Analyzing ferroresonance in transformers and distribution networks

Key Takeaways

• • Ferroresonance comes from a specific circuit state, so switching sequence and capacitance matter as much as transformer rating.

  • Lightly loaded distribution transformers and single phase switching cases deserve first attention in any study plan.

  • Reliable prevention depends on nonlinear modelling, explicit grounding detail, and operating steps that avoid partial energization.

Ferroresonance will damage transformers when an ordinary switching step leaves a saturated core interacting with enough system capacitance to hold a distorted overvoltage.

That’s why the event so often looks mysterious after the fact. You can have no visible fault, modest current, and a transformer that still comes out of service with insulation stress, noise, heating, or repeated fuse operations. A steady state load flow will miss it, and a simple transient model with linear inductance will miss it too. You need the switching sequence, the capacitance, and the nonlinear magnetic behaviour in the same study. The United States has roughly 55 million to 69 million distribution transformers in service, which means even uncommon switching interactions deserve planning attention. That scale is why ferroresonance belongs in routine switching studies for transformers and distribution feeders instead of a folder reserved for rare failures.

What causes ferroresonance during routine transformer switching events

Ferroresonance starts when a saturable transformer inductance is left interacting with system capacitance after an unusual switching state. The circuit settles into a nonlinear oscillation instead of normal sinusoidal operation. Voltage stays elevated, distorted, and sustained until the topology changes or enough damping is added.

A common case appears after one phase of a three-phase transformer bank opens while cable capacitance, breaker grading capacitance, or line to ground capacitance still couples voltage into the disconnected phase. The transformer is no longer fed in a balanced way, but it is not fully isolated either. That partial connection is enough to push one limb into saturation and keep feeding the oscillation. Operators often read the event as a nuisance switching incident because the current doesn’t look like a conventional fault.

You should treat ferroresonance as a circuit condition because the switching arrangement controls the outcome. The same transformer can behave normally in one switching sequence and fail badly in another. That’s why post-event inspection rarely gives a clean answer on its own. 

“The cause sits in the interaction between network capacitance, switching status, grounding, and the magnetization curve.”

The highest risk appears in lightly loaded distribution transformers

Lightly loaded distribution transformers are most exposed because damping is weak and the magnetizing branch dominates current. Small capacitances from cables, bushings, breaker contacts, or open conductor lengths can then support an oscillation that would collapse quickly under heavier loading. No load conditions are the first cases you should screen.

Consider a pad mounted distribution transformer at the end of an underground feeder during maintenance switching. The secondary side is carrying almost no load, one phase is opened upstream, and the feeder cable still contributes phase to ground capacitance. That combination creates exactly the kind of weakly damped circuit in which ferroresonance will settle and persist. The transformer looks idle, and that’s part of the problem.

This is why ferroresonance in distribution transformers is so often missed during planning. Engineers tend to focus on peak load, fault duty, or voltage drop, while the dangerous case sits at the opposite end of the operating range. Rural feeders, seasonal loads, spare transformers, and standby banks deserve extra attention. A lightly loaded unit will not protect itself just because system power is low.

Single-pole switching creates the circuit conditions ferroresonance needs

Single-pole switching creates ferroresonance when one or two phases remain coupled through capacitance while another phase pushes the transformer core into saturation. That partial energization is more dangerous than a clean three-phase open or close because the circuit keeps an uneven voltage reference and a weak path for oscillation.

One familiar utility case starts with a blown fuse on one phase of a three-phase overhead bank. Two phases remain connected, the transformer core no longer sees balanced flux, and the open phase still picks up voltage through stray capacitance. Another case appears when a switch doesn’t operate as a true gang device and poles separate at different times. Neither event looks dramatic from the yard, yet both create the exact topology ferroresonance needs.

You should read single-pole status as a trigger flag because it shapes how you assess switching exposure. Protection settings, fuse practices, and switching procedures all matter here because they shape how long the circuit sits in that awkward partial state. The longer that state exists, the more likely the transformer will settle into sustained abnormal voltage instead of passing through a short transient.

Sustained overvoltage follows a nonlinear inductance interacting with system capacitance

Sustained overvoltage appears because transformer inductance is not fixed. Once the core saturates, effective inductance shifts with voltage and flux. System capacitance keeps feeding energy into that nonlinear branch, so the circuit can settle into subharmonic, quasi periodic, or chaotic states instead of fading out after a brief transient.

During a ferroresonance event, you will often see one phase to ground voltage climb while waveform shape deteriorates badly. A study case on a lightly loaded grounded wye transformer can show a phase sitting near 1.8 per unit with strong low order distortion, loud noise, and heating that does not match the measured current. Published technical guidance reports ferroresonance overvoltages from about 1.25 to 6 per unit in power systems.

That mismatch between voltage severity and current magnitude explains why the problem hides so well. Standard overcurrent protection won’t tell the whole story, and a short RMS snapshot can make the condition look less severe than it is. You need time domain voltage, flux, and neutral shift to judge actual stress. Persistence does the damage, and the first crest is only part of the problem.

Core saturation detail decides if a simulation can reproduce ferroresonance

Core saturation detail decides if your simulation will show ferroresonance at all. A transformer model with fixed inductance cannot reproduce the nonlinear exchange that sustains the event. You need a magnetization curve, proper saturation knee, residual flux treatment, and the surrounding phase to ground capacitances represented explicitly.

Take two models of the same feeder transformer. The first uses a linear magnetizing branch and a generic ideal transformer. The second includes a nonlinear excitation branch, residual flux from the prior de-energization, and cable capacitance on each phase. Only the second model will reproduce the overvoltage and distorted waveform that field crews actually report after an odd switching sequence.

That modelling detail is where execution matters. SPS SOFTWARE gives you a transparent way to inspect the nonlinear transformer representation, the feeder capacitances, and the switching states in one physics-based study, which is exactly what ferroresonance work requires. If any one of those pieces is hidden or simplified away, the case won’t show the risk clearly.

A ferroresonance study should test credible switching cases first

A useful ferroresonance study starts with the switching cases your system will actually see. You should not begin with extreme contingencies or rare protection failures. Start with no load energization, single phase interruptions, fuse loss, staggered pole operation, and maintenance isolation steps that leave capacitance connected.

A practical workflow is short and disciplined:

• Model the transformer at no load and at very light load first.
• Test single phase open states and staggered pole operations.
• Include phase to ground capacitance from cables, bushings, and switchgear.
• Represent the nonlinear magnetization curve and residual flux explicitly.
• Record phase voltages, neutral shift, distortion, and event duration

A feeder study becomes far more useful when you rank cases by operating credibility. A maintenance cutover on a cable-fed pad mount will deserve attention before an unlikely multi device malfunction. The checkpoint table below helps keep that focus clear.

Prevention starts with controlled switching sequences during transformer energization

Prevention works best when switching doesn’t leave the transformer in a partially energized state. Three pole operation, clear isolation of all phases, and procedures that avoid lightly loaded single phase exposure will remove the conditions ferroresonance needs. Operational discipline usually matters more than adding hardware after failures appear.

Picture a crew energizing a spare three-phase bank after repairs. If poles close together as a true gang operation and the bank carries a modest temporary burden, the circuit passes through a short transient and settles. If one pole lags, one fuse is left open, or a long cable remains connected on an unloaded phase, the same task can turn into a sustained overvoltage event. The transformer did not change. The sequence did.

You should turn that observation into switching rules. Review fuse application on three-phase banks, identify devices that can leave one phase open, and write maintenance steps that do not strand transformer windings behind feeder capacitance. Some sites will justify damping resistors or alternate device selection, but most exposure is removed earlier through sequence control and explicit study of credible field operations.

Missed neutral grounding details keep ferroresonance risk hidden

Neutral grounding detail often decides if a questionable switching case stays harmless or turns into ferroresonance. Ground reference controls voltage balance, neutral shift, and the way capacitance returns current. If the grounding path is omitted, simplified, or misconnected in the model, you will underestimate both overvoltage severity and event persistence.

A grounded wye primary transformer with an unintended high impedance neutral path can behave very differently from the same nameplate unit with a solidly grounded neutral. That difference shows up most clearly after partial switching, when one phase is weakly fed through capacitance and the neutral point starts to move. Field teams often read the outcome as random transformer trouble because the grounding detail is easy to miss on drawings and even easier to omit in simplified studies.

“That level of modelling discipline is what keeps routine operations from becoming unexplained transformer damage.”

The useful judgement is straightforward. Ferroresonance is rarely a mystery once the model includes the actual switching path, the nonlinear core, the feeder capacitance, and the neutral connection. SPS SOFTWARE fits that job because you can inspect each assumption directly and test the switching steps that planners and crews will actually use.