A temporary overvoltage is a power-frequency rise in voltage that lasts from a fraction of a second to several hours. It comes from earth faults, load rejection, resonance and synchronizing operations. Insulation co-ordination represents it with the 1-minute power-frequency voltage shape and an amplitude that depends on the cause.
Short answer
A temporary overvoltage (TOV) is a rise in voltage at power frequency. It lasts from a fraction of a second to several hours. The usual causes are an earth fault, the rejection of a load, resonance or ferroresonance, and a synchronizing operation. For insulation co-ordination, IEC 60071-2 represents it with the standard 1-minute power-frequency withstand voltage shape. Its amplitude is chosen according to the cause.
What characterizes a temporary overvoltage
Three quantities define it: amplitude, waveshape and duration. All three depend on the origin, and the amplitude and the shape can change while the overvoltage lasts.
For insulation co-ordination, the standard takes the standard short-duration power-frequency voltage, 1 minute, as the representative shape. The amplitude can be given as a single value (the assumed maximum), as a set of peak values or as a full statistical distribution.
Two things are considered when fixing the amplitude: the real amplitude and duration of the overvoltage in service, and the power-frequency voltage versus time withstand characteristic of the insulation. If that characteristic is not known, the standard allows a simplification: take the maximum real overvoltage that lasts less than 1 minute and assume a duration of 1 minute.
The per-unit (p.u.) base is the peak of the highest phase-to-earth system voltage: 1.0 p.u. = Us·√2/√3, where Us is the highest voltage of the system. Every p.u. value in this guide uses that base.
Earth fault
A single phase-to-earth fault raises the phase-to-earth voltage on the other two phases. Phase-to-phase overvoltages, or overvoltages across a single insulation, do not normally appear. The shape is that of the power-frequency voltage.
The amplitude depends on the neutral treatment and on the fault location. How far the voltage rises on the healthy phases depends on the earth fault factor, and each neutral treatment is explained in isolated versus earthed neutral. For normal configurations the maximum amplitude is taken.
The duration is that of the fault, until it is cleared:
| Neutral earthing | Usual overvoltage duration |
|---|---|
| Solidly earthed neutral | Less than 1 s |
| Resonant (Petersen coil) earthing with fault clearing | Less than 10 s |
| System operated with a sustained earth fault | Up to several hours |
In the last case it may be necessary to treat the voltage during the fault as if it were the continuous operating voltage.
The highest voltage on a healthy phase does not depend only on the earth fault factor. It also depends on the operating voltage at the instant of the fault, which is normally taken equal to the highest voltage of the system.
Load rejection
When a load is disconnected, the three phase-to-earth voltages rise by the same amount. For that reason the same relative overvoltage appears phase-to-earth and phase-to-phase. It depends on the load that is lost, on the network configuration after the disconnection and on the source: its short-circuit power and the speed and voltage regulation of the generators.
It affects mainly the equipment on the source side of the breaker that opened. It is strongest when load is lost at the end of a long line, because of the Ferranti effect.
The standard gives these reference values for the analysis:
- Medium-extent systems: a full load rejection normally produces less than 1.2 p.u. The duration depends on the voltage-regulating equipment and can reach several minutes.
- Extended systems: it can reach 1.5 p.u. or more when there is a Ferranti effect or resonance, with durations in the order of seconds.
- Longitudinal overvoltage: if only static loads remain on the disconnected side, it is normally equal to the phase-to-earth overvoltage. If there are motors or generators, a system separation can produce two phase-to-earth components in phase opposition, with a maximum normally below 2.5 p.u.
The highest values occur in systems with long lines and low short-circuit power at the generation, typical of the early stages of an extra-high-voltage network.
Resonance and ferroresonance
They appear when energizing circuits that combine large capacitive elements (lines, cables, series-compensated lines) with non-linear magnetizing inductive elements (transformers, shunt reactors), or after a load rejection.
They can reach extremely high values. For that reason the standard requires avoiding or limiting them with their own measures: detuning the network from the resonance frequency, changing its configuration or adding damping resistors. They should not be used as the basis for choosing the rated voltage of an arrester or for designing the insulation, unless those measures are not enough. A specific case is ferroresonance in voltage transformers.
Synchronization
Closing a breaker between two parts that are not in synchronism produces a longitudinal overvoltage. The standard derives it from an expected service value: twice the phase-to-earth operating voltage, lasting from a few seconds to a few minutes.
If synchronizing is frequent, the probability of an earth fault is added. In that case the representative amplitude is the sum of the assumed maximum earth-fault overvoltage on one terminal and the operating voltage, in phase opposition, on the other.
When causes are combined
They are not simply added. The probability that they occur at the same time is examined first, because adding them leads to arresters with a higher rated voltage and to higher protection levels and insulation levels. That is justified technically and economically only if the probability is high enough.
One case that cannot be neglected: a line fault followed by a breaker failing to open, with a generator connected through a transformer to a long faulted line. The events are not statistically independent. The standard recommends network studies when a combination is considered likely.
What to do to limit them
Each cause is controlled differently. Earth-fault overvoltages are controlled through the design of the network and of the neutral, with fast earthing of neutrals that become separated. Sudden load changes are controlled with shunt reactors, series capacitors or static compensators. Resonance is avoided with the measures in the previous section.
Once known, temporary overvoltages are the basis for choosing the rated voltage of the surge arrester. The overvoltage function 59 monitors them in the network.
To go further, see Slow-front overvoltages: switching and fault.
When to contact us
If your network has long lines, remote generation, transformers whose neutral can end up isolated, or a new high- or extra-high-voltage project, the representative value of each temporary overvoltage is defined in the insulation co-ordination study. We carry out your insulation co-ordination study and quote the scope for your case, remotely and worldwide.
Related guides
- Insulation co-ordination under IEC 60071: the steps
- Earth fault factor and overvoltage on healthy phases
- Slow-front overvoltages: switching and fault
- Arrester protection level: Upl and Ups
- Insulation co-ordination: what it is and what it is for
References
- IEC 60071-2 (1996 edition), Insulation co-ordination — Part 2: Application guide, clauses 2.3.2 to 2.3.2.7 and annex B.
- IEC 60071-1:2006, Insulation co-ordination — Part 1: Definitions, principles and rules.
Frequently asked questions
What is the difference between a temporary overvoltage and a switching overvoltage?
A temporary overvoltage keeps the power-frequency waveform and lasts from cycles to hours. A switching overvoltage is an oscillatory, slow-front transient that lasts from tens to thousands of microseconds. They are handled separately because each calls for a different withstand test and a different protection.
Why does load rejection raise the voltage at the end of a long line?
When the breaker at the receiving end opens, the line capacitance is still fed from the source, and the capacitive current raises the voltage along the line. This is the Ferranti effect. It is stronger when the line is long and the short-circuit power at the source is low.
Does a surge arrester limit temporary overvoltages?
Not in any practical way. The arrester has to withstand them, which is set by its rated voltage and its power-frequency voltage versus time characteristic. The exception is resonance-type overvoltages, which an arrester can limit if the heating it imposes on the arrester is studied carefully.