Power System Fundamentals · Training Guide

Sources of Temporary Overvoltages (TOVs) Different causes and mechanisms in power systems

Reading time ≈ 25 min

What is a temporary overvoltage?

A temporary overvoltage (TOV) is an overvoltage that lasts longer than a fast switching or lightning transient and is usually associated with power-frequency or low-frequency system behaviour. It may follow switching, fault clearing, load rejection, islanding, transformer energisation or resonance excitation. This page focuses on TOVs that are caused or amplified by low-order network resonance; some other TOVs are dominated by a power-frequency voltage rise, but resonance can increase both the magnitude and the duration of the overvoltage.

Key terms used on this page

01TOV

Temporary overvoltage — a longer-duration overvoltage at or near power frequency.

02LC network

A network containing inductance (\(L\)) and capacitance (\(C\)), which together set its resonance frequencies.

03VT / CVT

Voltage transformer / capacitive voltage transformer — iron-core or capacitive voltage-measuring devices.

04Inrush current

The high, distorted magnetising current drawn when a transformer or iron-core reactor is energised.

05Pseudo-inrush

Inrush-like magnetising current caused by the voltage recovery after fault clearing.

06DC offset

A slowly decaying non-sinusoidal offset component in the current or flux.

07Point on wave

The instant on the voltage waveform at which switching or fault clearing occurs.

08Remanent flux

Residual magnetic flux left in a transformer or reactor core after de-energisation.

09Bifurcation

A sudden jump from one stable operating state to another in a nonlinear system.

10Weak grid

A grid with high source impedance and low short-circuit level — a less “stiff” supply.

Core mechanism

Why temporary overvoltages occur

Power systems contain both inductive and capacitive elements.

Inductive elements include transformers, reactors, and the system inductance. Capacitive elements include cables, capacitor banks, and line capacitance.

The interaction between inductance L and capacitance C creates resonance frequencies in the network. There are two main types of resonance.

Parallel resonance means the network impedance becomes very high at the resonant frequency. This is the main concern for TOVs because high impedance can amplify harmonic voltage.

Series resonance means the network impedance becomes very low at the resonant frequency. It can allow large harmonic currents to flow and may cause high voltage stress across individual components.

A TOV can occur when a switching event, fault-clearing event, or similar network event produces harmonic currents and these currents excite a network resonance.

For each harmonic order h, the relationship is:

\[ V(h) = Z(h) \times I(h) \]

\(V(h)\): harmonic voltage at order h
\(Z(h)\): network impedance at that harmonic frequency
\(I(h)\): harmonic current injected at order h
\(h\): harmonic order, where \(h=2\) is the 2nd harmonic, \(h=3\) the 3rd, and so on

On a 50 Hz system the 2nd harmonic is 100 Hz, the 3rd harmonic is 150 Hz and the 5th harmonic is 250 Hz.

This means: harmonic voltage = network impedance at that harmonic × harmonic current.

Therefore, if the network impedance is high at a certain harmonic frequency, even a moderate harmonic current can produce a high harmonic voltage. For example, if the network has a strong parallel resonance near the 3rd harmonic and a switching event produces 3rd-harmonic current, a high 3rd-harmonic voltage can appear.

Network sensitivity

Why weak and cable-rich systems are more sensitive

For a simple LC circuit, the natural frequency is:

\[ f_r = \frac{1}{2\pi\sqrt{LC}} \]

\(f_r\): natural resonance frequency, in hertz (Hz)
\(L\): equivalent system inductance, in henries (H)
\(C\): equivalent system capacitance, in farads (F)

This is a simplified relationship. Real power networks can have several resonance frequencies, but the basic idea remains useful.

A large cable network increases capacitance C. A weak grid, or a grid with low short-circuit level, has a higher effective inductance L. Both of these move the resonance frequency lower. This is important because the resonance may move closer to low-order harmonics such as the 2nd, 3rd, or 5th harmonic.

A low short-circuit level means the upstream network has a higher equivalent impedance — in simple terms, the grid is less “stiff”, so the same harmonic current or switching disturbance produces a larger voltage change.

These low-order harmonics matter because switching events, especially transformer energisation, can produce them strongly.

This type of TOV is more likely in systems with one or more of the following characteristics:

transmission systems with long cables
systems with very long overhead lines
weak system conditions
network restoration
network contingencies
islanding conditions
light loading conditions

Light loading is important because load provides damping. When there is little load, there is less damping, so the overvoltage can last longer.

Event types

Main triggering events

A TOV normally needs two things together: a disturbance, and a network condition that allows voltage amplification. The disturbance may inject harmonic current, change the voltage suddenly, or change the network configuration.

The main events that can cause this type of TOV are:

energisation of power transformers
energisation of magnetic-core shunt reactors
cable energisation
capacitor bank energisation
fault clearing
circuit breaker auto-reclose
system islanding
load rejection

These events can either inject harmonic currents into the system or cause sudden voltage changes that excite a resonance. Cable and capacitor bank energisation are included because they can change the network voltage and interact with the system capacitance and resonance conditions.

Major trigger

1. Transformer and shunt-reactor switching

When a transformer is energised, the core can saturate. When the core saturates, the magnetising current becomes non-sinusoidal. This current is called inrush current, and it contains harmonics.

Transformer inrush current is usually dominated by the 2nd harmonic, with DC offset and also significant 3rd, 4th, and 5th harmonic components.

The DC offset matters because it shifts the flux in the core and can push it further into saturation. The 2nd harmonic is typically dominant during energisation, and is one of the reasons transformer inrush looks so different from normal load current.

If one of these harmonic currents meets a high network impedance at the same harmonic frequency, a TOV can occur. The process is:

The same principle applies to magnetic-core shunt reactors — a shunt reactor built with an iron core. Like a transformer, it can saturate during energisation and draw a non-sinusoidal magnetising current rich in harmonics.

Factors affecting the TOV

Resonance frequency and impedance peak. A resonance close to an inrush harmonic can be dangerous. In a 50 Hz system, 100 Hz is the 2nd harmonic and 150 Hz is the 3rd harmonic. If the network has a high impedance peak near one of these frequencies, the TOV can be higher and last longer. A tall and sharp impedance peak means the system has low damping.

Saturation characteristic. The transformer core design affects how much harmonic current is produced during energisation. A transformer that saturates more strongly can produce higher harmonic current.

Remanent flux. When a transformer is switched off, some magnetic flux can remain in the core. When the transformer is energised again, this remanent flux can increase the saturation level. The worst case happens when the remanent flux and the new flux caused by switching are in the same direction. In that case, the core can go deeper into saturation and produce a larger inrush current.

Adjacent transformers. Other transformers connected nearby can also be affected during energisation. They may saturate through the shared coupling impedance. This is called sympathetic inrush. Sympathetic inrush can make the TOV worse.

Switch closing instant. The instant when the breaker closes affects the maximum flux in the transformer core. For three-phase breakers, the three poles may not close at exactly the same time. This non-simultaneous closing can make the inrush current and the TOV more difficult to predict.

Loading. A loaded transformer usually has lower inrush current and better damping. Therefore, no-load or light-load energisation is usually more severe.

Important point: dynamic resonance

The worst case is not always exactly at an integer harmonic. During transformer saturation, the effective inductance of the transformer changes. Because of this, the resonance frequency can shift during the event.

This means that a resonance slightly below an integer harmonic can still be dangerous. For example, a resonance slightly below the 3rd harmonic may still be excited during transformer energisation.

Takeaway

Transformer energisation injects low-order harmonic currents (mainly the 2nd harmonic); these produce a high TOV when the network impedance is high at the same frequency.

Fault response

2. Fault clearing

During a fault the voltage collapses; when the fault is cleared, the voltage recovers suddenly. A connected transformer core can experience this sudden recovery like a new energisation event — even though the transformer was already connected — so it draws a pseudo-inrush current whose harmonic content is similar to true inrush. This is why fault clearing can produce inrush-like effects.

The process is:

Three-phase faults are especially important because they can cause the largest voltage collapse.

Items to consider in fault-clearing studies

The following items are important when assessing fault-clearing TOVs:

point of fault initiation on the voltage waveform
point on wave at fault clearing
distance to the fault
fault clearing time
delay before any auto-reclose

Fault location is very important. A fault close to the bus causes a larger voltage collapse. When the fault is cleared, the voltage recovery is stronger, so the TOV can be higher. The severity of the recovery also depends on the point on wave at which the fault is cleared.

A fault further away usually produces a lower TOV and the overvoltage damps more quickly.

Takeaway

Fault clearing can behave like a sudden re-energisation — especially when the voltage collapse and recovery are severe — producing pseudo-inrush current and a resonance-driven TOV.

Nonlinear resonance

3. Ferroresonance

Ferroresonance is a nonlinear oscillation between an iron-core inductance and a capacitance.

Ferroresonance should not be treated as ordinary linear LC resonance, because the iron-core inductance is not constant — it changes with the applied voltage and the saturation level of the core.

Examples of iron-core inductances include power transformers, inductive voltage transformers, and reactors. Examples of capacitances include long lines, cables, capacitive voltage transformers, series capacitors, shunt capacitors, and circuit breaker grading capacitors.

Ferroresonance is nonlinear. This means the same system can have more than one possible stable operating condition. A small change in the system can cause a sudden jump into a ferroresonant condition. This sudden jump is called a bifurcation.

Damping can help suppress ferroresonance. One known solution is to connect a resistor to the secondary side of the voltage transformer. An open-delta VT secondary is a voltage-transformer secondary connection commonly used to detect residual (zero-sequence) voltage; adding damping resistance in this circuit can suppress ferroresonant oscillations in some configurations. In the case study, a damping resistor in the open-delta VT secondary eliminated ferroresonance during transformer energisation.

Takeaway

Ferroresonance is nonlinear and can appear suddenly after switching, especially where iron-core devices interact with network capacitance; damping is the primary mitigation.

Grid separation

4. Islanding

Islanding happens when part of the network becomes separated from the main grid while generation is still connected. The process is:

If generation inside the island is higher than the load, the voltage rises. This is worse when distributed generators operate at constant power factor: at constant power factor a generator does not absorb or control reactive power in response to the rising island voltage, so it does nothing to bring the voltage back down. The overvoltage therefore persists until the generator protection disconnects the generation.

In islanding the voltage rise can come from two different mechanisms:

generation greater than load, which causes a power-frequency voltage rise; and
breaker opening and network reconfiguration, which can excite a resonance — a transient resonance-frequency component on top of the power-frequency rise.

A weak system has higher source impedance. This can lower the resonance frequency and make the overvoltage higher.

Takeaway

An islanded area with surplus generation drives a power-frequency voltage rise — worsened by constant-power-factor generators — and can also excite a resonance, persisting until protection clears the island.

Switching event

5. Load rejection

Load rejection happens when a transmission line or a large inductive load is suddenly disconnected. When this happens, the generator can speed up and the busbar voltage can rise.

The duration of this overvoltage depends on the generator’s voltage control (AVR), the turbine/governor response, the system damping and the protection action — which is why a load-rejection TOV is usually slower than a switching or lightning transient.

The TOV is usually lower when the system has short lines and high short-circuit power at terminal stations. The TOV is usually higher when the system has long lines and low short-circuit power at generating sites.

If the voltage rises enough, transformers can saturate. This can produce harmonic currents, which can then excite low-order resonances.

Takeaway

Suddenly losing load lets the generator speed up and the busbar voltage rise; the duration is set by voltage control, governor response, damping and protection, and a large rise can saturate transformers and excite resonance.

Summary

A TOV can occur when a switching event, fault-clearing event, islanding event, or load-rejection event produces harmonic currents or sudden voltage changes that excite a low-order network resonance. The core mechanism is:

\[ V(h) = Z(h) \times I(h) \]

Harmonic Current at a Given Frequency × High Network Impedance at the Same Frequency = High Harmonic Voltage

Key takeaways

The core mechanism is V(h) = Z(h) × I(h). Harmonic current at a given frequency, multiplied by high network impedance at the same frequency, produces high harmonic voltage. Parallel resonance is the main concern because it raises network impedance at the resonant frequency.
Weak, cable-rich, and lightly loaded systems are most at risk. Higher capacitance and higher source inductance lower the natural resonance frequency, moving it closer to low-order harmonics that switching events can excite. Light loading reduces damping, so overvoltages last longer.
Transformer energisation is one of the most common triggers. Inrush current is dominated by the 2nd harmonic with additional 3rd, 4th and 5th harmonic content. Remanent flux, non-simultaneous pole closing, and sympathetic inrush from adjacent transformers can all increase the severity.
Fault clearing can produce pseudo-inrush current. The sudden voltage recovery after fault clearing behaves like a re-energisation. Three-phase faults cause the largest voltage collapse and therefore the strongest recovery. Fault distance from the busbar has a significant effect on TOV severity.
Ferroresonance is nonlinear and can be triggered by small changes. It involves an iron-core inductance and a capacitance in the same circuit. Bifurcation — a sudden jump into a ferroresonant state — can occur with small system changes. Damping, such as a resistor in the VT secondary open-delta circuit, is the primary mitigation.
Islanding and load rejection cause power-frequency TOVs driven by generation surplus. When generation exceeds local demand in an islanded area, or when a large load is suddenly dropped, the voltage rises until protection operates. These events may also excite resonance if the voltage rise saturates transformers.

Part Two Reading now

Sources of Temporary Overvoltages (TOV)

A focused companion on temporary overvoltages — earth faults and the earth-fault factor, load rejection, resonance, transformer energisation and ferroresonance, and the events that drive power-frequency TOV.

Series progress 2 of 2

Part One

Types of Overvoltages

The IEC 60071-1 classification — temporary, slow-front, fast-front and very-fast-front overvoltages: their shapes, durations, the events that produce them, and the insulation-coordination studies each one drives.