to be valid?

For the magnitude shortcut, no. At 0.1 power factor the inrush lags roughly 84 degrees, and a high-X/R source impedance sits near 90 degrees, so the two angles cancel and the I × Z phasor lands almost in phase with the bus voltage — nearly the whole magnitude subtracts. You need X and R split out only when a software phasor run must also give the retained-voltage angle.

Why does my screen show a sag over 100 percent?

Because the fixed-current assumption has broken down. A 1.0–1.2 per-unit drop would drive the bus to or below zero, and the collapsing voltage would throttle the very current causing the drop; the real event settles well short of both extremes. Treat any screen result at or above 1.0 pu as an instruction to run a transformer-energization transient study, not a number to report.

How long does an energization sag actually last?

Use 0.1 s at full assumed depth for screening — the peak occupies the first few cycles and the dip shallows over later cycles as the core flux symmetrizes. Judge sensitive loads on depth and duration together: a contactor that rides through 0.1 s at 88 percent voltage may still drop out if an unfavorable closing angle and residual flux stretch the decay.

Does it matter which bus or source configuration energizes the transformer?

Yes, the screen is only as good as Ssc at the actual switching point. Closing a parallel feed raises the short-circuit level and shrinks the dip; energizing over long cables, an upstream transformer or local generation lowers it. Refer impedances through every intervening transformer by the square of its turns ratio before dividing St by Ssc.

What fixes an energization dip a study says is too deep?

Stiffen the source or cut the inrush: energize from the stronger bus with a parallel feed in service, use controlled point-on-wave closing, or fit pre-insertion resistors. Size these from the transient study — the optimum depends on the core saturation curve and the switchgear pole scatter, none of which the screen sees.

Keep reading

  • How-to

    Estimating transformer inrush current for protection coordination

    The rules of thumb are screening gates, not coordination inputs: reported inrush runs 8 to 30 times full-load current and lasts 10 cycles to a minute. Plot the decay curve Ipu = 6 × t^-0.31 with t in seconds — it passes through 25× at the first peak, 12× five cycles in and 3× at 10 s — keep the protective device curve above it and below the transformer damage curve, and handle cold-load pickup separately.

    Protection: breakers, fuses and SCCR

  • Selection guide

    Stopping breaker trips on 25 kVA transformer inrush

    Set a breaker with a programmable electronic trip unit so its instantaneous pickup sits above the transformer's inrush peak, and the energization trips stop with no added power hardware: for a 25 kVA, 400 V primary that is 36 A rated current and up to about 540 A first-cycle inrush. A motor soft starter is the fallback only if its documentation permits a transformer magnetizing load; resistor pre-insertion is the engineered option when the supply itself cannot tolerate the dip.

    Protection: breakers, fuses and SCCR

  • How-to

    Closed delta transformer fault current calculation

    Compute each delta branch from its own kVA and impedance, then take phasor differences at the line terminals: for a 150 kVA leg with two 100 kVA legs at 240 V and 3.3% Z each, the bolted three-phase terminal fault is highest at each end of the 150 kVA winding and lower at the junction of the two 100 kVA windings. Size secondary equipment on the higher terminal value.

    Protection: breakers, fuses and SCCR