11.1.1 Mild Electric Shocks

Electrical currents as low as 0.5 mA can provoke a reaction from a person touching a conductive surface, although this does depend on the force applied to the surface, whether the skin is wet or dry and other physiological characteristics of the individual concerned. Higher currents, of the order of 3–6 mA, though considerably more uncomfortable, are not debilitating and the person can release the conductor concerned in order to stop the current flow. However, if the current flowing through the body increases beyond this, then the person concerned may not be able to let go.

Dalziel conducted an experiment on a group of 28 women and 134 men to determine the current threshold beyond which letting go was not possible, the let‐go threshold. The results of this work are summarised in the graph of Figure 11.1. He determined an average let‐go threshold for men of 16 mA, and 10.5 mA for women. These values can be read from the graphs in Figure 11.1 at the 50th percentile. Half of the people in the sample tested were able to let go at currents above these values, and half were not. Today IEC 60479 assumes a let‐go threshold of 10 mA for adult males. At higher shock currents ventricular fibrillation and electrocution become more likely.

11.1.2 Electrocution and Ventricular Fibrillation

Electrocution is defined as death occurring as a result of electric current flowing through the body. The most common electrocution effects relate either to the disruption of the heart’s natural pumping rhythm, known as cardiac arrest, or to a cessation of the respiratory system, i.e. the breathing response. Ventricular fibrillation is the most likely cause of cardiac arrest through electrocution. Both these effects prevent the body from receiving the oxygen it requires, either because (1) when in cardiac arrest, the heart fails to effectively pump blood, or (2) if breathing ceases insufficient oxygen enters the bloodstream. In either event the end result is the same.

Figure 11.2 shows a view of the heart and a typical electrocardiogram (ECG), which is a graph representative of the electrical potentials developed within the muscles of the heart. The numbers in Figure 11.2 designate the propagation of the muscle contraction with time throughout the heart.

Of particular interest is the T wave on the right‐hand side of the ECG. The ventricular muscles are recovering during this period, and at this time in the cycle the heart is particularly vulnerable to ventricular fibrillation, triggered by an external shock current. Figure 11.3 shows the effect of fibrillation, arising from such a disturbance, introduced early in the T wave.

Following a shock of sufficient magnitude and duration, the regular ECG waveform is destroyed and as a result the ventricular muscle begins to contract erratically. The pumping action of the heart ceases and, as shown, the blood pressure falls. Unless the heart can be quickly restarted, death is very likely.

Through his research, Dalziel was able to show that the body responds to high amplitude short duration shocks in a similar way to low amplitude long duration ones. He summarised his results in an empirical equation that predicts the shock current flowing through the heart, at the threshold of ventricular fibrillation:

(11.1)

where I = the threshold ventricular fibrillation AC current (mA), t = shock duration in seconds (0.03 < t < 3.0 sec), K is a constant relating to the body weight of the person concerned, (K = 116 for a 50 kg person and K = 157 for a 70 kg person).

The energy delivered as a result of any shock is proportional to the I²t product, and to the body resistance R_b, thus we can write:

(11.2)

The shock energy required to trigger ventricular fibrillation is therefore proportional to:

(11.2a)

Shock energies in excess of 50 joules can be sufficient to trigger ventricular fibrillation in individuals, the prevention of which is vital if injury or death is to be avoided. Therefore a well‐designed earthing system must ensure that prospective shock energy is kept below the threshold of ventricular fibrillation.

While Equation (11.1) strictly applies for AC currents between 15 and 100 Hz, a similar result also applies for direct current. Direct current is less effective in causing fibrillation than is AC current. This is partly because the stimulation of nerves and muscles is associated with changes in the magnitude of the current flowing. Once established, direct current is roughly constant in magnitude, and as a result it is less likely to trigger ventricular fibrillation.

Experimental results suggest that the magnitude of the DC current required to trigger fibrillation is about 3.75 times greater than its AC equivalent. IEC 60479 defines a DC/AC equivalence factor (k) as follows:

(11.2b)

Equations (11.1) and (11.2b) can be combined to yield the graph in Figure 11.4, which shows the AC and DC ventricular fibrillation threshold currents as a function of shock duration. Clearly, as the duration of the shock increases, considerably less current is required to trigger ventricular fibrillation. It is interesting to note, however, that for shocks of 3 seconds duration, DC currents around 300 mA are required to trigger fibrillation, whereas AC currents of the order of only 80 mA will achieve the same result.

11.1.3 Electrical Resistance of the Human Body

The severity of a shock depends upon many things, including the magnitude of the shock current, its duration and timing with respect to the cardio rhythm, the points of entry and exit from the body, the associated body resistance between these points, and the physical characteristics of the person concerned.

Given all these variables, it is not surprising that people respond very differently to mild electric shocks. What may seem painful and unpleasant to one person may scarcely be felt by another. One parameter that is useful in estimating the shock current a person may receive is the body’s electrical resistance. Figure 11.5 shows average values of body resistance for fractions of the general population as a function of the shock touch voltage. These are plotted for 5, 50 and 95 percentiles. The body resistance of 95% of the population lies below the upper curve, 50% of the population exhibit resistances below the middle curve, and only 5% of the population have body resistances below the lowest curve. These curves apply to either hand‐to‐hand or hand‐to‐foot current paths, both of which are possible shock scenarios.

The 5th percentile curve in Figure 11.5 is of interest in determining the worst case prospective AC or DC shock current, since it represents a body resistance lower than that possessed by 95% of the population. This curve suggests that an asymptotic body resistance of 1000 ohms is appropriate for AC or DC voltages up to 700 volts. Therefore the worst case prospective shock current can be estimated by measuring that drawn by a 1000 ohm resistance placed across any prospective shock potential.

The IEEE Standard 80 Guide for Safety in AC Substation Grounding assumes a value of 1000 ohms for body resistance. It also conservatively neglects the effects of shoe and glove resistances in the calculation of touch and step potentials.

11.1.4 Electrical Shock Current: (How Much is Too Much?)

With the ability to estimate shock currents in this way, it is reasonable to ask ‘How much shock current is too much?’ IEC 60479 provides guidance on this, in the form of a graph showing time–current zones for AC (Figure 11.6). This is somewhat similar to Figure 11.4 in that it considers the shock current, as a function of shock duration, but here the current spectrum is divided into four zones. The physiological effects of each are as follows:

Zone AC‐1: There are usually no reaction effects, owing to the small shock current flowing.

Zone AC‐2: Usually no harmful physiological effects.

Zone AC‐3: Usually no organic damage to be expected. Likelihood of muscular contractions and difficulty in breathing, reversible disturbances of formation and conduction of impulses in the heart, including atrial fibrillation and transient cardiac arrest without ventricular fibrillation, increasing with current magnitude and time.

Zone AC‐4: In addition to the effects of zone 3, probability of ventricular fibrillation increasing up to about 5% (zone AC‐4.1), up to about 50% (zone AC‐4.2) and above 50% beyond (zone AC‐4.3). Increasing with magnitude and time, pathophysiological effects such as cardiac arrest, breathing arrest and heavy burns may occur.

The body can tolerate the effects of quite large AC currents (200–500 mA), provided the shock duration is kept very short. Unfortunately, as the shock current magnitude increases, it becomes more difficult for victims to extricate themselves from the source of the current, and the longer the duration of the shock the more likely is the onset of ventricular fibrillation.

The technical report IEC 61200‐413 entitled Explanatory Notes to Measures of Protection against Indirect Contact by Automatic Disconnection of Supply defines permissible LV touch voltage exposure limits, illustrated in Figure 11.7. In order for the time–current zones defined in Figure 11.6 not to be violated as the prospective touch voltage rises, the duration of exposure must be reduced. In practice, this is achieved through the operation of fast‐acting protection devices isolating the LV supply.

Figure 11.7 shows two permissible LV touch voltage curves, L and Lp. The former relates to normal operation in dry conditions, and corresponds to a sustained touch voltage of 50 volts, known as the safety voltage. The latter is designed to cater for wet areas in which the body’s skin resistance is considerably lower. This corresponds to a sustained touch voltage of 25 volts.

These curves can be justified by assuming that the asymptotic body resistance of 1000 ohms in Figure 11.5 applies. The permissible touch voltage can then be plotted against duration on the time–current graph shown in Figure 11.8, since the touch voltage is approximately equal to the product of the shock current and the body resistance. Figure 11.8 shows that the resulting exposure curves avoid the likelihood of ventricular fibrillation.

Provided that protection devices are chosen whose clearance times are less than the maximum permitted duration, then the likelihood of severe injury is remote. The prospective touch voltage that occurs as a result of a phase‐to‐ground fault depends to a large extent upon the LV earthing system used.

Electric shocks result from either direct or indirect contact with an exposed source of current. Direct contact occurs when contact is made with a live phase conductor. This is generally prevented by the use of both insulation materials and physical barriers in electrical installations.

Indirect contact occurs through contact with metallic components that have been accidentally elevated above earth potential as the result of an insulation failure elsewhere in an installation. Thus from the viewpoint of electrical safety, we are primarily interested in the effects of phase‐to‐ground faults.

Protection against indirect contact is provided by the earthing system, which limits the touch voltage and by the protection system which clears the fault within an acceptably short period of time. The latter is accomplished through the use of fuses or magnetic circuit breakers, or in the case of small ground fault currents, through the application of residual current devices.

11.2.1 Operation of the RCD

The RCD operates on the principle of a differential transformer, as shown in the two‐wire circuit of Figure 11.9. It consists of a magnetic circuit upon which there are three windings. Two of these carry load current, one from the phase terminal to the load, the other from the load back to the transformer neutral connection. These windings are wound anti‐phase with respect to each other, so that when they carry the same current their ampere‐turns cancel, resulting in zero flux developed in the core.

The third winding senses any current imbalance between the other two. Thus, in the event that leakage current does find its way to ground, the core imbalance will result in a residual flux, proportional to the residual current flowing in the earth conductor(s). The resulting sense voltage is rectified, filtered and applied to the trip coil of the isolation switch, which breaks both source conductors.

The isolation switch is held closed by a permanent magnet contained within its magnetic circuit. When the trip current threshold is exceeded, the DC flux in the circuit is nulled, and the isolator opens under the action of a spring. Once open, the air gap in the magnetic circuit is sufficient to prevent re‐closure, and the isolator must therefore be reset manually.

The test button causes a small current to be bled from the outgoing active terminal to the neutral, without it passing through the neutral winding. This creates an imbalance in the core, and the resulting voltage in the sense coil trips the isolator. This facility enables users to periodically check that the RCD is functioning correctly.

In addition to the two‐wire RCD described above, three‐ and four‐wire circuits are also possible, so long as the protective earth conductor is not one of them. Under normal operation the vector sum of the currents flowing through the toroid adds to zero. Therefore any imbalance that does occur represents a genuine residual ground current, on which the unit will trip.

There are a couple of items relating to RCDs that should be stressed. Firstly, an RCD will not limit the magnitude of the ground fault current flowing; it will only limit the duration for which it flows. Secondly an RCD will not trip in the event of a phase‐to‐phase or a phase‐to‐neutral fault, since these do not affect the ampere‐turn balance within the core.

11.2.2 RCD Sensitivities and Operating Times

The sensitivity of an RCD is the rated residual operating current, given the symbol I_Δn. Table 11.1 shows high, medium and low sensitivities commonly used in RCDs. Low and medium sensitivities are used for the protection of equipment and to assist in the prevention of fires ignited during low level insulation faults. Medium and low sensitivity RCDs are also used together to allow discrimination between circuits in order to isolate the faulty circuit.

The high sensitivity RCDs are used for the purpose of shock prevention. In general residential applications, 30 mA is generally the trip threshold setting, but in hospitals and medical centres 6 or 10 mA may be used instead. In the USA, trip thresholds as low as 5 mA are also used.

The IEC standard 61008 defines two types of RCD based on tripping times. Type G (general) use RCDs will trip within 300 ms for ground fault currents less than 2I_Δn while type S (selective) devices are fitted with a short operating delay which can also be helpful in achieving discrimination between circuits. Figure 11.10 shows the tripping characteristic for both these types, as multiples of the RCD operating current.

Finally, Figure 11.11 shows a residential RCCB rated at 30 mA which can carry a through current of 40 A. The test button is located at the top of the device, and the isolation switch is below it.

11.2.3 Limitations of RCDs

RCDs are very good at doing what they’re designed for, but there are several situations in which they will not operate, and these should be understood. Firstly, a three‐phase RCD will not operate in the event of a phase‐to‐phase or a phase‐to‐neutral fault, regardless of its magnitude. This is because such faults do not generate residual currents that flow back to the star point, via one or more earth conductors. Secondly, a fault on the secondary side of a transformer will not cause an RCD on its primary winding to trip, for precisely the same reason.

There are also some situations in which a single‐phase RCD may trip unexpectedly. These situations arise when a parallel path from the appliance neutral to the supply star point is inadvertently created. This can occur, for example, when using an oscilloscope, if the earth terminal of either channel is connected to the neutral conductor of an appliance protected by an RCD. By so doing, a parallel path is created back to the transformer’s star point, via the oscilloscope’s earth conductor. This path is likely to be slightly higher in impedance compared with the neutral conductor, but it may permit just enough neutral current to flow to unbalance and trip the RCD. Unexplained RCD trips should always be investigated.

11.3.1 TT Earthing System

The TT earthing system sees both the transformer neutral and the customer’s switchboard connected to earth, as shown in Figure 11.12. Therefore any phase‐to‐ground fault current is limited by the earth impedances both at the customer’s premises and at the transformer. Assuming the fault impedance to be zero, the fault current, I_f is approximately given by:

where V_ph is the faulted phase potential, Z_f  = the fault loop impedance = , Z_S is the source impedance (largely that of the local MV‐LV transformer), Z_L is the impedance of the active conductor, Z_PE is sum of the protective earth conductor impedance and the customer’s earth impedance and Z_T is the transformer earth impedance.

Of these impedances, Z_PE and Z_T are generally the largest. Therefore the associated touch potential V_t is approximately:

Z _PE and Z_T are likely to have similar magnitudes and therefore . Since this is in all cases beyond the permissible continuous touch voltage of 50 volts (or in wet areas 25 volts), then the protection equipment must limit the exposure duration to a value less than that specified in Figure 11.7.

The fault current in a TT system, being limited by earth impedances, may not be sufficient to operate the installation’s short‐circuit protective devices (fuses or magnetic circuit breakers), which must be dimensioned according to the rated load current. Therefore residual current devices like those shown in Figure 11.13, must be used to detect and clear the fault, in which case we require:

where IΔn is the operating current of the RCD and V_Lt is the maximum sustained touch voltage (also known as the safety voltage).

11.3.2 TN Earthing System

The TN system of LV earthing has the customer’s switchboard and all other exposed metallic components solidly connected to the LV transformer neutral, which itself is earthed at the transformer. The phase‐to‐neutral fault impedance is therefore relatively low, and large fault currents can flow as a result. These are often well in excess of normal load currents and therefore the short circuit protective devices (SCPDs) will quickly clear such a fault. The TN earthing system is pictured in Figure 11.14.

Because of the heavy fault currents flowing through the transformer and supply conductor impedances, the phase voltage V_ph seen by the customer will fall during the fault, typically by at least 20%.

The fault current for a phase to neutral fault is therefore given by:

And the touch voltage V_t is equal to the drop in the protective earth conductor, thus:

If these impedances are approximately equal then:

Therefore in 230/400 volt systems the touch voltage will be about 90 volts, and since this is in excess of the sustained touch voltage limit of 50 volts, then according to Figure 11.7 the SCPD must clear the fault in less than 450 ms. In cases where the impedance of the supply conductors is high or unknown, a combined circuit breaker/RCD may be used, i.e. an RCBO.

The satisfactory operation of SCPDs requires a fault current of sufficient magnitude that will ensure operation of the fuse or magnetic circuit breaker within the required clearance time. The fault current in turn depends upon the impedance of the faulted circuit, known as the fault loop impedance. This impedance, defined in Figure 11.15, includes the source and active conductor impedances to the fault, as well as the return path via the protective earth and neutral (PEN), back to the neutral terminal of the transformer, i.e. path 1‐2‐3‐4‐5‐6. The fault impedance is assumed to be zero.

The fault loop impedance, Z_f is sufficiently small if the following equation is met:

where Z_f is the fault loop impedance, V_ph is the faulted phase voltage and I_a is the current that will ensure operation of the SCPD (fuse or magnetic circuit breaker), within the required clearance time. (This is often taken to be the instantaneous trip current of a circuit breaker.)

Due to the relative size of the distribution conductors in the power network compared to those within the customer’s premises, it is assumed that at least 80% of the faulted phase potential is dropped within the customer’s premises. As a result, there exists a maximum route length for customer circuits, beyond which the SCPD will not operate within the required time. The maximum route length can be calculated using the following equation, which assumes negligible inductive conductor impedance (i.e. that the phase and PEN conductors run in close proximity to each other and are therefore predominantly resistive).

where L_max is the maximum route length in metres, ρ is the resistivity at the working temperature Ωmm²/m, (22.5 × 10⁻³ for Cu or 36 × 10⁻³ for Al), S_ph is the customer’s phase conductor cross‐sectional area in mm² and S_pen is the customer’s PEN conductor cross‐sectional area in mm².

Should a proposed circuit length exceed L_max, then the cross‐section of the conductors must be increased or the circuit’s current rating must be reduced. Larger conductors may also be demanded by the minimum voltage drop requirements for circuits approaching this length.

Where local regulations permit, an RCD may be used to protect long circuits. This makes the calculation of the fault loop impedance unnecessary. In such cases, the RCD threshold must be chosen so that:

(This solution cannot be applied in the case of TN‐C earthing, since here the neutral and the protective earth are the same conductor, see the following section.)

Variations on the TN Earthing System

Figure 11.16 shows several variations on the TN earthing system. These are designated TN‐C, TN‐S and TN‐C‐S. The TN‐C system (Figure 11.16a) uses a common protective earth and neutral (PEN) conductor, whereas the TN‐S system (Figure 11.16b) separates these conductors. While a combined neutral and PEN saves one conductor, it has the disadvantage that the return current flowing back to the neutral terminal of the transformer generates a small potential drop between the earth potential at the transformer and that at the customer’s premises. Further, in the presence of a significant triplen harmonic component¹, the neutral may become overloaded, particularly if its dimensions are smaller than those of the associated phase conductors. In France, for example, the TN‐C system is only permitted when the size of the PEN conductor exceeds 10 mm².

In the TN‐S system the functions of the neutral conductor and the protective earth are separated by the provision of a distributed earth conductor, shown in Figure 11.16b and 11.16c. The TN‐S system offers the opportunity of isolating the neutral conductor via the SCPD, while maintaining the protective earth connection. (It is not permitted to connect a TN‐C system downstream of a TN‐S.)

Earthing of the Neutral Conductor

Because of the possibility that the PEN may carry sizeable fault currents, in many countries it is common to earth this conductor at frequent intervals along its length, thereby providing a global grounding system in which the division of earth fault currents provides a reduction in the earth potential rise (EPR) that otherwise might occur. In the USA and the UK when a TN‐C system is used, multiple earth connections are made at regular intervals to the PEN in the public network, while in Germany this conductor is earthed upstream of each customer connection.

In Australia and New Zealand a variation to the TN‐C earthing system is used, one known as the multiple earthed neutral (MEN) system, in which each customer connection includes an earth rod on the customer’s premises, to which the PEN conductor is connected via an MEN connection between the earth bar and the neutral bar in the customer’s switchboard, as depicted in Figure 11.17 Thus the MEN system is effectively TN‐C upstream of the customer’s switchboard and TN‐S downstream of it.

All exposed metal surfaces are connected to the switchboard earth conductor, including water and gas reticulation piping. The earthing of the neutral conductor at each customer’s earth electrode ensures the existence of a global grounding system both in residential and city areas.

As with all TN systems, a phase‐to‐neutral fault in the MEN system will result in large fault currents that will be cleared by the associated protective device(s). Assuming that the phase and neutral conductors have the similar cross‐sectional areas, the touch potential V_t will generally be about 0.4 V_ph, as derived earlier. Provided the fault is cleared quickly (typically within 400 ms), there should be minimal risk of fire or electrocution. RCDs can be successfully used within the customer’s premises, since the neutral and the earth conductors are separate, and RCDs are now incorporated into all new Australian MEN installations as well as in older ones as they are upgraded.

11.3.3 IT Earthing System

In the IT earthing system, the LV transformer neutral is not earthed, or alternatively it is earthed through a high impedance. This may be from a few ohms in the case of resistively earthed systems to several hundred kilo‐ohms for fully floating systems, as illustrated in Figure 11.18. Exposed metal surfaces within the customer’s premises are connected to ground. Despite the lack of a solid earth connection, the LV neutral point remains close to ground potential. This is due to the presence of distributed network capacitances between each phase and ground (Figure 11.18), which effectively tie the LV neutral‐to‐ground potential. These capacitances arise in LV cables between the phase conductor and the earthed sheath, in transformers between phase windings and the earthed core, and to a lesser extent between overhead lines and ground. These capacitances are small and are approximately equal on each phase; they generally lie in the range 0.1–5 μF depending on the size of the LV network. As a result, under normal conditions the total positive sequence capacitor current flowing back to the transformer neutral is close to zero.

Because of the high impedance of both the earth and the distributed capacitances, the presence of a single‐phase‐to‐ground fault will not generate significant fault current, and therefore the resulting touch potential will be very low. As a result, the system can remain operational in the presence of the first fault, but this must be located and rectified before a second fault develops, which may result in high fault currents and therefore high touch voltages as well. The ability of the LV system to remain operational in the presence of the first fault is a particular advantage of IT earthing; for this reason, it is often used in applications where the AC supply is critical, such as hospitals and medical centres.

The fault current that flows for the first phase‐to‐ground fault can be calculated with reference to Figure 11.18 as follows. Because the neutral normally lies very close to ground potential, the current I_f will also be close to zero. However, in the event of a phase‐to‐ground fault on phase a, as shown, the capacitances associated with phases b and c will now each see a line voltage. The phasor diagram in Figure 11.19 summarises this situation. Ic_b, Ic_c and Ic_n will flow through ground towards the transformer star point. The magnitudes of Ic_b and Ic_b will be given by:

and their vector sum will equal:

Assuming that the neutral‐to‐ground capacitance is of a similar size to the phase capacitances, then the current Ic_n will equal V_phωC and the total fault current will then be:

For a 230/400 V, 50 Hz system, with C = 5 μF then I_f = 1.4 A and the resulting touch voltage V_t, will be given by:

Since V_t is less than the continuous touch voltage of 50 volts, even for relatively high earthing resistance and phase capacitances, the LV system can remain in operation while the fault is located and repaired. However, the occurrence of a second fault before the first has been repaired can lead to hazardous touch voltages being generated.

For example, consider a second fault, this time between neutral and ground, as shown in Figure 11.20. The second fault effectively generates a phase‐to‐neutral fault with fault current I_f flowing in the PE conductor. The associated voltage drops in the phase and PE conductors can create dangerous touch voltages of the order of 0.4 V_ph, which must be cleared quickly. The magnitude of the phase‐to‐neutral fault will generally be such that the associated SCPD will clear the fault, but if the distance between the two faults is sufficiently large, then the cabling impedances may reduce the fault current below the threshold of the SCPD and then RCDs may be required for fault clearance instead.

Where the second fault to ground occurs on another phase, the interrupting circuit breaker must now interrupt a phase‐to‐phase fault which must also be cleared quickly. It is therefore a necessary requirement that all circuit breakers used in IT earthing systems be capable of interrupting a single pole phase‐to‐phase fault.

A maximum route length can also be calculated for circuits within an IT earthing system, in a similar way to that already discussed for the TN system. In this case, the following equations apply:

1. For IT circuits where the neutral is distributed, the fault potential will be a phase voltage, and the maximum route length is given by:
   
2. For IT circuits where the neutral is not distributed the fault potential will be a line voltage, and the maximum route length is given by:
   

(Note that a factor of 1/2 has been applied to both these equations since there is now twice the length of each phase and protective earth conductor within customer premises, as compared with the TN earthing system. This is shown in Figure 11.20.)

Locating the first fault in an IT earthing system is not easy, due to the small capacitive currents involved and the need to determine which LV feeder is responsible for the fault. One obvious method is to disconnect feeders in turn and note which carries the fault current, but this technique requires interruption to the supply, which defeats the main advantage of the IT system in being able to operate safely through the first fault. More recent approaches include the injection of a low frequency signal between the transformer neutral and ground. The resulting current can then be traced to the offending feeder with handheld or permanently installed detectors, tuned to the injected signal frequency. It is expected that an insulation monitoring device will be installed between the transformer neutral and ground in an IT system, to highlight the existence of the first fault.

11.5.1 Un‐Earthed MV Neutrals

In some countries, the MV neutral is unearthed, in a similar fashion to the IT system of LV earthing. In such cases the network capacitance between each phase and ground, ties the MV neutral terminal close to ground potential. As with the IT earthing system, the first phase‐to‐ground fault will generate a small residual capacitive fault current which, depending on the nature of the fault, may result in an arc. From the point of view of disruptive breakdown, such a fault will not generate large currents, and as a result the transfer potential seen in the LV network will generally be low. The MV supply can therefore remain in service until the fault is found.

Should an arc occur, however, significant over‐voltages may be generated. This is as a result of the intermittent nature of the arc current, the stray inductance associated with the fault path and the capacitance associated with the faulted phase. The arc current flowing in the fault extinguishes as it passes through zero; this occurs at the instant when the voltage across the fault capacitance is at its maximum (since a capacitive current leads the voltage by 90°). Thus when the fault first occurs, the faulted phase is pulled to ground potential, and the neutral terminal therefore rises by an amount equal to the peak voltage of the faulted phase. The arc current extinguishes at this time and the associated capacitance retains this charge thereafter. The system of phase voltages now operates one peak phase voltage above ground potential. Then, 180° later, the faulted phase achieves a potential equal to twice the peak phase voltage above ground, which may be sufficient for the arc to restrike. Arc fault current now flows through the capacitor to ground via any stray inductance, included within the fault path. This LC circuit rings at its resonant frequency, and the potential on the capacitor becomes reversed. As before, the arc current extinguishes as it passes through zero, having existed for about a half cycle of the LC resonant frequency. The faulted phase now lies at a potential of −2 times the peak phase potential, and after a further 180° this phase voltage reaches −4 times this level. The arc will again likely reignite and flow in the LC circuit, once again reversing the potential on the capacitance. This process repeats, and thus the neutral to ground voltage alternates in sign and grows in magnitude until eventually an insulation failure occurs somewhere in the circuit. This may be in the transformer itself or in MV cabling associated with it. The insulation requirement of transformers and associated equipment that are to be used on ungrounded systems must therefore be considerably higher than those used in solidly grounded systems.

Figure 11.23 shows the capacitive current paths as a result of a fault on phase c of an MV transformer. Under this condition the distributed capacitances C_a and C_b are each excited by a line voltage, and the resulting capacitive currents I_Ca and I_Cb therefore flow in the fault.

As for the IT LV earthing system, the capacitive current magnitudes are given by:

and their vector sum equals:

Similarly with IT earthing networks, insulation monitoring is required between the neutral terminal and ground to detect the presence of the first fault, which may be located while the MV system is in operation.

Unearthed MV neutrals are sometimes used in Germany, Italy, Ireland and Japan, but in order to remove the risk of over‐voltages generated through arcing faults, either impedance earthing or resonant earthing is frequently used instead.

11.5.2 Resonant Earthing (Compensated Earthing)

Resonant earthing is a novel system devised by Professor W Petersen in 1916. It uses a reactance inserted between the MV transformer neutral and ground (the Petersen coil) to resonate with the distributed network capacitance during phase‐to‐ground faults. Such an arrangement appears in Figure 11.24.

The capacitive current flowing to ground during a phase‐c‐to‐ground fault is given by:

Also flowing in the fault is the inductive current, set up by the voltage on the faulted phase and the neutral reactance X_L. It is given by:

By choosing the Petersen coil reactance appropriately, the total fault current can be reduced to near zero. Thus:

Therefore

For this to occur we require that:

This reduction in the magnitude of an earth fault current has the desirable characteristic of extinguishing most arc faults and thereby virtually eliminates the over‐voltage phenomenon that often accompanies them. It also allows the MV distribution system to continue operating until the fault can be located and possibly even be repaired. For this reason the Petersen coil is also referred to as an arc suppression coil or a ground fault neutraliser.

To achieve complete compensation, the LC circuit must be resonant at times the system frequency. Because the network’s distributed capacitances vary with its configuration and with the connected load, it is necessary for the Petersen coil to be dynamically tuned to the network. This is achieved by altering a motor‐controlled magnetic shunt within the reactor’s magnetic circuit. Since the distributed capacitances are seldom exactly balanced, a small neutral current flows even during normal operation, creating a small zero‐sequence voltage across the Petersen coil. This voltage is monitored, and the Petersen coil is tuned until it is minimized.

The cancellation of the residual capacitive current by the inductive current from the Petersen coil is seldom exact, and losses in the Petersen coil also result in an additional small current, in phase with the faulted winding potential flowing in the fault. Losses associated with the Petersen Coil can be modelled by the inclusion of a parallel connected resistance. This resistance is often augmented with an external resistance, and these collectively generate a resistive component of fault current which aid in fault detection. Since only the inductive component of the fault current flows in the Petersen Coil, the EPR is small, and therefore the potential transferred to the LV network will be as well.

Ground fault detection in resonant earthed systems can be difficult. It is relatively easy to detect the presence of a phase‐to‐ground fault by monitoring the zero‐sequence voltage across the Petersen coil, but finding its exact location is not. One method is to progressively open each feeder in turn, until the fault condition disappears, but this is generally too disruptive and defeats the ability of a resonant earthed network to continue operating in the presence of the first fault. Detecting the small fault currents that occur in resonant systems may also be difficult, particularly if the compensation is nearly complete. One approach is to install a zero‐sequence wattmetric relay on each feeder that originates from the transformer. (See problem 1, Chapter 12.)

11.5.3 Impedance Earthing

Impedance earthing of MV systems requires the inclusion of either a resistor or a reactor between the MV neutral terminal and ground. This earthing impedance can be used to limit the magnitude of the ground fault currents and hence the damage they do, and it is usually dimensioned so that the resulting fault currents are less than 1000 A. This limits the EPR in the case of faults in MV‐LV transformers described above, and thus the potential transferred to the LV network.

Ground fault currents in impedance earthed systems are sufficiently large as to permit easy detection, by measuring either the zero‐sequence current or voltage at the neutral terminal; however, these generally require the immediate disconnection of the supply. Resistive grounding impedances offer the advantage of damping over‐voltages, but must be physically large to cater with the energy dissipated. On the other hand, inductive impedances dissipate very little energy but can create large over‐voltages during fault clearance.

11.5.4 Over‐Voltages

While the inclusion of an impedance between the neutral terminal and ground will reduce the phase‐to‐earth fault level, it does generate an over‐voltage problem during a fault. This is because the voltage dropped across this impedance forces the neutral potential to rise. In the case of a solid phase‐to‐ground fault, the neutral potential will rise by an amount almost equal to the voltage of the faulted phase, and as a result the potential between healthy phases and ground becomes almost equal to the pre‐fault line voltage. This over‐voltage has implications for the winding insulation within the transformer, since it must now withstand an elevated potential during fault conditions.

Effectively Earthed Systems

IEC60044.2 (inductive voltage transformers), defines earth fault factor as the ratio between the maximum phase‐to‐ground voltage of a healthy phase during an earth fault, to the pre‐fault phase‐to‐ground voltage. Distribution systems where the neutral point is solidly grounded are called effectively earthed, and in such systems the earth fault factor is less than 1.4. This is because in a solidly earthed system during a ground fault, the maximum voltage to ground of the healthy phases cannot exceed 80% of the phase to phase voltage, thus:

For an effectively earthed system we require that and , where X₁ is the positive sequence impedance of the system at the fault and X_o and R_o are the zero‐sequence reactance and resistance respectively. From the foregoing discussion ungrounded, resonant and impedance earthed systems are not effectively earthed.